Effect of Different Solvents on Morphology and Gas-Sensitive Properties of Grinding-Assisted Liquid-Phase-Exfoliated MoS₂ Nanosheets

Abstract

Grinding-assisted liquid-phase exfoliation is a widely used method for the preparation of two-dimensional nanomaterials. In this study, N-methylpyrrolidone and acetonitrile, two common grinding solvents, were used during the liquid-phase exfoliation for the preparation of MoS₂ nanosheets. The morphology and structure of MoS₂ nanosheets were analyzed via scanning electron microscopy, X-ray diffraction, and Raman spectroscopy. The effects of grinding solvents on the gas-sensing performance of the MoS₂ nanosheets were investigated for the first time. The results show that the sensitivities of MoS₂ nanosheet exfoliation with N-methylpyrrolidone were 2.4-, 1.4-, 1.9-, and 2.7-fold higher than exfoliation with acetonitrile in the presence of formaldehyde, acetone, and ethanol and 98% relative humidity, respectively. MoS₂ nanosheet exfoliation with N-methylpyrrolidone also has fast response and recovery characteristics to 50–1000 ppm of CH₂O. Accordingly, although N-methylpyrrolidone cannot be removed completely from the surface of MoS₂, it has good gas sensitivity compared with other samples. Therefore, N-methylpyrrolidone is preferred for the preparation of gas-sensitive MoS₂ nanosheets in grinding-assisted liquid-phase exfoliation. The results provide an experimental basis for the preparation of two-dimensional materials and their application in gas sensors.

1. Introduction

Given the special structure and potential applications, two-dimensional (2D) materials such as graphene, boron nitride, and molybdenum disulfide (MoS₂) draw plenty of concerns. Among them, MoS₂ as the frontrunner in transition metal dichalcogenides (TMDCs) materials has gained the most attention [1,2,3,4] and is used in a wide variety of applications [5,6,7,8,9,10,11] due to its unique properties [12,13,14]. MoS₂ is at the forefront in the race of an ideal gas-sensing material because of its large surface-to-volume ratio, enormous number of active sites, and favorable adsorption sites [15,16]. MoS₂ manifests two possible crystal phases, including trigonal and hexagonal structures, with metallic and semiconducting properties, respectively [17]. The presence of weak Van der Waals force facilitates the isolation of layers from bulk MoS₂. The indirect bandgap of 1.2 eV in bulk MoS₂ is converted to a direct bandgap of 1.8 eV for monolayer MoS₂ [3,14,18]. The absence of dangling bonds provides stability to pristine MoS₂ flakes in liquid and gaseous media in the presence of oxygen, thereby facilitating its gas-sensing application [19,20]. Therefore, a reliable and low-cost technique is needed to produce 2D-MoS₂ for gas-sensing applications. Currently, several methods including vapor deposition [21], mechanical exfoliation [22], lithium-ion intercalation [23], liquid-phase exfoliation [24,25], and RF sputtering [26] have been utilized to fabricate 2D-MoS₂ nanomaterials. Although high-quality MoS₂ nanosheets were prepared by mechanical methods for fundamental research, it is difficult to meet the need for mass production. Meanwhile, hydrothermal and solvothermal methods yield few-layer MoS₂ nanosheets on a large scale. However, they often require high temperature and high pressure. Lithium intercalation into the layered structure of 2D-MoS₂ is limited by long lithiation time, high temperature, and sensitivity to environmental conditions. Instead, grinding-assisted liquid-phase exfoliation is not air-sensitive, does not entail chemical reaction, and generally has acceptable yield [3,27,28]. Coleman et al. [29] evaluated multiple solvents for ultrasonic exfoliation of materials. They found that the most effective solvent was N-methylpyrrolidone (NMP), followed by acetonitrile (ACN). Yao et al. [25] reported that relatively high yields up to 26.7 mg/mL were obtained by incorporating NMP as a grinding solvent into the exfoliation procedure because the surface energy of NMP is similar to that of MoS₂. As a result, NMP is the preferred solvent in liquid-phase exfoliation to obtain single or multilayered 2D nano-materials [25,27,28,29,30,31].

The studies suggest that grinding solvent plays an important role in grinding-assisted exfoliation because its physical properties, such as boiling point [32], surface tension and energy [29,33], as well as solubility parameters [24], can affect the final exfoliation materials. Emily et al. reported that the lateral size and thickness of the exfoliated flakes of MoS₂ nanosheets are highly dependent on the solvent. NMP yielded flakes of the highest quality based on lateral size and flake thickness. The NMP remained on the surface of the MoS₂ nanosheets when ACN was completely removed [34]. Good yields were obtained when using NMP as a grinding solvent. However, whether the NMP residue affects the performance of electronic devices is unknown. It may adversely affect the application of MoS₂ nanosheets in gas sensing. To our knowledge, there is no report on the research of the effects of a grinding solvent on the gas-sensing properties of MoS₂.

Therefore, we evaluated the effects of residual NMP on the morphology and gas-sensing properties of liquid-phase-exfoliated MoS₂ nanosheets. We selected ACN, which has a lower boiling point and easier solvent removal compared with NMP. The morphology and structure of MoS₂ nanosheets were analyzed by scanning electron microscopy (SEM), X-ray diffraction (XRD), and Raman spectroscopy. The effects of grinding solvents on the gas-sensing performance of MoS₂ nanosheets were investigated for the first time.

2. Materials and Methods

2.1. Preparation of Materials

MoS₂, with a purity of 99% and particle size less than 2 μm, was purchased from Sigma-Aldrich. ACN, NMP, and absolute ethanol (C₂H₆O) were purchased from Tianjin Zhiyuan Chemical Reagent Co. Ltd. as analytically pure reagents. The preparation of MoS₂ nanosheets via grinding-assisted liquid-phase exfoliation is described as follows:

MoS₂ powder (100 mg) was manually ground in a mortar for 2 h, and 0.5 mL of the chosen solvent was added during the grinding. The sample was then dried in a vacuum oven at 60 °C for 12 h. The dried sample was dispersed in 40 mL of 45 vol% absolute ethanol and sonicated for 1 h at 120 W with stirring. The dispersion was centrifuged for another 20 min (1500 r / min) to obtain the MoS₂ nanosheets, and the supernatant was dried in air for further use. For convenience, the MoS₂ nanosheets obtained by grinding with ACN were designated as S1, and those ground with NMP were called S2.

2.2. Characterizations

The morphology of MoS₂ nanosheets was observed with a field emission scanning electron microscope (SEM, JSM-7610F Plus). The crystal structure of MoS₂ nanosheets was characterized by X-ray diffraction (XRD, Bruker D8 Advance, with Cu-Kα radiation). Raman spectroscopy (Renishaw inVia, Gloucester, Britain) was used to characterize the defects and functional groups of samples. The I-t and I-V curves of the sensing chip were measured by Keithley 2636B at room temperature.

2.3. Device Fabrication and Testing

The MoS₂ nanosheets were dispersed in absolute ethanol at 10 mg/mL. Dispersions (2 μL) were uniformly coated to fabricate a MoS_2-based sensing chip with Ag-Pd fork-finger electrodes. The minimum width and spacing of electrodes was 0.2 mm. The interdigital electrode was dried at 25 °C and aged for 24 h at a voltage of 4 V to obtain a sensing chip with good stability. The target vapor was produced by thermal evaporation, according to our previous work [35], and a calculated amount of target liquid was dropped onto a hot plate in a 1 L container to generate target vapor in the container. Next, 98% relative humidity was obtained by saturating salt solution (potassium sulphate-K₂SO₄). Then, by transferring the sensing chip from the air to the target gas at room temperature, the Keithley 2636B recorded the change of the current signal of the sensing chip (Figure S1). The response was defined using the formula $\left(\frac{I_G-I_R}{I_R}\right)\times 100\%$, where I_R and I_G are the currents of the sensor in the reference gas and target gas, respectively. The response time and recovery time were defined as the response values of 90% and 10% of the current of the sensor in contact with the target gas, respectively.

3. Results and Discussion

The XRD patterns of the two types of MoS₂ prepared by different grinding solvents are shown in Figure 1. Compared with JCPDS Card No. 73-1508, the lattice constants were: a = 3.15938 Å, b = 3.15938 Å, and c = 12.28962 Å. The diffraction peaks 14.39°, 29.02°, 32.69°, 33.51°, 35.88°, 39.56°, 44.27°, 49.81°, and 56.01° in the figures correspond to (002), (004), (100), (101), (102), (103), (104), (105), and (106) crystal planes of MoS₂, indicating that the materials were well-crystallized MoS₂. The peak intensity of MoS₂ nanosheets weakened, and the FWHM broadening (Figure S2) of the peaks appeared after liquid-phase exfoliation, indicating that the MoS₂ nanosheets were able to be exfoliated, and thus, the size of MoS₂ decreases [36,37,38,39].

Raman spectroscopy is effective in distinguishing bulk from exfoliated 2D materials. Figure 2 shows the Raman spectra of bulk MoS₂: S1 and S2. The two Raman peaks correspond to the high-energy $A_{1g}$ mode and lower-energy $E_{2\mathrm{g}}^1$ mode. As shown in Figure 2a, all the samples displayed the $E_{2\mathrm{g}}^1$ and $A_{1g}$ peaks of MoS₂. Comparing with peaks of bulk MoS₂, a red shift of $E_{2\mathrm{g}}^1$ peak and a blue shift of the $A_{1g}$ peak were observed for both S1 and S2, respectively. These shifts are associated with nanosheets obtained with NMP and ACN [40,41]. Figure 2b presents two very broad and intense Raman peaks (1360 and 1580- cm⁻¹) of S2, which may be assigned to NMP [31,36] that was not completely removed from the surface of MoS₂ nanosheets although it was heated and reduced at 60 °C for several hours. In contrast, S1 showed no broad peaks, indicating that ACN was almost removed.

We next investigated the effect of grinding solvents on the morphology of MoS₂ nanosheets. The SEM image shown in Figure 3a,b reveals the morphology of the starting MoS₂ powder as a thick layer with dimensions ranging from about 1 to 6.4 μm. The SEM images presented in Figure 3c,d clearly indicate that the lateral sizes and thicknesses of layered MoS₂ were reduced by combined grinding and sonication. The MoS₂ nanosheets were obtained by grinding with ACN (S1), as shown in Figure 3c,d, and the nanosheets were uniform in size and well-dispersed, with the majority measuring between 0.1 and 0.5 μm. As shown in Figure 3e,f, exfoliation with NMP (S2) also produced nanosheets with good dispersion with lateral dimensions of 0.4–1.6 μm. The MoS₂ nanosheets obtained by grinding with ACN were smaller than NMP-ground MoS₂ nanosheets, which is consistent with the results reported in the literature [34] and the results of XRD patterns (Figure 1 and Figure S2).

The gas-sensitive properties of MoS₂ nanosheets loaded on ceramic substrates were tested at room temperature. The results shown in Figure 4a,c indicate gas-sensitive properties and response time (Figure 4b,d) of S1 and S2 at 98% relative humidity (RH) and 1000 ppm of formaldehyde (CH₂O), acetone (C₃H₆O), and ethanol (C₂H₆O). The MoS₂ layers exfoliated with both the grinding solvents showed good stability in three continuous response–recovery cycles at room temperature. Both of them completed a response–recovery cycle in 40 s and returned completely each time with almost no drift.

Figure 5 shows the average response, response time, and recovery time of S1 and S2 for the target analyte. As can be seen from Figure 5a, the sensitivities of MoS₂ nanosheets exfoliation with NMP (S2) were 2.4, 1.4, 1.9, and 2.7 times higher than exfoliation with ACN (S1) to CH₂O, C₃H₆O, C₂H₆O, and 98%RH, respectively. These results prove that the MoS₂ nanosheets obtained by grinding with NMP have higher gas-responsive properties than the MoS₂ nanosheets with ACN although NMP was not removed completely. At the same time, it can be seen from Figure 5b,c that both samples have faster response time to the four analytes, which did not exceed 35 s, and the recovery time did not exceed 4 s.

In order to further evaluate the real-time monitoring capability of MoS₂ nanosheets obtained by grinding with NMP (S1), the responses of the S2-based sensor under different concentrations (50–2000 ppm) of CH₂O vapor were evaluated (Figure 6a). The response of S2 increased with the increase of CH₂O concentration. Figure 5b shows a linear response to changing CH₂O concentration, and the correlation coefficient R2 was 0.99, which facilitated gas-sensing application. Figure 6a,b show that the response time and recovery time of S2 were only 18 s and 0.5 s to 50 ppm CH₂O, respectively, and only 11 s and 0.6 s to 100 ppm CH₂O.

In order to comprehensively evaluate the gas-sensing performance of MoS₂ nanosheets obtained by grinding with NMP, the performances of the MoS₂ nanosheet-based sensors were compared (

Table 1. Sensing performances of recently reported CH₂O sensors.

Materials	Structure	Sensor Types	Con. (ppm)	Response (%)	LoD (ppb)	Temperature (°C)	Response Time (s)	Recovery Time (s)	Ref.
ZnS	0D nanosphere	Resistance	50	9440	-	295	11	8	[42]
In2O3/MoS2	Nanocubes/nonfilm	Resistance	50	75	200	RT	14	22	[43]
In2O3	Nanospheres	Resistance	1	3.5	1000	180	180	1000	[44]
In2O3/WS2	Nanocomposites	Resistance	5	7.5	-	RT	98	137	[45]
Au/TiO2	Hybrid films	Resistance	5	8.5	100	RT	36	110	[46]
rGo/MoS2	Hybrid films	Resistance	10	2.8	-	RT	-	-	[47]
Ni-doped In2O3/WS2	Nanocomposites	Resistance	20	32	150	RT	76	123	[45]
rGO/SnO2	Nanocomposites	Resistance	0.5	3200	10	125	31	62	[48]
MXene/NH2-MWCNTs	Hybrid films	Self-powered voltage	5	35	10	RT	51	57	[49]
MXene/ Co3O4	Hybrid films	Self-powered voltage	10	9.2	10	RT	83	5	[50]
MoS2	Nanosheets	Resistance	50	66.4	-	RT	18	0.5	This work

). As shown in Table 1, the response time and recovery time of MoS₂ nanosheets obtained by grinding with NMP for 50 ppm CH₂O were 18 s and 0.51 s, respectively, which were close to the shortest response time (11 s) and recovery time (8 s) shown by ZnS and In₂O₃/MoS₂ [42,43]. Nevertheless, compared with the operating temperature (295 °C) of ZnS, the operating temperature of MoS₂ nanosheets was at room temperature (25 °C). Therefore, the MoS₂ nanosheets exhibited a robust sensing performance at a low working temperature, with rapid response and recovery. However, the sensitivity and limit of detection (LoD) of the sensor based on pure MoS₂ nanosheets need to be improved.

Figure 7 shows the I–V curves of S1 and S2 measured with an applied bias voltage ranging from −2 to 2 V at 1000 ppm CH₂O. The I–V curves demonstrated a good ohmic contact between the sensing layers and the electrodes for both samples, which indicates that the sensor response was attributed to the sensitive material and not the metal–semiconductor contact.

The conductivity of the sensing material depends on the adsorption and desorption of gas molecules on the surface. When the MoS₂ sensor is exposed to air, the oxygen molecules are adsorbed on the surface of the MoS₂ nanosheets. Because of the strong electronegativity of oxygen atom, the adsorbed oxygen molecule captures electrons from the conduction bands of MoS₂ nanosheets and generates ionized oxygen radicals, such as ${\mathrm{O}}_2^{-}, {\mathrm{O}}^{-}$, and ${\mathrm{O}}^{2-}$ [51]:

$${\mathrm{O}}_2\mathrm{gas}\to {\mathrm{O}}_2\mathrm{ads}$$

(1)

$${\mathrm{O}}_2\mathrm{ads}+{\mathrm{e}}^{-}\to {\mathrm{O}}_2^{-}\left(100°\mathrm{C}\right)$$

(2)

$${\mathrm{O}}_2^{-}\mathrm{ads}+{\mathrm{e}}^{-}\to 2{\mathrm{O}}^{-}\left(100-300°\mathrm{C}\right)$$

(3)

$${\mathrm{O}}^{-}\mathrm{ads}+{\mathrm{e}}^{-}\to {\mathrm{O}}^{2-}(>300°\mathrm{C})$$

(4)

The sensing mechanism of MoS₂ nanosheet to CH₂O, C₃H₆O, C₂H₆O, and 98%RH have been well-studied and described elsewhere [52,53,54,55]. According to these references, MoS₂-nanosheets-based gas sensors exhibit n-type characteristics in our work. The possible sensing mechanism is as follows: The transfer of electrons from the conduction band to chemisorbed oxygen decreases the carrier density and increases the depletion layer, thereby increasing the resistance of the MoS₂ nanosheets. At room temperature, when the MoS₂-nanosheet-based sensor is exposed to the target gas, for example, CH₂O, the gas is adsorbed on the surface of the MoS₂ nanosheets. These chemisorbed molecules react with ${\mathrm{O}}_2^{-}$ (ads) to form H₂O and CO₂. Therefore, the trapped electrons are released back into the MoS₂ nanosheets, which increases the number of conductive channels, leading to a decrease in sensor resistance (Figure 8).

4. Conclusions

MoS₂ nanosheets were prepared with two grinding solvents via grinding-assisted liquid-phase exfoliation. The effects of grinding solvents on the structure of MoS₂ nanosheets as well as the gas-sensing performance were studied. The structural and gas-sensing properties of MoS₂ were investigated using XRD, SEM, and Raman spectroscopy. The sensing performance of MoS₂ toward four target gases, including CH₂O, C₃H₆O, C₂H₆O, and 98% RH, was analyzed at room temperature. The experimental results proved that the MoS₂ nanosheets exfoliated with NMP responded better than the MoS₂ nanosheets exfoliated with ACN although NMP was not removed completely. The MoS₂ nanosheet-based sensor also exhibited excellent response. However, the sensitivity and LoD of the sensor need to be improved. Accordingly, although NMP cannot be removed completely from the surface of MoS₂, NMP exhibits good gas sensitivity compared with other materials. Therefore, NMP is preferred for the preparation of gas-sensitive materials in grinding-assisted liquid-phase exfoliation. The results provide an experimental basis for the preparation of two-dimensional materials and their application in gas sensors.