Highly sensitive two-dimensional MoS2 gas sensor decorated with Pt nanoparticles

A two-dimensional molybdenum disulfide (MoS₂)-based gas sensor was decorated with Pt nanoparticles (NPs) for high sensitivity and low limit of detection (LOD) for specific gases (NH₃ and H₂S). The two-dimensional MoS₂film was grown at 400°C using metal organic gas vapour deposition. To fabricate the MoS₂ gas sensor, an interdigitated Au/Ti electrode was deposited using the electron beam (e-beam) evaporation method with a stencil mask. The MoS₂ gas sensor without metal decoration sensitively detects NH₃ and H₂S gas down to 2.5 and 30 ppm, respectively, at room temperature (RT). However, for improved detection of NH₃ and H₂S gas, we investigated the functionalization strategy using metal decoration. Pt NP decoration modulated the electronic properties of MoS₂, significantly improving the sensitivity of NH₃ and H₂S gas by 5.58× and 4.25×, respectively, compared with the undecorated MoS₂ gas sensor under concentrations of 70 ppm. Furthermore, the Pt NP-decorated MoS₂ sensor had lower LODs for NH₃ and H₂S gas of 130 ppb and 5 ppm, respectively, at RT.

1. Introduction

Metal oxide-based gas sensors possess many merits, such as relatively high sensitivity and low cost, which has garnered much attention to the material [1]. However, to obtain such gas-sensing properties for a specific gas, these gas sensors generally suffer from several issues, such as thermal instability and high power consumption, due to poor performance at low operating temperatures [1]. As gas detection relies strongly on the properties of material, new materials that possess excellent characteristics for adsorption of gas molecules at low operating temperatures are in great demand as a way to overcome these deficiencies.

Among the new sensing materials, the two-dimensional materials [2–22] with novel and unique electronic, optical and mechanical properties have attracted much attention. In addition to their excellent properties, the novel two-dimensional gas-sensing materials possess structural advantages, such as a high surface-to-volume ratio and the tunable functionality of the surface for decoration species or functional groups, so they have attracted immense interest as new gas-sensing materials [2–4] that can sensitively react upon exposure to a lower level of a specific gas at low temperature. Even though graphene [5–7]—with its outstanding characteristics—is the most widely known two-dimensional material, transition metal dichalcogenides (TMDCs) [8–13,15–22] have great potential to serve as new materials for advanced electronic devices owing to their excellent characteristics, such as high electron mobility, high flexibility, high elasticity and low power consumption. The charge transfer mechanism between gas molecules and molybdenum disulfide (MoS₂)—which is a typically known two-dimensional TMDC—was previously reported for gas detection [15]. In addition, to improve the performance of the gas sensor, various studies on functionalization using metal [23–30] or metal oxide [31,32] and structural modification [17,33] have been reported. Especially, functionalization [23–32] using metal decoration [23–30] on two-dimensional materials can open an avenue for gas detection [23–32].

In this work, we investigated the performance of a two-dimensional MoS₂-based gas sensor and the effects of functionalization using Pt nanoparticle (NP) decoration. Metal organic chemical vapour deposition (MOCVD), which can control the number of two-dimensional material layers and synthesize scalable two-dimensional-layered materials with high quality, was employed to grow the two-dimensional MoS₂ film. Moreover, MOCVD is capable of synthesizing high-quality two-dimensional MoS₂ directly on the sensor substrate to easily fabricate the MoS₂ gas sensor. The MoS₂ sensor consisted of the MOCVD-grown MoS₂film as the active channel and an interdigitated Au/Ti electrode fabricated using the electron beam (e-beam) evaporation method with a stencil mask. At room temperature (RT), the MoS₂ sensor could sensitively detect down to 2.5 and 30 ppm for NH₃ and H₂S gas, respectively. However, the detection of NH₃ and H₂S gas still requires higher sensitivity and a lower limit of detection (LOD) owing to the inhalation toxicology of both gases.

Because both gases exhibit inhalation toxicology that leads to lung and organ injury or even death as well as skin irritation, several international standards have been developed as guidelines to assist in the control of these health hazards. For example, the permissible exposure limits, i.e. regulatory limits, in the work place stipulated by the Occupational Safety and Health Administration (OSHA) are 50 ppm for NH₃ and 10 ppm for H₂S. However, in many cases, the recommended limits of NH₃ and H₂S gas, which are characterized by an extremely irritating odour, lie below the regulatory limit stipulated by OSHA, so studies for high sensitivity and low LOD are still required. Thus, we investigated a functionalization strategy, i.e. metal decoration, to obtain excellent gas-sensing characteristics (e.g. higher sensitivity and lower LOD) for NH₃ and H₂S gas. To improve the sensitivity and LOD for NH₃ and H₂S gas, which act as electron donors on the MoS₂, Pt metal NPs, which resist corrosion and oxidation, yielding stable doping as a noble metal [27], are used as the decorating material. Furthermore, Pt NPs double the p-type doping effect compared with Au NPs [27], which commonly act as a p-dopant [12], so Pt NPs could improve the sensing characteristics for NH₃ and H₂S gas. Even though various studies have already been reported [15–19,34] on the detection of various toxic gases (e.g. NO₂, NH₃, H₂ and CO₂) using two-dimensional TMDC-based sensors, most use flake or bulk two-dimensional TMDC instead of two-dimensional MoS₂ film at RT. In this work, H₂S detection using a two-dimensional MoS₂ sensor in addition to the detection of well-known gases (e.g. NH₃, NO₂, H₂ and volatile organic compounds) is reported. The H₂S response is lower than the NH₃ response on the MoS₂ sensor, confirming that the MoS₂ sensor has lower charge transfer for the detection of H₂S than NH₃ gas. Compared to an undecorated MoS₂ sensor, the Pt NP-decorated MoS₂ sensor detected NH₃ and H₂S gas with high sensitivity down to 600 ppb and 5 ppm, respectively. In addition, NH₃ detection experiments were conducted at concentrations of 600 ppb or less in order to determine the LOD of the Pt-decorated MoS₂ sensor for NH₃ gas.

2. Material and methods

2.1. MOCVD growth of two-dimensional MoS₂ monolayer film with bilayer islands

As previously reported [35], the two-dimensional MoS₂ film was grown using MOCVD. The MoS₂ film was grown by a showerhead-type reactor using Mo(CO)₆ (better than 99.9% purity, Sigma-Aldrich), high-purity H₂ (99.999%), Ar (99.999%) and H₂S (99.9%) with an operating temperature of 400°C. A p-type silicon wafer (1–10 Ω cm) with 300 nm silicon dioxide was used as the substrate for growth. The piranha solution (H₂SO₄ : H₂O₂ = 3 : 1) was used for pretreatment to passivate dangling bonds [35] and make a hydrophilic substrate, resulting in the control of more nucleation sites on the substrate; the piranha treatment was performed by immersing the substrate in the piranha solution for 10 min. After immersion in the piranha solution, the substrate was thoroughly rinsed in deionized water through immersion and sonication, and it was blow-dried with high-purity N₂ (99.999%). After cleaning this, the substrate was placed on a silicon carbide-coated graphite susceptor and immediately loaded into a load-lock chamber, which was connected with the MOCVD process chamber to prevent any contamination. Subsequently, the substrate on the graphite susceptor was loaded into the process chamber to grow the two-dimensional MoS₂ film. MoS₂ film was grown using sublimed Mo(CO)₆ precursor vapour with high-purity H₂S, Ar and H₂ flow for high-quality MoS₂ at 300°C by the reaction between sulfur and gas with the Mo-precursor under a pressure of 5 Torr. Consequently, the fully covered monolayer MoS₂ film was synthesized with bilayer islands without empty spaces of MoS₂.

2.2. Fabrication of MoS₂-based devices and decoration with metal NPs

To prevent any contamination by chemical substances during the fabrication of the MoS₂ gas sensor, we chose a solution-free process, as shown in figure 1a. The MoS₂ gas sensor, which consisted of the MOCVD two-dimensional MoS₂active channel and the interdigitated electrode, was fabricated using the e-beam evaporation method with a stencil mask. To use the MoS₂ as the active material of the gas sensor, half of the MoS₂ film was removed by the scotch-tape method. The interdigitated electrode was patterned on the remaining MoS₂ film using the e-beam evaporation method and a stencil mask with the interdigitated pattern. Additionally, the electrode pad for resistance measurement was patterned on Si with SiO₂ where the MoS₂ was removed. The interdigitated electrode region was made from 5 nm Ti and 50 nm Au using an e-beam evaporation method under a pressure of approximately 10⁻⁷ Torr. Figure 1a also shows an optical image of the active channel (MoS₂ film, 70 µm in width) and the interdigitated Au/Ti electrode of the MoS₂ gas sensor. The active channel material detects gas molecules, and the interdigitated electrode collects charge. In this work, two MoS₂-based devices (one decorated with Pt NPs and one without Pt NPs) were made to investigate the effect of Pt NP decoration. The e-beam evaporation method was used to randomly distribute Pt NPs on the MoS₂. The Pt NPs (approx. 1 nm in thickness) were monitored with a crystal thickness sensor of e-beam evaporation under a pressure of approximately 10⁻⁷ Torr.

2.3. Gas-sensing experiments

A gas-sensing system that allows the users to customize experimental conditions according to their need was set up, as shown in figure 1b. The system consisted of analyte gas (NH₃ and H₂S), dilution (or purging) gas (N₂), a pneumatic valve for opening/closing gas flow, a mass flow controller for regulating analyte gas and dilution gas, a sensing chamber for the gas detection process and a rotary pump for purging the atmosphere after gas-sensing measurements. The gas-sensing experiments were conducted under analyte gas (NH₃ or H₂S) diluted with N₂ in a sensing chamber, and the concentration of analyte gas was controlled by modulating the flow rate of analyte gas and N₂dilution gas. The fabricated MoS₂ gas sensor was placed in the sensing chamber for gas detection. After loading one of the MoS₂ gas sensors (i.e. bare MoS₂ or Pt-decorated MoS₂), the gas-sensing system was sufficiently stabilized under N₂ gas in the chamber and then the gas-sensing experiments were repeatedly performed by automatically controlling the purging step and analysis step for 10 and 5 min, respectively. The experiments were done with analyte concentrations of 600 ppb and 1, 2.5, 5, 15, 30 and 70 ppm. To measure the LOD of NH₃ gas for the Pt-decorated MoS₂ sensor, the experiments were done with concentrations of 130 ppb. The resistance value of the MoS₂ sensor was measured with a Keysight B2985A high-resistance meter. Because the MoS₂ gas sensor has excellent gas-sensing characteristics at high operating temperatures and RT, all experiments were performed at RT under ambient pressure.

3. Results and discussion

To examine the morphology and grain size of the as-grown MoS₂, the MoS₂ film was characterized using scanning electron microscopy (SEM) (Hitachi S-4800), as shown in figure 2a–d: figure 2a shows smaller monolayer MoS₂ clusters (growth time: 1 h), figure 2b shows larger monolayer MoS₂ clusters with a little bit of space (growth time: 2 h) and figure 2c shows monolayer MoS₂ with bilayer islands (growth time: 3 h), which were obtained by controlling coverage using the growth time without a change of other process conditions. As the growth time increased, the coalescence of these monolayer islands led to a fully covered MoS₂ film. Under a highly strong sulfiding condition, the layer atoms are more strongly attracted to the substrate than to themselves, thereby accelerating two-dimensional growth. [35] Many small nucleation sites of MoS₂were made on the SiO₂/Si substrate by pretreatment with the piranha solution, as shown in figure 2a,b. Triangular monolayer MoS₂ clusters with S-edge termination are more stable, and figure 2a,b shows the initially grown triangular MoS₂ monolayer islands on the SiO₂/Si substrate. Before the monolayer MoS₂film had fully formed, bilayer islands began to grow on the merging monolayer islands of MoS₂ on the substrate. After sufficient growth of the MoS₂, the fully formed monolayer MoS₂ film with bilayer islands was deposited after a growth time of 4 h 30 min, as shown in figure 2d. Figure 2d shows that a monolayer two-dimensional MoS₂ film with bilayer islands was grown using the MOCVD method. Thus, MOCVD two-dimensional MoS₂ was successfully grown at 300°C by the reaction between the sulfur and gas reaction with the Mo-precursor under the sulfiding conditions. The grain size of the bilayer islands on the fully formed monolayer film was about 100 nm. Because the adsorption of gas molecules can be attributed to the influence of the many activation sites caused by the presence of the many edges site of the bilayer islands, the fully formed monolayer MoS₂ film with bilayer islands was used as the active material of the gas sensor in this work.

To examine the number of the MoS₂ layers, photoluminescence and Raman spectroscopy (Renishaw inVia Raman microscope) were used. A laser with an excitation wavelength of 488 nm was used. The spectra were also normalized at the Si peak (520.7 cm^–1). As a result, the photoluminescence spectrum of the MoS₂ has an adsorption peak at 659 nm (1.88 eV). The Raman spectrum represents the in-plane vibration (E¹_{2 g}) and the out-of-plane vibration (A_{1 g}) from the MoS₂. The peak position difference between E¹_{2 g} and A_{1 g} was 21.81 cm^–1, which indicates that the MoS₂ has two layers based on the laser wavelength.

To investigate the effects of Pt NP decoration, we made two sensors: an undecorated MoS₂ sensor and a Pt NP-decorated MoS₂ sensor. Henceforth, the undecorated MoS₂ and Pt NP-decorated MoS₂ sensors will be denoted as bare MoS₂ and Pt : MoS₂, respectively. Before the gas-sensing experiments, a transmission electron microscope (TEM) (FEI Tecnai G2 F30 S-Twin) was used to characterize the morphology and structure of the as-obtained Pt : MoS₂. The decorated Pt NPs on MoS₂ were confirmed by TEM images. As shown in figure 3, a number of Pt NPs indicated by yellow circles and the TEM image clearly show small NPs with a diameter around 5 nm. The TEM image of the Pt : NP-decorated MoS₂ reveals the monolayer MoS₂ film (brightest), the triangular bilayer islands (brighter) and the Pt NPs (darker). Moreover, since Pt NPs were physically decorated by the e-beam evaporation method, the TEM image showed that the Pt NPs were randomly distributed on the MoS₂.

The transient NH₃ or H₂S gas-sensing characteristics of the two MoS₂ gas sensors are shown in figure 4. With incrementally increasing concentrations (600 ppb and 1, 2.5, 5, 15, 30 and 70 ppm) of analyte gas, the gas responses of two sensors were investigated under NH₃ and H₂S gas. The sensitivity to the analyte gas was calculated using ΔR/R_p = (R_a − R_p)/R_p, where R_p and R_arepresent the resistances of the device under the N₂ gas (purging gas) and the analyte gas, respectively. When exposed to the NH₃ or H₂S gas, the MoS₂sensors showed decreasing resistance (negative sensitivity). This means that both NH₃ and H₂S gas act as electron donors, transferring electrons to the conduction band of MoS₂ [15]. Thus, this leads to increased electron concentration and conductivity, and the resulting n-doping brings the Fermi level closer to the conduction band edge. Figure 4a shows that the LOD of NH₃ gas was as low as 2.5 ppm and the absolute sensitivity of NH₃ gas ranged from 2.5 to 70 ppm on the bare MoS₂ gas sensor at RT. Moreover, the LOD of H₂S gas was as low as 30 ppm, and the sensitivity of H₂S gas ranged from 30 to 70 ppm at RT, as shown in figure 4b. The two-dimensional MoS₂ gas sensor is the first film sensor to exhibit H₂S gas-sensing characteristics (figure 4b). We concluded that detection of H₂S gas can be attributed to the influence of many activation sites caused by the presence of bilayer islands on the MoS₂ gas sensor. As the concentration of analyte gas (NH₃ or H₂S) increased, the absolute sensitivity of the MoS₂ sensor increased at RT (figure 4c,d). Because the bare MoS₂ gas sensor has higher sensitivity and lower LOD for NH₃ than H₂S gas sensor, it experiences a smaller charge transfer from H₂S than from NH₃. To confirm the LOD of NH₃ gas for the Pt : MoS₂ gas sensor, we performed an additional gas-sensing experiment under analyte concentrations of 600 ppb or less. This experiment demonstrates that the Pt : MoS₂ gas sensor has an LOD of NH₃ gas down to 130 ppb, as shown in figure 5. Table 1 summarizes the gas-sensing characteristics of many TMDC-based gas sensors for NH₃ or H₂S gas at RT. It shows that the Pt : MoS₂ gas sensor has high sensitivity and low LOD for both NH₃ and H₂S gas. Figure 6 shows a schematic of the energy band diagram of MoS₂ and Pt, where the work function (Φ_m) of Pt is approximately 5.65 eV [36] and the electron affinity (χ) of MoS₂ is approximately 4.2 eV [37]. Metal decoration is generally known to enhance sensitivity to a target gas, and Pt NPs act as a p-dopant, increasing sensitivity to gases that act as electron donors on the MoS₂ sensor. Overall, the Pt : MoS₂ sensor reduced the LOD of NH₃ and H₂S gas to 130 ppb and 5 ppm, respectively, which are far lower than those of the bare MoS₂ sensor (2.5 and 30 ppm, respectively). Specifically, the Pt : MoS₂ gas sensor exhibited 5.58× improved NH₃ sensitivity under a concentration of 70 ppm, and the H₂S sensitivity improved by 4.25× at a concentration of 70 ppm. Even though all gas-sensing experiments were performed at RT, the bare MoS₂and Pt : MoS₂ gas sensors exhibited excellent sensing characteristics for the target gases (NH₃ and H₂S).

4. Conclusion

In summary, Pt NPs were decorated on MOCVD two-dimensional MoS₂ to obtain high sensitivity and low LOD for NH₃ and H₂S gases. The electronic properties of MoS₂ were tuned through decoration with Pt NPs, improving the sensitivity and lowering the LOD of the gas sensor for the target gases. The Pt NPs act as a p-dopant, depleting the electron carriers of the MoS₂ film, thereby improving the sensitivity to NH₃ and H₂S gas by 5.58× and 4.25×, respectively, compared to the bare MoS₂ gas sensors under concentrations of 70 ppm. The Pt : MoS₂ gas sensor exhibited LODs of 130 ppb and 5 ppm for NH₃ and H₂S gas, respectively, which are far lower than those for the bare MoS₂ gas sensor (2.5 and 30 ppm, respectively). Even though all gas-sensing experiments were carried out at RT, the MoS₂-based gas sensors showed excellent gas-sensing characteristics. Therefore, functionalization using metal decoration on a two-dimensional gas sensor can improve the sensitivity and LOD for a specific gas by modulating the electronic properties of the two-dimensional material. This strategy opens an avenue for effective toxic gas detection with two-dimensional gas sensors.