Near Room Temperature Light-Activated WS₂-Decorated rGO as NO₂ Gas Sensor

Abstract

The NO₂ response in the range of 200 ppb to 1 ppm of a chemoresistive WS₂-decorated rGO sensor has been investigated at operating temperatures of 25 °C and 50 °C in dry and humid air (40% RH) under dark and Purple Blue (PB) light conditions (λ = 430 nm). Few-layers WS₂, exfoliated by ball milling and sonication technique, with average dimensions of 200 nm, have been mixed with rGO flakes (average dimension 700 nm) to yield WS₂-decorated rGO, deposited on Si₃N₄ substrates, provided with platinum (30 μm gap distance) finger-type electrodes. TEM analysis showed the formation of homogeneous and well-dispersed WS₂ flakes distributed over a thin, continuous and uniform underlying layer of interconnected rGO flakes. XPS and STEM revealed a partial oxidation of WS₂ flakes leading to the formation of 18% amorphous WO₃ over the WS₂ flakes. PB-light irradiation and mild heating of the sensor at 50 °C substantially enhanced the baseline recovery yielding improved adsorption/desorption rates, with detection limit of 400 ppb NO₂ and reproducible gas responses. Cross sensitivity tests with humid air interfering vapor highlighted a negligible influence of water vapor on the NO₂ response. A charge carrier mechanism between WS₂ and rGO is proposed and discussed to explain the overall NO₂ and H₂O response of the WS₂–rGO hybrids.

1. Introduction

The intrinsic merits of Transition Metal Dichalcogenides (TMDs), including their high surface-to-volume ratio and semiconducting properties, have accelerated the development of a diverse range of applications of these materials as chemical sensors [1,2]. Two-dimensional (2D) mono- or few-layer TMDs produced by different exfoliation procedures, expose both plate and edge atoms of a single layer capable of adsorbing gas molecules, providing the largest sensing area per unit volume. Among a large variety of gaseous species, NO₂, H₂ and NH₃ are the most investigated gases, considering their high chemical reactivity with mono- or few-layer MoS₂ [3,4,5], WS₂ [6,7], MoSe₂ [8] and MoTe₂ [9].

In need to decrease the operating temperature of the sensing materials, a common drawback of TMDs operating at room temperature, is their slow response and recovery times, or even no recovery when used at low temperatures [3,5,10]. The increase of the operating temperature up to 150 °C greatly improves adsorption/desorption rates and baseline recovery, but causes partial oxidation of TMDs into their metal oxide counterparts, as previously demonstrated for MoS₂ and WS₂-layered materials [5,7,11]. It turns out that both low and high temperatures hinder the practical use of TMDs due to kinetic (i.e., slow recovery rates) and thermodynamical (i.e., spontaneous oxidation) reasons, respectively. A possible strategy to avoid irreversible adsorption and ageing phenomena, and thus, to enhance the long term stability of the sensors, is to operate them at low temperature utilizing light irradiation as an external source of energy, as previously reported for metal oxide sensors [12,13], rGO-Metal Oxide nanocomposite [14], graphene [15], MoS₂ [3] and WS₂ [16].

Considering the variety of different preparation techniques for producing mono- or few-flake TMDs, consistent with standard electronic processes, some other aspects must be considered. A first issue is the reproducibility of the exfoliation procedure with respect to both microstructure (i.e., number of layers, lateral size, surface area, etc.) and chemical composition (i.e., defects concentration and surface oxidation). A second aspect is the small allowable average lateral size of the TMDs flakes, which may range from 100 to 300 nm, depending on the exfoliation technique [17]. A final issue is the surface coverage of the substrate, needed to create percolation paths for charge carriers between TMDs flakes, bridging metal electrodes. The possibility of using hybrid nanostructures, mostly focused on MoS₂/graphene, making use of large-size, conducting-flakes-pathways of Graphene Oxide (GO) or reduced Graphene Oxide (rGO) with dispersed smaller-size, lesser-conductive-flakes of MoS₂ has been recently reported as an effective solution to enhance the fabrication of this new class of gas sensors [18,19,20]. Beside some reports on the utilization of WS₂/GO hybrids as electrocatalysts for hydrogen evolution reactions [21,22] and WS₂/GO as humidity [23] and NH₃ [24] sensors, no applications of WS₂–rGO nanocomposite for NO₂ gas sensing applications have been reported so far.

In this paper, we report the exfoliation of WS₂ powders by a combined ball milling and sonication technique, which leads to average dimensions of 200 nm and “aspect ratios”, i.e., lateral dimension to the thickness of 27, to produce mono to few-layer WS₂ with controlled morphology and chemical composition. WS₂ flakes are mixed with rGO flakes with average dimensions of 700 nm to yield WS₂-decorated rGO as chemoresistive NO₂ thin films deposited on large-area Si₃N₄ substrates, provided with platinum finger-type patterned electrodes.

The aim of this paper is firstly to demonstrate the reliability of the decoration process leading to the deposition of thin films of well dispersed WS₂ flakes over large-size, interconnected rGO flakes, secondly, to demonstrate and discuss the influence of purple blue light (λ = 430 nm) to detect NO₂ gas in air in the operating temperature range of 25 °C to 50 °C, and lastly, to investigate the influence of water vapor on the NO₂ gas response.

2. Materials and Methods

Materials Preparation. GO was prepared via a modified Hummers method [25] starting from graphite flakes of 500 μm maximum size. Monolayers with average sizes of several micrometres and thicknesses of less than 2 nm were obtained and dispersed in water to reach a final concentration of 0.05 mg/mL. Two-dimensional WS₂ flakes, with average lateral size of 200 nm and average thickness under 15 nm, were obtained by starting from WS₂ commercial powder (Sigma-Aldrich 243639-50G) exfoliated by ball milling assisted sonication, and dispersed in ethanol and subsequently centrifuged as described in the Supporting Figure S1. Finally, equal amounts of the two GO/water and WS₂/ethanol solutions were mixed together and sonicated for 10 min to homogenize the dispersion and avoid agglomeration.

Microstructural and chemical characterization. TEM and STEM pictures of the WS₂-decorated GO were acquired using a TEM—JEOL 2100 Field Emission Transmission Electron Microscope operating at 200 kV by drop casting the dispersion on a lacey grid. Samples prepared by drop casting the WS₂-GO dispersion on Si₃N₄ substrates were further analyzed by X-Ray Photoemission Spectroscopy (XPS) using a PHI 1257 spectrometer equipped with a monochromatic Al Kα source (hν = 1486.6 eV) with a pass energy of 11.75 eV (93.9 eV survey), corresponding to an overall experimental resolution of 0.25 eV. Thin-film XRD measurements were performed at 3° incidence using a Philips PW1710 diffractometer equipped with grazing-incidence X-ray optics, using Cu Kα Ni-filtered radiation at 30 kV and 40 mA.

Sensor fabrication and gas sensing measurements: Thin layers have been prepared by drop casting 8 μL of the water/ethanol dispersed WS₂-GO solution on Si₃N₄ substrates provided with 30 μm spaced Pt interdigitated electrodes, followed by annealing at 70 °C for 30 min to partially reduce the GO to rGO in order to fix the film resistivity in the range of 10⁴–10⁵ Ohm. The electrical resistance of the sensors was measured by an automated system. The sensors placed inside a Teflon chamber (500 cm³), provided with Teflon tubings have been exposed to different gas concentrations in the range 200 ppb-10 ppm NO₂ obtained by mixing certified NO₂ mixtures with dry air carrier gas at 500 sccm/min flow rate, by means of an MKS147 multi gas mass controller. Electrical resistance was measured by a volt-amperometric technique (AGILENT 34970A) under dark and Purple Blue (PB) light irradiating conditions (PB λ = 430 nm) and different power densities at 270 μW/cm², 468 μW/cm² and 668 μW/cm². The device temperature has been controlled by heating elements and temperature sensors (thermocouples) integrated on the device backside, guaranteeing that the gas pressure was not affected by the changes in the OT.

In this paper the relative response RR is represented by the ratio RR = Ra/Rg where Ra and Rg are the resistances in dry air and in gas respectively and τ_ads and τ_des (respectively adsorption and desorption times) represent the time required to reach 90% of the full response at equilibrium, during both gas adsorption and desorption.

3. Results and Discussion

3.1. Morphological and Compositional Characterization of the WS₂-Decorated rGO

Thin films of WS₂-decorated rGO deposited on Si₃N₄ substrates provided with Pt finger type electrodes have been characterized. We have firstly deposited by drop deposition the minimum amount of WS₂–rGO solution, corresponding to the formation of a continuous percolation path of GO flakes. The development of a continuous percolation path was assessed by recording the electrical resistance of the film corresponding to the onset of an electrical contact between the electrodes (30 µm apart). By further annealing at 70 °C in air for 30 min, the GO flakes have been partially reduced to rGO to yield baseline resistances in air in the range of 10⁴–10⁵ Ω.

The WS₂-decorated rGO morphology was first characterized by low-resolution TEM, as shown in Figure 1a. WS₂ flakes (darker regions) are distributed over a thin, continuous and uniform layer made of interconnected rGO flakes (light-grey background) as revealed by the presence of grey lines attesting the formation of rGO folded edges (yellow arrows in Figure 1a). Statistical image analysis carried out on differently prepared samples over an area of 80 µm², as shown in Figure 1b, exhibits a log-normal WS₂ average particle size distribution with an average particle size dimension of 200 nm and an average WS₂ flakes coverage percentage of 6%. Notably, the dispersion of the WS₂ flakes is homogeneous over the investigated area, meaning that the sonication step after mixing represents an effective strategy to avoid agglomeration of the WS₂ flakes.

Figure 1.

(a) Low-resolution TEM image of WS₂-decorated GO deposited on a lacey grid. WS₂ flakes (darker regions) distributed over interconnected rGO flakes (light-grey background). The occurrence of grey lines (highlighted by yellow arrows) attest the formation of rGO folded edges; (b) Lateral size distribution of WS₂ flakes and corresponding Log Normal fit.

The TEM analysis shown in Figure 2a reveals the presence of large (i.e., hundreds of nanometers) transparent, irregularly shaped GO flakes with folded edges (yellow arrow) distributing according to a continuous planar underlying layer. From Figure 2a it is shown that WS₂ flakes may also align vertically, as highlighted by the darker needle-shaped formations inside the red circles of Figure 2a. Statistical analysis carried out on the vertically aligned WS₂ flakes revealed an average thickness of 15 nm, corresponding approximately to 25 layers. TEM image depicted in Figure 1b illustrates the formation of well-shaped, stacked-few-layers WS₂, with edge angles of 120°, deposited over the underlying GO flakes. The crystalline nature of few-flakes WS₂ is confirmed by the selected area electron diffraction (SAED) pattern shown in the inset of Figure 2b, which clearly exhibits the formation of WS₂ nanosheet with hexagonal atomic arrangements, assigned to WS₂ (100) plane [26].

Figure 2.

High-resolution TEM of WS₂-decorated rGO showing: (a) light-grey background of interconnected GO flakes. Darker grey lines (yellow arrow) corresponding to folded GO edges and some WS₂ flakes vertically placed on GO flakes (needless inside the red circle); (b) a big hexagonal WS₂ flake. The inset shows a magnification of the flake’s edges with related SAED analysis pattern.

Figure 3 shows the atomic distribution of sulphur (a), tungsten (b), carbon (c) and oxygen (d) elements, as measured by STEM technique, respectively on the sample displayed in Figure 2b. From Figure 3a,b, it turns out that the distribution of tungsten and sulphur exactly replicates the shape of the WS₂ flake. The carbon signal over the WS₂ particles of Figure 3c, may be possibly be attributed to a partial contamination of the sample, whereas the carbon signal deriving from the background, clearly replicates the morphology of the rGO underlying layer. The chromatic signal intensity of oxygen corresponding to the WS₂ flake (Figure 3d), is slightly brighter than the background, indicating the occurrence of a higher oxygen concentration over the WS₂ flake with respect to the rGO background, suggesting the occurrence of an oxidation process of WS₂, as it will be confirmed in the next XPS section.

Figure 3.

STEM elemental mapping of the flake shown in Figure 2b showing the atomic distribution of: (a) sulfur, (b) tungsten, (c) carbon and (d) oxygen.

Figure 4 shows the XPS C1s (a) and W4f (b) core level photoemission spectra of WS₂-decorated rGO air-annealed at 70 °C. According to the literature [27,28], the C1s spectrum shown in Figure 4a has been successfully fitted by the sum of five components assigned to C sp² atoms belonging to aromatic rings and hydrogenated carbon (C=C/C−C 284.7 eV), hydroxyl groups (C−OH, 285.8 eV), epoxy groups (C−O−C, 286.8 eV), carbonyl groups (C=O, 288.0 eV), and carboxyl groups (C=O(OH), 289.2 eV). Upon thermal annealing at 70 °C, the intensity of all the oxygen-containing groups is lowered, with regard to sp² carbon containing ones (i.e., (C=C/C−C), signifying a loss of oxygen in favor of sp² carbon. Compared to previous results [29], the XPS spectrum shown in Figure 4a is located halfway between the XPS signals of as deposited GO and the one corresponding to 200 °C UHV-annealed rGO. Regarding the tungsten W4f of Figure 4b, the four peaks fitting the spectrum can be located, according to the literature [30,31], to WS₂ and WO₃. These results suggest, as previously reported for MoS₂ and WS₂ [5,7,11], that crystalline WS₂ is partially oxidized to WO₃ with an associated content of approximately 18%. Gracing incidence XRD measurements on the WS₂-decorated rGO annealed at 70 °C for 30 min, as shown in the supporting Figure S2, revealed that the WO₃ is amorphous.

Figure 4.

XPS spectra of the WS₂-decorated rGO showing (a) C1s core level spectrum of rGO flakes and (b) W 4f core level spectrum of WS_2.

3.2. NO₂ Gas Response

The reacting surface of the WS₂-decorated rGO comprises, as attested by microstructural characterization, a flat underlying layer made of large and interconnected rGO flakes covered with dispersed, partially oxidized, WS₂ flakes. Given that WS₂ flakes do not form a continuous layer (see Figure 1a), it is the rGO layer which mostly determines the baseline resistance of the WS₂–rGO hybrid. The NO₂ gas responses in dry air of a single rGO film and a WS₂-decorated rGO film at 25 °C and 50 °C operating temperatures (OT) are compared in Figure 5.

Figure 5.

Electrical responses of single rGO (red line) and WS₂-decorated rGO films (blue line) in dry air and NO₂ concentrations in the range 2–10 ppm at 25 °C and 50 °C operating temperature.

By increasing the OT to 50 °C, the baseline resistance (BLR) of the WS₂-decorated rGO (dotted lines in the figure) decreases almost of one decade as compared to rGO, attesting the substantial contribution of the WS₂ semiconductor to the overall resistance response. At 25 °C OT, by increasing the NO₂ gas concentration, it is noted that: (i) the baseline resistance (BLR) is not recovered after gas desorption; (ii) baseline resistance steadily drifts, decreasing its resistance; (iii) the electrical signal never reaches equilibrium under adsorption/desorption conditions within the time schedule of the experiment (i.e., 60 min). Conversely, by increasing the OT to 50 °C, WS₂-decorated rGO shows a faster response, improved equilibrium conditions and reduced baseline drift. Besides the positive effects related to the OT, it may be concluded that irreversible adsorption and baseline drift phenomena are still evident at 50 °C OT, representing a serious drawback for the exploitation of near room temperature WS₂–rGO hybrids sensors.

According to previous research which demonstrated the positive effect of the combined action of light irradiation and thermal activation on the NO₂ response of WO₃ and NiO semiconductor sensors [12,13], Figure 6 shows the rGO’s and WS₂–rGO hybrid’s NO₂ gas responses at 25 °C in dry air under Purple Blue light (PB λ = 430 nm @ 2.88 eV) and different power densities at 270 μW/cm², 468 μW/cm² and 668 μW/cm². Bare rGO exhibits neither significant changes of the electrical response at different power densities (in Figure 6 the response at 668 μW/cm² is shown) nor any appreciable gas response to 1 ppm NO₂. WS₂-decorated rGO yields relative responses (RR = Ra/Rg at 1ppm NO₂) of approximately 1.21, more or less independent of the power density. Response and recovery times, in contrast, improve with increasing the power density.

Figure 6.

Response to 1 ppm NO₂ in dry air of WS₂-decorated rGO and bare rGO film (red curve) irradiated by Purple–Blue (λ = 430 nm) light at different power densities (270 μW/cm², 468 μW/cm², 668 μW/cm²).

By selecting the 668 μW/cm² light power density, Figure 7 shows the electrical response of the WS₂-decorated rGO to 1 ppm NO₂ in dry air under “dark” and “purple–blue” conditions at 25 °C and 50 °C OT respectively.

Figure 7.

The 1 ppm NO₂ gas responses of the WS₂-decorated rGO under “dark” and “light” conditions (dotted lines) at 25 °C operating temperatures compared to the “dark” and light” conditions (solid lines) at 50 °C operating temperature.

If we now define the recovery percentage (RP) as the percentage ratio (Δ_D/Δ_A) × 100, where Δ_D and Δ_A (see Figure 7) are the desorption/adsorption resistances’ variations (Δ), measured at the end of each desorption/adsorption cycle (i.e., 60 min), it turns out that light illumination strongly enhances the recovery of the baseline resistance during desorption.

As shown in Figure 7 and Table 1, recovery percentage (RP) values increase from 23% (dark) to 65% (light) at 25 °C OT, and from 60% (dark) to 70% (light) at 50 °C OT. Moreover, considering features and shapes of the curves displayed in Figure 7, at 25 °C OT regardless of “dark” or “light” exposures, no equilibrium conditions are achieved within the timescale of the experiment (i.e., 60 min). In contrast, as attested by the horizontal slopes of the two bottom curves of Figure 7, equilibrium conditions are achieved under adsorption/desorption conditions when light irradiation is performed and when the OT is increased to 50 °C. Notably, according to Table 1, given the associated uncertainty of the measurement (± 0.02), no substantial changes of the relative responses’ values are recorded with respect to 25 °C, either by increasing the OT to 50 °C or by illuminating the sensor with purple blue light.

Table 1.

Comparison of RR, RP, τads and τdes to to 1 ppm NO₂ in dark conditions and PB light illumination (λ = 430 nm at 668 µW/cm²) at different OT (25–50 °C). Notably (-) means that no equilibrium conditions have been reached within the time scale of the experiment.

Response TO 1 ppm NO2
Operating Conditions	RR Ra/Rg	RP ΔD/ΔA	τ ads	τ des
	(–)	(%)	(min)	(min)
25 °C DARK	1.18 ± 0.02	23	-	-
25 °C LIGHT	1.21 ± 0.02	65	-	-
50 °C DARK	1.20 ± 0.02	60	22	26
50 °C LIGHT	1.27 ± 0.02	70	16	18

The results shown in Table 1 are in line, both in terms of sensitivity, using comparable definitions, and time constants with those reported in literature for hybrid graphene/MoS₂ structures for NO₂ sensing [18,19] working in dark conditions and at higher operating temperatures (at least 150 °C).

Figure 8 compares the normalized dynamic gas responses of the WS₂–rGO sensor at 50 °C OT under dark and light conditions, respectively. By increasing the NO₂ concentration from 200 ppb to 1 ppm, purple blue light irradiation promotes full baseline recovery after each NO₂ pulse, as demonstrated by the horizontal slope of the dotted line of Figure 8. WS₂-decorated rGO exhibits an experimental detection limit of 400 ppb NO₂ in dry air. The inset of Figure 8 shows the sensitivity plot of the response under PB illumination, with associated standard deviations. Reproducibility tests of the electrical response under pulse and cumulative NO₂ adsorption/desorption exposures, shown in the supporting Figure S3, exhibit no substantial irreversible adsorption phenomena, as well as a fairly good reproducibility of the electrical response. Additionally, long-term stability properties of the baseline and saturation resistances to 1 ppm NO₂ over a period of 12 months were also recorded. Supporting Figure S4 shows baseline resistances (upper curve) and saturation resistances corresponding to 1 ppm (lower curve), randomly collected over a period of 52 weeks, with associated standard deviations calculated over a set of five consecutive measurements. No remarkable fluctuations of both baseline and resistances at saturation are detectable, demonstrating good long-term stability properties of the WS₂ films.

Figure 8.

Comparison of WS₂-decorated rGO’s electrical responses to increasing NO₂ concentrations at 50 °C operating temperature under dark and light conditions respectively. The inset shows the sensitivity plot corresponding to light conditions. The bars in the inset represent the standard deviation calculated over a set of five measurements performed for each gas concentration. Dotted lines indicate the baseline resistance in dry air.

The effect of 40% relative humidity (RH) to the NO₂ response, under dark and light conditions, at 50 °C OT, is shown in Figure 9. Regardless of the illumination: (i) the baseline resistance slightly increases from dry to humid (40% RH) conditions; (ii) the NO₂ relative responses (Ra/Rg) at 40% RH, do not change appreciably neither under dark or light conditions; (iii) the baseline resistance is fully recovered after each adsorption/desorption cycle. This behavior demonstrates that humidity, regardless of the illumination conditions, does not appreciably interfere with the NO₂ gas adsorption mechanism, suggesting, as it will be discussed in the next paragraph, that NO₂ preferentially adsorbs on WS₂, since energetically favored as respect to water vapor.

Figure 9.

The influence of 40% Relative Humidity (40% RH) on the NO₂ response of WS₂-decorated rGO under dark and light illumination at 50 °C operating temperature.

Finally, the sensors’ selectivity among other interfering gases such as H₂, NH₃, acetone and ethanol has been recorded and the results shown in Supporting Figure S5. The WS₂-decorated rGO sensor exhibits an excellent selectivity for NO₂ gas as compared to the other investigated species, making it suitable for being a selective NO₂ sensor.

Gas Sensing Mechanism

Discussing the gas sensing mechanism of the WS₂–rGO hybrid sensor with respect to water vapor and NO₂, the single contribution of the rGO and WS₂ species and WS₂–rGO hybrid to the overall gas response have to be considered.

Regarding the single rGO in our previous work [29] we demonstrated that both as-deposited GO and partially reduced rGO (annealed in vacuum at 200 °C) exhibit a p-type response to NO₂, resulting in a decrease of the resistance in the operating temperature range of 25–150 °C. Considering that the rGO prepared in this work shows a degree of reduction, located halfway between these two extremes, we would have expected a decrease of the resistance in Figure 5, rather than no response when exposing bare rGO film to NO₂ gas. The lack of any gas response may not be entirely attributed to the decrease of functional groups induced by the mild reduction process (i.e., air annealed at 70 °C), but mostly to a substantially smaller amount of the deposited rGO, which amounts to approximately 1/10 of the quantity previously utilized [29]. It may be concluded that the deposition procedure adopted here yields an rGO film which does not significantly contribute to the NO₂ response, while it guarantees the formation of a continuous, conductive layer “bridging” distant platinum finger-type electrodes.

Regarding single WS₂, which is reported to increase its resistance to NO₂, exhibiting an n-type response [7,16], some considerations apply. According to the XPS section, WS₂ flakes deposited on rGO partially oxide to amorphous WO₃ (approx 18%). Amorphous WO₃, as previously discussed [7,16], acts as a non-conductive phase, eventually inhibiting the charge–carrier transfer mechanism within the flakes. It turns out that the formation of WO₃, while not contributing to the overall gas response, has the merely negative effect of partially covering the underlying reacting surface of the WS₂ flakes, eventually decreasing the relative gas response.

Regarding crystalline WS₂, the variation of the electrical resistance induced by Air–NO₂ mixtures, is part of a complex mechanism involving the combined effects physisorption, chemisorption, the role of edges and surface defects and the transduction mechanism. First principle calculations on MoS₂ sulphur-vacancy-defective monolayers [32,33] demonstrated that O₂ firstly chemisorbs on existing sulphur vacancies, secondly, that sulphur vacancies are passivated (i.e., “healed”) by the dissociative chemisorption of O₂ molecules, leading to the formation of two Mo-O bonds. In case of direct NO₂ molecules interaction with sulphur vacancies, a dissociative chemisorption of NO₂ takes place, leading to oxygen atoms passivating the vacancies (i.e., “secondary healing”) and NO molecules eventually to be physisorbed on the MoS₂ surface. Considering that both MoS₂ and WS₂, are susceptible to spontaneous oxidation in air [11], the mechanism of sulphur vacancies suppression operated by O₂ and NO₂ claimed for MoS₂, can be reasonably extended to sulphur defective WS₂. Under these circumstances, literature reports based on first principle calculations on “healed” WS₂ surfaces pointed out that O₂, NO₂ and H₂O physisorb on defect-free monolayer WS₂ surfaces [34,35].

Regarding WS₂ –rGO hybrids, they decrease their overall resistance when exposed to NO₂ (see Figure 5), exhibiting, as reported for MoS₂/Graphene composites [18,20], an overall p-type response. Bearing in mind that the WS₂–rGO hybrid’s gas response depends on both charge transfer values (positive charge transfer values between adsorbing molecules and the material surface stand for electrons withdrawal from the material) and adsorption energies (negative adsorption energy means that the adsorption process is exothermic and energetically favourable), the following discussion applies.

According to Table 2, adsorption of a single NO₂ molecule on rGO yields charge transfer values of approximately 0.029e and associated adsorption energies of −0.80 eV [32,34,35]. By contrast, the adsorption of a single NO₂ molecule on defect free WS₂ indicate that NO₂ physisorbs on WS₂ yielding charge transfer values of 0.178e and associated adsorption energies of −0.41 eV [34,35].

Table 2.

Adsorption energy on rGO and WS₂ of O₂, NO₂ and H₂O molecules. Note that negative adsorption energy means that the adsorption process is exothermic and energetically favorable.

Physisorbed Molecules	Adsorption Energy on rGO (eV)	Reference	Adsorption Energy on WS2 (eV)	Reference
O2	−0.16	[36]	−0.24	[36]
NO2	−0.80	[32,34,35]	−0.41	[34,35]
H2O	−0.04	[37]	−0.23	[34]

Considering that calculated NO₂ charge transfer values are 0.029e for rGO and 0.178e for WS₂, it turns out that WS₂ yields a net charge transfer 6.14 times higher than on of rGO per NO₂ physisorbed molecule. Given these premises, the electrical response of WS₂-decorated rGO as compared to single rGO shown in Figure 5, accounts for the larger WS₂′s charge carrier exchange with respect to rGO.

Figure 10, which is a schematic illustration of the WS₂-decorated rGO film, deposited between two platinum electrodes 30 µm apart, highlights that NO₂ molecules adsorbs on both rGO and WS₂ causing electrons to withdraw from both materials. Considering the smaller charge transfer induced by NO₂ adsorption on rGO (i.e., 0.029e) with respect to WS₂ (i.e., 0.178e), by exposing the WS₂–rGO hybrid to NO₂ gas, the NO₂ gas molecule withdraws electrons mainly from the WS₂ surface. As a consequence, electron-depleted n-type WS₂ flakes drain electrons from the underlying p-type rGO, thanks to a rapid electron transport from the highly conducting rGO to the less-conducting WS₂ [19,22]. The increase of hole concentration in p-type rGO due to physisorption of NO₂ molecules on WS₂ flakes, explains the decrease of the overall resistance of the p-type WS₂-decorated rGO shown in Figure 5. It may be concluded that rGO flakes, with excellent transport capability, serve as highly conductive channels bridging distant electrodes, whereas WS₂ decoration eventually modulates the NO₂ gas response.

Figure 10.

Schematic illustration of the proposed sensing mechanism of WS₂-decorated rGO hybrid during NO₂ exposure.

Regarding the influence of purple–blue light, it is well known that response and recovery rates depend on the adsorption energy of the adsorbed gas molecules (see Table 2). Light irradiation and thermal activation represent, therefore, two alternative or complementary modes to improve response/recovery rates [13]. By irradiating WS₂-decorated rGO with purple–blue light (PB λ = 430 nm) the associated photon energy of 2.88 eV and power density of 668 μW/cm² provide: (i) a large quantity of photo generated electrons/holes; (ii) the required energy to desorb physisorbed O₂, NO₂ and H₂O molecules from both rGO and WS₂ flakes. Under NO₂ gas and PB light illumination, physisorbed oxygen partially desorbs from its adsorption site (☐) according to Reaction (1) while NO₂ physisorbs on free sites (☐) left behind from oxygen desorption according to Reaction (2). At equilibrium the overall reaction is represented by Equation (3).

$$\left(O_2^{-}-☐\right)+h^{+}+e^{-}\rightleftharpoons \uparrow O_2+☐+e^{-}$$ (1)

$$N{O}_2+☐+e^{-}\rightleftharpoons \left(N{O}_2^{-}-☐\right)$$ (2)

$$\left(O_2^{-}-☐\right)+h^{+}+e^{-}\rightleftharpoons \left(N{O}_2^{-}-☐\right)+\uparrow O_2$$ (3)

Considering that NO₂ molecules withdraw electrons with an associated charge transfer of 0.178e as compared to that of Oxygen of 0.136e [34], under NO₂ adsorption, Equation (3) is shifted to the right, meaning an excess of holes, which explains the overall resistance decrease of the WS₂–rGO hybrid. Moreover, the improved recovery percentages and response times of Figure 7, account for the extra light-photogenerated carriers which speed up the time to reach equilibrium (Equation (3)) both during gas exposure and recovery.

Regarding water interaction with WS₂–rGO hybrid the reason why humidity, regardless of the illumination conditions, does not appreciably interfere with the NO₂ gas response, can be tentatively explained (see Table 2) evaluating that the adsorption energy of a NO₂ molecule on WS₂ is approximately twice (−0.41 eV) the corresponding energy of water (−0.23 eV), indicating a stronger attitude of NO₂ molecules to adsorb on WS₂ compared to water. Finally, the initial resistance increase of the WS₂–rGO when exposed to humid air can be explained taking into consideration that water is a reducing agent, which adsorbs on both rGO and WS₂ injecting electrons, thus decreasing the hole concentration in rGO and WS₂, causing an overall resistance increase of the WS₂–rGO hybrid.

4. Conclusions

We have exfoliated, by a combined grinding and sonication technique, WS₂ commercial powders into mono-to few-layer flakes of WS₂, with an average dimension of 200 nm, which have been successfully dispersed with rGO flakes with average dimensions of 700 nm, to yield WS₂-decorated rGO as chemo-resistive NO₂ thin film sensor. Operating at near room temperature conditions and providing an extra source of energy, by purple–blue light illumination, we have proposed a possible strategy to improve adsorption/desorption rates and to suppress water vapour cross sensitivity. The deposition procedure adopted here yields WS₂ flakes which mostly drive the NO₂ gas response, while the underlying rGO film guarantees the formation of a continuous, conductive layer “bridging” distant platinum finger-type electrodes. By retrieving literature data about charge carriers and adsorption energies deriving from the interaction of NO₂, O₂ and water molecules with WS₂ and rGO, we have proposed a gas sensing mechanism which accounts for the overall gas response of the WS₂–rGO hybrids.

Supplementary Materials

The following are available online at https://www.mdpi.com/1424-8220/19/11/2617/s1, Figure S1: Schematic illustration of WS2 exfoliation process, Figure S2: Grazing incidence XRD spectrum of WS₂-decorated rGO film, Figure S3: Reproducibility and base line recovery features of the WS₂-decorated rGO by exposing the film to both dynamic and cumulative NO₂ concentrations (5–10 ppm) under light irradiation (purple blue), Figure S4. Long term stability properties of the WS₂/rGO hybrid sensor, Selectivity response of the the WS₂/rGO hybrid sensor to different oxidizing and reducing gases, measured at 50 °C operating temperature under purple-blue illumination.

Author Contributions

Conceptualization, L.O., C.C and V.P.; investigation, V.P. and S.M.E.; data curation, V.P.; Methodology, L.O. and C.C.; Formal analysis, V.P.; writing—original draft preparation, C.C. and V.P.; writing—review and editing, L.O., C.C.

Funding

This research was funded by REGIONE ABRUZZO Dipartimento Sviluppo Economico, Politiche del Lavoro, Istruzione, Ricerca e Università Servizio Ricerca e Innovazione Industriale for financial support throug progetto POR FESR Abruzzo 2018-2020 Azione 1.1.1 e 1.1.4—“Studio di soluzioni innovative di prodotto e di processo basate sull’utilizzo industriale dei materiali avanzati” CUP n. C17H18000100007.

Conflicts of Interest

The authors declare no conflict of interest.