A Room Temperature High‐Performance Visible‐Light‐Assisted NO2 Gas Sensor Based on Ultrathin Zinc Oxysulfide

Metal oxysulfides are an emerging group of sensitive materials for high‐performance NO2 sensing owing to their room‐temperature operation capacity, excellent selectivity of NO2, the part‐per‐billion‐leveled limit of detection, as well as high stability against the ambient environment. Here, the room‐temperature NO2 sensing performances of zinc oxysulfide 3D micro‐self‐assembly composed of ultrathin nanoflakes are investigated. The combined hydrothermal–annealing approach is applied to first synthesize zinc sulfide micro‐self‐assembly and then transform it into zinc oxysulfide in a controlled environment. As a result, the majority of S atoms in zinc sulfide are replaced with O atoms, leading to the crystal structure variation from cubic/hexagonal to an orthorhombic configuration. Simultaneously, the corresponding optical bandgap is reduced from ≈3.6 – 3.8 eV to ≈1.92 eV, enabling the visible light harvesting capability. The sensor demonstrates a fully reversible and repeatable sensing response toward 1.26 ppm NO2 gas at room temperature with a response magnitude of ≈2.27 under the 460 nm excitation, a limit of detection (LOD) of 294.8 part‐per‐trillion (ppt), and almost an order of magnitude larger compared to other commonly used gas species. This work demonstrates the great potential of the metal oxysulfide framework for developing next‐generation room‐temperature NO2 gas sensors.


Introduction
Nitrogen dioxide (NO 2 ) is a hazardous gas within the category of air pollutants, posing adverse effects on the environment and DOI: 10.1002/adsr.202300007human health even at the part-permillion (ppm) level. [1,2]To achieve effective monitoring of NO 2 as a trace gas, various sensing technologies, including catalysts-based electrochemical, [3] semiconductors-based chemiresistive, [4] absorption-based optical, [5] and luminescence-based optical, [6] have been developed and implemented in applications of environmental remediation, air pollution monitoring, automobile, as well as disease diagnostics.Amongst these, chemiresistive NO 2 sensors based on semiconducting metal oxides present unique advantages such as high sensitivity, fast response/recovery, as well as low cost, while relatively low selectivity and elevated operating temperatures are the two major drawbacks. [7]Despite that SnO 2 has been the core sensitive material for most commercially available products, the exploration toward other nanostructured metal oxides such as MoO 3 , [8] WO 3 , [9] In 2 O 3 , [10] Fe 2 O 3 , [11] and Bi 2 O 3 [12] has demonstrated certain improvements in the sensing performances by lowing operating temperatures and enhancing the selectivity of NO 2 .
[15] Such peculiar features originate from the unique NO 2 gas physisorption behaviors of 2D metal sulfides, which are different from most of the nanostructured metal oxides that rely on the catalytic reaction of NO 2 with surface adsorbed oxygen at elevated temperatures (i.e., so-called "chemisorption"). [16,17]Physisorption of paramagnetic NO 2 gas molecules over the material surface typically occurs at room temperature or relatively low elevated temperatures, leading to the formation of electric dipoles at the gas-matter interface. [13,18]][21] A wide range of sulfides of metal elements which the oxides also share, including WS 2 , [22] MoS 2 , [23] SnS 2 /SnS, [24] In 2 S 3 , [25] and Ga 2 S 3 /GaS, [26] have been investigated, in which some of them demonstrate fully reversible NO 2 response at room temperature with and without external stimulation (e.g., light and gating).Nevertheless, the surface of most metal sulfides is sensitive to oxygen from the ambient air, indicating that the sensor performances are degrading overtime during the practical implementation. [27]etal oxysulfides, on the other hand, can potentially harness both the advantages of metal oxides and sulfides on NO 2 sensing. [28]As the intermediates during the transformation from oxides to sulfides or vice versa, nanostructured metal oxysulfides are generally realized through the mild oxidation of metal sulfide crystals in either the liquid [19] or gas phase. [29]Improved NO 2 sensing performances with part-per-billions (ppb) or sub-ppb limit of detection are found in oxysulfides of In, [19] Ag, [30] and Pd, [31] while maintaining the unique selectivity, roomtemperature operation potential, as well as excellent long-term stability all at the same time.
In this work, we investigate the room-temperature NO 2 sensing performances of zinc oxysulfide with the appearance of 3D self-assembly made from ultrathin nanoflakes.The utilization of 3D self-assembly has been proved to favor gas sensing performances over those of randomly stacked low-dimensional nanostructures, owing to its relatively ordered pores and enhanced surface area. [18,20,29]With the excitation of visible light, the electrical resistance of zinc oxysulfide self-assembly deposited over the interdigitated electrode (IDE) substrates is measured in a realtime manner upon the exposure of NO 2 with different concentrations at room temperature.Subsequently, the critical sensing performance parameters of response magnitude, response/recovery time, the limit of detection, and selectivity are extracted.

Material Characterization
The synthesized zinc oxysulfide exhibits amalgamated spherical structures with sizes ranging between 5 and 50 m as shown in a scanning electron microscope (SEM) image (Figure 1a).According to the low-resolution transmission electron microscopy (TEM) image in Figure 1b, the 3D spherical structures are found to be assembled from clusters of round-shaped zinc oxysulfide nanoflakes with lateral dimensions ≈50 nm and ≈8 nm thickness obtained from Atomic Force Microscopy (AFM) measurement (Figure S1, Supporting Information).Furthermore, the nanoflakes are of an ultrathin flake-like morphology, depending on their highly transparent features (Figure 1b), which may originate from the high-temperature calcination treatment during the synthesis process (details are presented in the Material Synthesis and Preparation section). [18,29]From energy-dispersive X-ray spectroscopy (EDS) measurements in Figure S2 (Supporting Information), the original zinc sulfide powder is found to be composed of Zn and S atoms only.Given an annealing treatment at 600 °C with a constant airflow supplied, the S atoms of the pure zinc sulfide are largely replaced with O atoms in company with a substantial crystal reconstruction, forming the zinc oxysulfide nanoflakes with the elemental composition of Zn, S, and O as shown in EDS results in Figure 1c.The crystal structures of the materials are investigated using X-ray diffraction (XRD).From Figure 1d, the diffraction patterns of the zinc oxysulfide nanoflakes demonstrate a vast difference compared to that of the pure zinc sulfide, implying a significant structural transition occurred during the annealing process.The diffraction peaks observed at 32.2°, 34.8°, 36.7°,[34][35] Such a lattice system is further confirmed using the selected area diffraction (SAED) measurement and the corresponding high-resolution TEM (HRTEM) image as shown in Figure 1e,f, in which the crystal planes of (002), (103), and (521) are identified with the corresponding d-spacing of 0.24, 0.16 and 0.15 nm, respectively.
The detailed chemical composition of the zinc oxysulfide is scrutinized using X-ray photoelectron spectroscopy (XPS).From the Zn 2p spectrum in Figure 2a, the Zn 2p 3/2 together with its spin-orbital coupler are found at binding energies (BE) of ≈1021.90 and ≈1044.90[38][39][40][41][42] Given the broad peak observed in S 2p spectrum in Figure 2b, the S 2p 3/2 and S 2p 1/2 doublet are deconvoluted at the BE of ≈167.95 and ≈166.79 eV, respectively, indicating the formation of S─O bonds as suggested in other metal oxysulfide reports. [18,19,29]][45][46][47] Figure 2c depicts the BE of O 1s spectrum, in which the dominated peak centered at ≈532.04 eV is attributed to the chemical bond between Zn atoms and O atoms, whereas the minor peak located at the higher BE of ≈533.86 eV is considered to be a trace of the absorbed oxygen species upon the material surface. [48,49]he bonding characteristics of the zinc oxysulfide are investigated using the Raman spectrum as shown in Figure 3a, in which the strong peaks found at the Raman shift of ≈330, ≈437, and ≈569 cm −1 can be ascribed to the Zn─O vibrational mode of 2E 2 , E 2 and E 1 (LO), respectively. [50,51]Besides, the weak peak located at 204 cm −1 is assigned to the Zn─S bond vibration due to the first-order zone-boundary phonons, which further proves the coexistence of the Zn─S bond and Zn─O bond inside the material.From the UV-Vis-NIR measurements in Figure 3b, the zinc oxysulfide exhibits a broad absorption peak between UV and NIR regions, revealing its excellent optical absorption behavior in the visible light spectrum, particularly in the blue light region (Figure 3b).[54] From the photoluminescence (PL) spectrum in Figure 3c, a sharp PL peak is found to center at ≈521 nm, which is also close to the estimation of the UV-Vis-NIR measurement.To further understand its optical properties, the time-resolved photoluminescence (TRPL) spectrum is measured using a timecorrected single-photon counting (TCSPC) technique as shown in Figure 3d. [55,56]The intrinsic excitonic radiative lifetime can be evaluated by fitting the measured data with a tri-exponential decay model (fitting parameters are shown in Note S1, Supporting Information). [14,19,21,57]As a result, the average excitonic radiative lifetime of the ultrathin zinc oxysulfide is calculated as 846.16 ns, which is dramatically prolonged compared to those of other metal oxysulfides, [14,19,21,58] suggesting that the synthesized zinc oxysulfide can be an ideal sensing material to incorporate into a visible-light-assisted gas sensor.

Reversible Room Temperature NO 2 Gas Sensor
To thoroughly explore its potential in gas sensing, a zinc oxysulfide sensor is prepared by depositing ≈40 L of material solution upon a silicon-based interdigital transducer (IDT) substrate (details are presented in the Experimental Section).An air-tight gas chamber is customized to contain the prepared sensor substrate, in which a set of programmable mass flow controllers (MFCs) is connected to the inlet to provide a tunable gas flow from the target gas cylinders to the chamber.The exhaust outlet of the chamber is vented in the fume hood through an ¼ inch Teflon tube.Moreover, a light source is set directly upon the sensor substrate as a light-emitting diode (LED) with changeable wavelengths of 460, 520, and 620 nm, providing a visible light excitation during the experiment.The electrical resistance of the sensor is constantly recorded using a high-resolution multimeter via the builtin probe stages inside the chamber.The gas sensing performance is assessed as a response factor based on Equation (1), where R g denotes the sensor's electrical resistance with the exposure of the analyte gas species, whereas R a represents its base resistance in the air. [13,59]sponse Factor = R g ∕R a (1) NO 2 is chosen as the target gas due to its unique paramagnetic nature. [60]Figure S4, (Supporting Information) depicts the influence of the preparation temperatures on the material's gas detection efficiency.It is observed that the material exhibits a response to NO 2 gas starting from a calcination temperature of 600 °C.This optimized temperature is consistently employed throughout the study.From Figure 4a, the zinc oxysulfide sensor demonstrates a fully reversible sensing response with a response factor of ≈2.27 toward 1.26 ppm of NO 2 under blue light illumination ( = 460 nm).Furthermore, although the zinc oxysulfide still responds to the NO 2 gas under other light conditions, the response factors are relatively low that are ≈1.49and ≈1.39 for green and red light, respectively, their recovery phases are all shown in an incomplete manner compared to that of blue light, which is in excellent agreement with the optical properties of zinc oxysulfide as discussed in Figure 3b,d.Such a sensing response mainly originates from a physisorptive behavior between the gas molecules and the material.As the blue light reaches the material surface, abundant electron-hole pairs are significantly generated from zinc oxysulfide and effectively separated to participate in the gas sensing interactions, owing to its substantially prolonged excitonic radiative lifetime.When paramagnetic NO 2 gas molecules are adsorbed upon the nanoflakes, interfacial electric dipoles are formed resulting in charge transfers between the NO 2 molecules and the material, causing a redistribution of surface charge density on the zinc oxysulfide surface. [13,14,20,30,31]Consequently, the base resistance of the zinc oxysulfide sensor increases in correlation with the reduction of the free carriers, forming a response phase as shown in Figure 4a.Reversely, the sensing recovery can be facilitated by the desorption of the NO 2 molecules in the air expose phase when the photogenerated carriers are set free back to the zinc oxysulfide body. [61]In addition, relatively weaker excitonic interactions with the light wavelength beyond the blue region cause insufficient electron-hole pairs, resulting in a smaller response magnitude.The small number of photogenerated electrons and holes including the intrinsic free carriers may be easily trapped in the material defects, possibly leading to the absence of the recovery phases. [62]Therefore, 460 nm is selected as an optimized excitonic light wavelength in the rest of the experiments.Additional tests using wavelengths of 520 nm and 620 nm demonstrated no change in initial resistance and no recovery kinetics following NO 2 exposure (Figure S5, Supporting Information).The sensor repeatability is continuously tested for an 8 h period with four repeated cycles of exposure to 1.26 ppm NO 2 , the test concludes with a minimized impact on the sensor performance as shown in Figure 4b.The equivalent resistance data supporting these findings can be found in Figure S6 (Supporting Information).From the dynamic sensing performance in Figure 4c, the sensor exhibits an almost linear response toward the NO 2 concentrations of 315, 630, 945, and 1260 ppb with response factors of 1.64, 1.73, 1.94, and 2.27, respectively.Moreover, the corresponding response and recovery time for each concentration are measured in Figure 4d, which are determined as the reacting time for 90% of full response magnitude and 10% of full recovery, respectively.As a result, the sensor demonstrates a relatively fast response toward the low-concentrated gas, entirely reacting to 315 ppb NO 2 in just ≈8 min.Given the linear regression of the response factors of the NO 2 concentrations (Figure 4c), the sensor's limit of detection (LOD), derived from three times the ratio of measuring noise level to NO 2 concentration slope, is calculated in 294.8 part-per-trillion (ppt) which is in the first rank of the semiconductors-based chemiresistive sensors (Figure 4e). [63,64]Additionally, we assessed the sensor's long-term stability by subjecting it to a daily exposure of 1.26 ppm NO 2 over a 5-day period.The sensor maintained an average response factor of 2.304 throughout this duration, as illustrated in Figure S7 (Supporting Information).The sensor is further confirmed in other commonly-used gas species with industrial meaningful concentrations including H 2 (1%), H 2 S (1120 ppb), CO 2 (2%), CH 4 (1%), and NH 3 (1000 ppb).As illustrated in Figure 4f, the zinc oxysulfide sensor is highly selective to the NO 2 gas over other tested analytes, indicating its excellent sensing selectivity.Additionally, the zinc oxysulfide sensor was evaluated in a 20% and 40% relative humidity (RH) humified environment in comparison with 0% humidity.According to Figure S8 (Supporting Information), the humidity exposure shows minor impact on the sensor NO 2 sensing performance.

Conclusion
We successfully synthesized the zinc oxysulfide ultrathin nanoflakes based on a two-step chemical reaction.The material forms into a 3D spherical structure due to a self-assembly agglomeration according to the SEM.The crystal and chemical composition of the obtained zinc oxysulfide are further investigated using XRD, TEM, XPS, and Raman spectroscopy, showing a ZnSO 4 lattice system with the chemical bonds of Zn, S, and O elements.Given its intrinsically strong optical absorption behavior and the substantial prolonged excitonic radiative lifetime, the ultrathin zinc oxysulfide is realized in a light-gas-matter interaction platform to test its room temperature gas sensing performance under visible light illumination.As a result, the zinc oxysulfide sensor demonstrates a fully reversible and repeatable sensing response toward 1.26 ppm NO 2 gas with a response magnitude of ≈2.27 under a blue light excitation source ( = 460 nm) at room temperature.The sensing linearity is scrutinized through a dynamic experiment, in which the sensor shows a linear response from low to high concentrated NO 2 with a LOD calculated as 294.8 ppt, which is in the first rank of chemiresistive sensors.The sensor also exhibits a relatively fast response to lowconcentrated gas, entirely reacting to 315 ppb NO 2 in just ≈8 min.In addition, the zinc oxysulfide sensor highly selects NO 2 gas in comparison with other commonly used gas species including H 2 , H 2 S, CO 2 , CH 4, and NH 3 .Therefore, we believe that this work may pave the way for the utilization of ultrathin metal oxysulfides in the new generation of high-performance room temperature visible-light-assisted gas sensors.

Experimental Section
Material Synthesis: The zinc sulfide powder was synthesized from a mixture of Zinc Chloride (ZnCl 2 ) (270 mg) and Thiourea (CH 4 N 2 S) (300 mg) into 50 mL of deionized water.Before autoclaving the solution at 110 °C for 24 h, the mixture was vigorously stirred (500 revolutions per minute (rpm)) at 30 °C for 30 min.After cooling it to room temperature, the obtained solution was centrifugal washed for 20 min at 500 rpm.Collecting the precipitate for another 20 min of centrifugal washing in deionized water.The zinc oxysulfide powder could be finally prepared after drying up at 50 °C for 24 h; the zinc sulfide powder was annealed at 600 °C for 6 h at a temperature ramp speed of 300 °C h −1 with a constant flow rate of 200 standard cubic centimeters per minute (sccm) of compressed dry air.This was then followed by a cool-down phase of a ramp-down temperature speed of 300 °C per hour, in which the collected zinc oxysulfide was prepared by dispensing 1 mg of the annealed powder into 10 mL ethanol solution.
Material Characterization: The structure of the zinc oxysulfide powder was examined using a scanning electron microscope with an accelerating voltage set to 3 kV (FEI Nova NanoSEM 200, FEI Company, Hillsboro, OR, USA).A transmission electron microscopy with a built-in EDS detector was used to study the crystal morphology of zinc oxysulfide's SAED patterns, crystal lattices, and chemical composition (JEOL JEM-F200, Jeol, Tokyo, Japan).Atomic force microscope (AFM) was used to measure the thickness of zinc oxysulfide (Bruker, Icon, Billerica, MA, USA).Crystal morphology was further inspected with XRD measurements with the equipment monochromatic radiation source Cu K at  = 0.154 nm (Bruker D4, ENDEAVOR, Bruker, Billerica, MA, USA).The chemical composition of zinc oxysulfide was further verified with XPS using a dual Al/Ag monochromatic X-ray source of Al K X-rays at 1486.7 eV (AXIS Supra XPS, Krato, Manchester, UK).UV-VIS-NIR microspectrophotometer was used for the absorption spectrum of zinc oxysulfide (CRAIC Apollo, CRAIC Technologies, San Dimas, CA, USA).Raman spectroscopy was analyzed at excitation wavelength of 532 nm (LabRAM HR Evolution, Horiba Scientific, Kyoto, Japan).The photoluminescence lifetime measurements were conducted with a fluorescence spectrometer equipped with a thermoelectrically cooled Si CCD detector using an excitation wavelength of  = 0.154 nm (QE Pro spectrometer, Orlando, FL, USA).
Sensor Fabrication and Measurements: An IDT containing 200 pairs of gold electrodes with a spacing of 10 m on a SiO 2 substrate acts as the sensing medium (HORX Sensortech, Melbourne, Australia).The prepared zinc oxysulfide solution in ethanol was sonicated for 10 min to agitate the material in the solution for homogeneity.40 L from the zinc oxysulfide solution was aspirated and drop-casted onto the IDT sensing substrate as the NO 2 sensor.For resistance measurements of the NO 2 sensor, an air-tight customized gas chamber was built to house the sensor where an internal probe stage connects to the sensor.The resistance was continuously measured throughout the gas sensor experiment via a digital multimeter (DMM 34465A, Keysight, Keysight Technologies Australia Pty Ltd, Australia).Meanwhile, a programmable multichannel gas calibration system was used to regulate and control the gaseous mixture in the gas chamber (EL-FLOW Mass Flow Controller, Bronkhorst, Leonhardsbuch, Germany).An LED was applied directly upon the sensor substrate with changeable wavelengths of 460, 520, and 620 nm, acting as a low-power excitonic light source with a power density ≈0.36 mW cm −2 .
Statistical Analysis: For Figure 2, a five-data-point average window smoothing was applied to the raw XPS data to generate the fitted curve.In Figure 3D, an Exponential Decay 3 function was used in OriginPro 2021 for PL data fitting, with fitting parameters disclosed in Note S1, (Supporting Information).The LOD was calculated using the formula (3.3 × )/S, where  is the standard deviation of the signal-to-noise response and S is the slope of the calibration curve for dynamic response.

Figure 2 .
Figure 2. Zinc oxysulfide XPS measurements show the raw and fitted data of a) Zn 2p, b) S 2p, and c) O1 s spectrum.

Figure 3 .
Figure 3. a) Raman spectrum for zinc oxysulfide, b) UV-VIS-NIR absorption spectrum with the corresponding Tauc plot shown in inset, c) PL, and (d) PL lifetime.

Figure 4 .
Figure 4. Zinc oxysulfide room temperature NO 2 gas sensing performance evaluated in air balance, a) reversible gas sensing performance in different visible-light wavelengths, b) repeatability, c) NO 2 dynamic response, d) response/recovery time, e) the linear fitting of the response based on the SD dynamic response and calculated NO 2 LOD of zinc oxysulfide, and f) selectivity.