Continuous hydrothermal flow-inspired synthesis and ultra-fast ammonia and humidity room-temperature sensor activities of WO3 nanobricks

Mesoporous tungsten oxide nanobricks (WO3 NBs) are successfully prepared via a simple and cost-effective hydrothermal synthesis method. The as-synthesized WO3 NBs demonstrate a high sensitivity and selectivity when used for liquid ammonia and humidity sensor activities at room temperature (27 °C). The monoclinic crystal structure has beencorroborated from thex-ray diffraction studies and the specific surface area is estimated to be 38.74 m2g−1. A larger specific surface area has significantly facilitated a fast gas adsorption/desorption process. The WO3 NBs notably exhibite gas sensitivity and selectivity for volatile organic compounds (VOCs) such as ammonia; however, a moderate performance is displayed with different oxidizing and reducing agents at room-temperature, namely: toluene, methanol, ethanol, and acetone. The sensor has offered a commercial potential with an extremely high response (75%), a 15-day operational stability at 100 ppm concentration of ammonia, and a practically remarkable ultra-high 8/5 s response/recovery time. The WO3 NBs-based humidity sensor endows a 32% resistance response at 20% relative humidity, with a quick response/recovery time of 10/8 s; which is due to unique surface architecture of these NBs.

properties, which have made it a appropriate material for various applications such as catalysts [3], gas sensors [4], and energy storage applications [5]. Up till, myriads of chemical and physical preparation methods like thermal evaporation, chemical vapor deposition, the sputtering, and colloidal suspension, have been applied for the synthesis of WO 3 nanostructures of different phases and morphologies; nanoparticles, nanofiber, nanorods and nanowires [6,7]. All reported synthesis methods could be either expensive, time consuming or tedious, with huge hazardous waste as toxic byproducts. Hydrothermal synthesis has received great consideration because of its simplicity, cost-effectiveness, soft chemical approach, low-temperature processing, and uniqueness in producing controlled shapes/sizes of various nanostructures [8,9].
Siciliano et al reported a direct growth of WO 3 film in an oxygen atmosphere using metal tungsten foil via a thermal evaporation method on a sapphire substrate which demonstrated a high sensitivity for NO 2 gas, and less for NO and NH 3 gases [10]. Ciateanu et al used a pulsed laser deposition method of mediated crystalline WO 3 films as oxygen gas detectors [11]. Zeng et al obtained porous WO 3 -based NO 2 gas sensors using anodic oxidation of DC magnetron sputtering method onto an alumina substrate [12]. Surveying the literature in depth, it is inferred that the WO 3 is a potential sensor material because of its large surface area and unique structural, optical, and electrical properties. However, WO 3 as a sensor described in the literature has endowed its effective sensing operation above 150°C; not in position of fabricating low-temperature operating sensor devices. Moreover, a high-temperature operation endows limited practical applications, viz., detection of flammable gases [13]. Several consecutive efforts have developed to reduce the operation temperature of these sensors likewise; the use of metal oxide-based gas sensors to detect various target gases at atmospheric temperature [14]. Teoh et al and Kim et al demonstrated WO 3 room-temperature-based sensor that detected NO 2 and NH 3 gases with moderate values of response and recovery time [15,16]. Moreover, above drawbacks might overcome by controlling the structure, morphology and surface area of WO 3, prepared by soft-chemical methods.
In present study, a facile, simple and low-cost hydrothermal method is used to obtain WO 3 NB gas sensors operating at room temperature. The WO 3 NB gas sensors approve higher selectivity towards ammonia, than other VOCs such as ethanol, methanol, acetone, and toluene. The sensor for ammonia gas endows an ultra-fast response/recovery time. The schematic of gas sensing mechanism using a band structure change is closely analogous to experimental findings. The WO 3 NBs gas sensor has also demonstrated remarkable sensing response for humidity.

Synthesis of WO 3 NBs
The analytical grade ammonium paratungstate (NH 4 ) 10 [H 2 W 12 O 42 ]·4H 2 O, 99%), hydrochloric acid (HCl, 37%), and hydrogen peroxide (H 2 O 2 , 30% w/w in water) were used as received. In a facile synthesis process, 1 g of ammonium paratungstate was mixed in 95 ml distilled water (Milli-Q water; 18.2 MΩ cm) and concentrated HCl (∼3 ml) and then stirred for a while. The stirring time was depending on the amount of HCl added. After the addition of nearly 2 ml of H 2 O 2 , the ammonium paratungstate solution initially turned to a transparent light yellow and then dark yellow by forming gelatinous precipitate, which was evidenced after continuously 1 h stirring to form transparent solution. A transparent solution was poured in a stainless-steel Teflon-lined autoclave. The hydrothermal synthesis reaction was conducted at a constant temperature of 160°C for ∼14 h (heating rate 5°C min −1 ). The whitish powder was taken off, rinsed with distilled water and calcinated at 500°C for 1 h to increase the crystallinity of the as-obtained WO 3 NBs (Scheme 1). The solid-state film was fabricated by taking 1 g of polyvinyl alcohol (PVA) in 10 ml of distilled water as a binder, which was kept at 90°C while constant stirring until the solution became viscous. Then, a few drops of viscous PVA solution was mixed with the WO 3 NBs in a pestle and mortar. This mixture was coated onto a soda-lime glass substrate and dried at 300°C to vaporize the excess water and binder. The obtained film was used as a room-temperature gas and humidity sensor.

Characterization details
The optical transparency of WO 3 NBs film was measured using a UV-visible spectrophotometer (V-530, Jasco, Japan). The DTA-TGA study measurement was carried out using a DSC-TGA standard instrument (SDT Q 600 v 20.9 Build 20). The surface morphology of the WO 3 was confirmed from FE-SEM (Nova-SEM 200-FEI) digital silhouettes. The HR-TEM image was recorded using a FEI TECNAI G2 20 STWIN instrument. The x-ray diffraction (XRD) spectrum was used to investigate the crystal phase (XRD-6000, x-ray diffractometer, Shimadzu) with a Cu-Kα radiation tube. The x-ray photoelectron spectroscopy profile was acquired for determining the elemental composition via a PHI 5000 Versa Probe (Ulvac-PHI). Moreover, the BET surfacearea and pore-size distribution analyses were measured on Belsorp II, BET, Japan Inc. instrument.

Gas sensor setup
The gas sensor performance was measured using 250 ml volume capacity stainless-steel chamber.
PID-controlled heater was fixed at the base of the cylindrical chamber to adjust the temperature. A dimmer stat was used as a fixed voltage source to protect the variations in the operating temperature. A Keithley 6514 programmable computer connected electrometer was employed to record change in the sensor resistance with respect to time after inserting target gases in the stainless-steel chamber. For sensor measurements, thin films of 1.5×1.5 cm 2 area were fixed and counted over the measurement unit (Scheme 2). The electrical contacts were drawn using commercial silver paste. the gas sensor response was calculated using following equation [16]: where, R a is stabilized resistance of atmospheric air, R g is stabilized resistance of target gas. A static liquid-gas distribution analysis method was preferred for calculating the concentration of VOCs by applying the following formula [8]: where, C (ppm) is the concentration of target gases; ρ is (g ml −1 ) is the liquid density; V is the volume of liquid ammonia (μl), T is temperature (K), M is the molecular weight of ammonia (g mol −1 ), and V is the stainless-steel chamber volume (L). In this study, the M, ρ and V for ammonia are constant for gas sensor measurement and are 17.03 g mol −1 , 0.68 g cm −3 , and 0.250 l, respectively.

Reaction mechanism
In hydrothermal synthesis a white-colored precipitate was settled at the bottom of the Teflon-lined autoclave. The plausible chemical reactions responsible for forming WO 3 NBs powder are proposed as equations (3) (6)). Finally, the as-obtained white-colored WO 2 .(OH) 2 powder was air-calcinated for 1 h at 500°C to form pale green WO 3 NBs (equation (7)).

Optical transparency and surface morphology analyses
A representative FE-SEM image of the WO 3 product has demonstrated uniformly distributed surface morphology of NB arrays. The side-view (figure 1(a)) clearly shows that the surfaces of the WO 3 NBs were smooth with the approximately 100 nm of diameter and the 0.5 μm of length. Figure S1(a †), available online at stacks.iop.org/MRX/7/015076/mmedia demonstrates the low magnification FE-SEM image of WO 3 NBs. Figure S1(b †) shows the energy dispersive spectroscopy spectra of WO 3 NBs and the analyzed peaks confirmed the phase purity and the presence of W and O atoms. The carbon peaks present in the energy dispersive spectroscopy spectra are due to carbon tape used during FE-SEM scanning. Furthermore, the HR-TEM image of the as-prepared WO 3 NBs has approved a single crystalline character with fine and regular lattice fringes over scanned area ( figure 1(b)). The 0.36 nm lattice spacing measured from the HR-TEM image, is closely supporting to (200) diffraction peak of monoclinic WO 3 [19]. The SAED pattern exhibited a regular sharp spot which is in accord with highly crystalline nature of the as-obtained WO 3 NBs. Notably, XRD spectrum evidenced the formation of polycrystalline WO 3 (discussed in the followed section), a little contradictory to SAED analysis, could be attributed to difference in the scanning areas of two measurements. The HR-TEM attached EDAX analysis highlights both W and O elements which are equivalently distributed throughout the product. The Cu peaks were also appeared due to the copper grade substrate used while scanning (figure 1(c)). Figure S2(a † ) shows the optical absorbance spectrum at around 460 nm, evidencing a visible light transparency signature. The inset of figure S2(a † ) presents the Tauc plot for WO 3 film (1.5 μm thickness estimated from cross-sectional image), from which the E g was estimated. The 2.82 eV E g value for the WO 3 NBs film is closely matching to a number results reported previously [20]. The DTA measurement (figure S2(b † )) was used to confirm the thermal behavior of the material. The DTA measurement depicted two peak positions, relatively small endothermic peak at 36.4°C and broad exothermic peak at 221.3°C for the removal of water molecule and the synthesized WO 3 phase from WO 3 ·H 2 O [21].

Structural elucidation
The XRD pattern was used to confirm phase purity and crystal plane of the WO 3 NBs. As-seen in figure 2(a), the XRD plane revealed the diffraction peaks of monoclinic crystal phase with a=0.7297 nm, b=0.7539 nm, c=0.7688 nm and β=90.91°which is in close agreement with known data (JCPDS card no. 43-1035).
Presence of broad and strong diffraction peaks is approving polycrystalline signature of the as-obtained WO 3 NBs film. A strongly preferential growth direction [200] is noticed. In figure 2 [22]. The XPS spectroscopy measurement provides a valuable quantitative and chemical state information of the material. Two 4f 7/2 (35.7 eV) and 4f 5/2 (37.6 eV) peaks, due to a dual state of tungsten, were corroborated in a wide XPS scan ( figure 3(a)). In consistent to literature value, the W4f core-level at 35.3 eV was 37.8 eV. The W4f spectrum with metal and oxide presents the 4f 7/2 -4f 5/2 doubles match to the W 6+ oxidation state from WO 3 [23]. The XPS as depicted in figure 3

Surface area and pore-size distribution studies
To determine the specific surface area, and pore-size distribution of the WO 3 NBs, the nitrogen adsorptiondesorption isotherm was measured and is shown in figure 4. The nitrogen adsorption-desorption isotherm  demonstrated type IV hysteresis loop at relative pressure (P/P 0 ) between 0.4 and 1.0, suggesting the presence of mesoporous nature [25]. According to the BET method, a typical value for the specific surface area was 38.74 m 2 g -1 . The inset plot of the as-obtained WO 3 NBs exhibited a small pore-size distribution centered at 10.76 nm. The average specific surface area and small pore-size distribution of WO 3 NBs would facilitate as easy mass/ charge transportation for gas molecules with a high sensitivity (discussed below) [26].

Gas sensing properties
Sensitivity and selectivity are two significant parameters for gas sensor measurement. From a commercial point of view, the sensors must possess a good selectivity and high sensitivity, as well as considerable stability. In this study, we measured the sensitivities of the WO 3 NBs film as a sensor material toward five VOCs at a fixed 100 ppm concentration, namely: ammonia, acetone, methanol, ethanol, and toluene. Figure 5(a) shows the gross sensitivity measurements towards VOCs, where the WO 3 NBs sensor demonstrated higher sensitivity at room temperature for ammonia ( figure 5(a)). The WO 3 sensor was almost insensitive for methanol and toluene.
In figure 5(b), the impact of various temperatures on the gas sensor performance for 100 ppm ammonia was evaluated. The WO 3 NBs sensor produced a maximum response of 75% for 100 ppm ammonia at roomtemperature. This can be elucidated with the help of following reasons. Firstly, the mesoporous structures could favor a higher surface area of WO 3 NBs that can facilitate easy adsorption of excessive gas molecules. The amount of O 2 − species present at material surface can be enhanced by increasing the surface area and pore engineering, which can provide more active sites for adsorption of gas molecules [27]. Secondly, the higher  temperatures might suppress the gas adsorption by rising the operation rate. Finally, as ammonia is unstable at higher temperatures, lower response values can be resulted [2]. Figure 5(c) demonstrates the dynamic response of WO 3 NBs sensors ranging from 10 to 1000 ppm levels of ammonia concentrations. The gas response enhanced with ammonia concentration and was sensitive even at less than 10 ppm of ammonia. In sensor measurement, the response transient curve showed that the resistance reached almost to its original value after the exclusion of ammonia gas, indicating excellent gas sensor reversibility [28]. The responses of the sensor upon exposure to 10, 50, 100, 200, 400, 800 and 1000 ppm of ammonia concentrations were respectively 35, 52, 75, 88, 108, 127 and 132%. The low response at 10 ppm ammonia concentration was due to lower adsorption of ammonia molecules, thereby there could a slow sensing reaction rate. The higher response at 1000 ppm ammonia concentration was because of a larger quantity of ammonia gas molecules adsorbed on the reaction surface. However, there was no significant increase in the response for 800 and 1000 ppm concentrations of ammonia, which might be because of the limited accessibility of active surface sites on the sensing material surface for adsorption reaction. Furthermore, figure 6(a) demonstrates the transient resistance versus time curve of WO 3 NBs sensors with 100 ppm of ammonia. Notably, on exposing the reducing ammonia, the resistance of the sensor material drastically decreased (for 8 s), then stabilized after some time. The gas sensor recovery was measured by exposing atmospheric air in sensing chamber, confirmed an excellent recovery in just 5 s. The 8/5 s response/recovery time of the WO 3 NBs film sensor represented its ultrafast sensing activity for potential commercial viability. The response/recovery time of gas sensors mainly depends on how quick the gas diffuses in/out [29]. The response/recovery time is an imperative parameter for calculating the gas-sensing performance of a material. Ideally, both should be equal. But, the presence of stacking faults, active sites and structural irregularities make them different. The response/recovery time achieves a maximum sensing response of 90% on exposure to target gases/time and decreases to 10% upon air exposure [30]. The fast response/recovery time of the WO 3 NBs film sensor was because of its special NB-typemorphology, allowing an easy and fast adsorption/desorption of ammonia gas molecules, which is one of the appealing measurements. Figure 6(b) shows transient response and recovery curves for 100 ppm of ammonia concentration at room temperature. From the gas response curve, the WO 3 NBs sensor demonstrate a highest response of 75% at 100 ppm concentration of ammonia gas within 8 s response time which is better than 92% at 250 ppm (4 min response time) reported by Mintcheva et al for Sn-Zn alloy-based room-temperature sensors [31]. The quick response/recovery time could be due to the higher surface area (38.74 m 2 g −1 ); this increased surface area is responsible for higher adsorption of the target gas because one-dimensional morphology promotes a faster charge transportation. The higher surface area and NB-type surface morphology of WO 3 could have assisted a better room-temperature gas sensing performance. A reproducibility test of the WO 3 NBs sensor at 100 ppm of ammonia was also studied and is revealed in figure 6(c). The sensor response remained nearly same during continuous-cycling stability tests. Gas sensor devices should avail not only a high sensitivity for the target gases, but also a better stability [32]. Figure 6(d) shows the cyclic stability test of WO 3 NBs sensor film measured at a concentration of 100 ppm ammonia for the period of 15 days time. The response remained closed to 75% at 100 ppm concentration of ammonia gas, suggesting industrial potential of WO 3 NB-based gas sensors.

Sensing mechanism
The gas sensing mechanism is shown in schematic diagram (scheme 3). The (metal oxide-based) gas sensing mechanism can be elucidated through changing depletion layer of oxygen adsorption. The WO 3 is n-type semiconductor material, and therefore electrons are the essential source of charge carriers [33]. When WO 3 sensor film is open to ambient air, the air molecules of oxygen adsorb on its surface by withdrawing electrons to form surface adsorbed oxygen species, which eventually can reduce the number of electrons from the conduction band of WO 3 . The oxygen adsorbed on surface of the WO 3 sensor film helps to manipulate the resistance of the gas sensor as the adsorbed oxygen is temperature dependent [34]. In our case, O 2 − is dominant as the WO 3 sensor is operated at room temperature. Before exposure to ammonia, care should be taken that the resistance of the WO 3 NBs sensor should be stabilized. It can be seen that, in presence of ammonia, resistance of WO 3 NBs sensor decreased quickly with respect to time, and stable resistance was accomplished within few seconds (8 s), suggesting a rapid response to ammonia. Upon the interface of ammonia with the WO 3 NBs sensor, the previously adsorbed O 2 − species could play a major role in the reaction. Consequently, adsorbed oxygen species get replaced with ammonia molecules because of their reducing character, decreasing the resistance of the WO 3 NBs sensor. A plausible ammonia sensing mechanism through chemical reactions can be as follows [35]: Further, the gas sensing reaction mechanism might be elucidated with the following points. Firstly, WO 3 , being a metal oxide, follows an electron transport mechanism similar to other metal oxides. The electron transportation takes place from WO 3 to O 2 , and potential barrier is taking place at the surface of sensor material. The resulting potential barrier hinders electron transportation throughout the material; subsequently, the sensing layer surface adsorbs a large number of oxygen species, which can enhance the gas sensor performance. Secondly, the capability of test molecules to arrive on the depletion region also determine the sensing performance of materials [36]. This is a wellknown phenomenon: if the size of the gas molecule is smaller there is more probability of penetrating through depletion region. The ammonia molecule has lowest (kinetic diameter of 0.26 nm [37] compared with ethanol (0.45 nm) and acetone (0.469 nm) molecules [38]. Therefore, the gas sensitivity towards ammonia is higher compared to the other VOCs that we considered. Thirdly, particle size can also be acknowledged as a parameter for gas sensing behavior. In the case of WO 3 , the Debye length is approximately 33 nm [39], which is much smaller than the size (i.e., 85 nm) of WO 3 NBs. Therefore, the gas sensing mechanism of WO 3 NBs follows the Schottky-barrier-controlled model. Moreover, porosity also plays an important role in the gas sensing behavior of sensor materials. The calculated average pore-size was in the range of 10-15 nm, suggesting a mesoporous nature of the materials [40].

Humidity sensing activity
The humidity-sensing performance of WO 3 NBs film sensor was carried out using a binary electrode under various (20−80%) relative humidity (RH) conditions. Saturated salt solutions were used to monitor RH conditions at room-temperature and variations in the resistance of WO 3 NBs film sensor under 20% RH conditions are shown in figure 7(a). The saturated salt solution contains H 2 O molecules that can act as electron donors, and thereby, the film resistance of WO 3 NBs sensors droped down with the RH conditions. The humidity-sensing mechanism, the Fermi level, of a material which is nearby to the conduction band edge could modify after adsorbing H 2 O molecules on the surface of the sensor [41]. The response/recovery time of the WO 3 NBs sensor was 10/8 s. The ultra-fast response (10 s)/ recovery time (8 s) was determined by the NBs type of WO 3 architecture and the fast desorption process of H 2 O molecules from the WO 3 NBs sensor surface, respectively; resulting in the fast adsorption/desorption of H 2 O molecules. Bharatula et al reported that the thermodynamic adsorption of analyte molecules may not be favorable due to strong adsorption/desorption at room temperature. During this process, the analyte molecules get desorbed easily due to the low absorption energy, which leads to an increase in the response time compared to the recovery time [42]. The result of various RH conditions on sensor sensitivity is shown in figure 7(b); thus, confirming an improvement in the response of the WO 3 NBs sensor to humidity with the RH conditions. It was concluded that, in addition to ammonia, hydrothermally-synthesized WO 3 NBs sensors directed an ultra-high response/recovery time for humidity as well. The room-temperature sensitivities of 75 and 33% were obtained for the WO 3 NBs sensor for ammonia and humidity, respectively; suggesting its practical and commercial perspectives. The gas sensing performance of WO 3 NBs sensor was correlated with previously reported data (shown in table 1), wherein higher response values for ammonia sensing were noticed. In previous studies, the sensor operating temperature was a major constraint; plus, doping of noble metals and metal oxides into the host WO 3 materials were introduced to enhance the sensor response. Present WO 3 NBs film sensors detect ammonia with considerable response at room temperature, which has its own creditials.

Conclusions
In conclusion, mesoporous WO 3 NBs were prepared via a simple and cost-effective hydrothermal synthesis method. A film of WO 3 NBs was fabricated directly onto a borosil-glass substrate using a doctor-blade method. The gas sensing performance of WO 3 NBs sensor was measured for the detection of ammonia and humidity at room-temperature (27 ℃). The highest response of 75% and ultrafast 8/5 s response/recovery time were obtained at 100 ppm concentration of ammonia. In addition, an excellent humidity response (32%) was displayed by the WO 3 NBs sensor at a relative humidity (RH) of 20%. The fast response time (10 s) of humidity sensing is due to the hydrophilic surface of the WO 3 NBs can be accounted for a slow H 2 O molecules adsorption and an ultra-fast recovery time of 8 s; this could be because of the rapid desorption of H 2 O molecules on the surface of WO 3 NBs.