NH3 Sensing Performance of Pt-Doped WO3·0.33H2O Microshuttles Induced From Scheelite Leaching Solution

WO 3 ·0.33H 2 O microshuttles (WMSs) self-assembled by numerous nanorods along the same direction were prepared based on a cheap tungsten-containing metallurgical raw material by combination processes of NaOH leaching and one-step hydrothermal method. The microstructures and gas sensing properties of various concentrations (0, 0.7, 1.0, and 1.3 mol%) of Pt-doped WMSs were investigated to improve their performance. The microstructural characterizations demonstrated that the WMSs assembled by one-dimensional WO 3 ·0.33H 2 O nanorods were approximately 0.8−1.9 µm in diameter. Such nanorods exhibited a single hexagonal structure with their diameters ranging from 17 to 62 nm. The gas sensing properties indicated that Pt-doped WMSs showed superior gas sensing performance in terms of the sensor response and NH 3 selectivity in the operating temperature range of 25−225 o C as compared with pure one, and simultaneously Pt doping could signicantly reduce the detection limit of NH 3 . Especially, 1.0 mol% Pt-doped WMSs exhibited highest response of 28.2 to 1000 ppm NH 3 at 175 o C, which was 4 times higher than pure one at 50 o C. The remarkably enhanced gas sensing performance of Pt-doped WMSs to NH 3 could be ascribed to the electronic and chemical sensitization mechanisms of noble metal nanoparticles.


Introduction
With the rapid development of society and economy, human health problem has gradually become an increasing topic. Ammonia (NH 3 ) with a strong pungent odor is a toxic and colorless gas, and easy to dissolve in water [1,2]. It is a very important chemical product and industrial raw material, which is widely used in decoration, chemical fertilizer, medicine, industrial refrigeration and other important elds [3][4][5][6][7][8]. In addition, NH 3 is also a kind of indoor air pollutant with great harmfulness, which is prevalent in our daily life, such as protein rich corruption, interior decoration materials and home adhesives, etc., which could cause serious harm to human health even under a low concentration by damaging throat, lungs, eyes, and skin [9][10][11]. NH 3 is also explosive, and there is a potential risk of explosion when mixed with air [12]. According to the relevant reports of some safety and health institutions, long-term inhalation or exposure of high concentration (more than 5000 ppm) of NH 3 can seriously endanger human health, and even cause death. The immediately dangerous to life or health concentration (IDLH) towards NH 3 is 300 ppm [13][14][15][16]. When the human body is in the low concentration of NH 3 environment such as 35 ppm for a long time, it will also cause the increase of blood ammonia concentration, which will lead to a serious of pathological changes of organs such as liver and kidney [14,17,18]. Therefore, it is very important to develop a new NH 3 gas sensor with high response and selectivity. WO 3 is a common transition-state semiconductor metal oxide, which is recognized as a sensing material with good application prospects in the eld of gas sensors due to its many excellent physical and chemical properties, such as unique non-stoichiometric, narrow band gap, strong acid resistance and high temperature resistance [19]. Tungsten trioxide hydrate (WO 3 ·0.33H 2 O) is a kind of hydrate of WO 3 , which is widely applied as gas sensing material because of its excellent physical and chemical properties. At present, there are relatively few researches focusing on WO 3 ·0.33H 2 O sensing materials. Single-component WO 3 ·0.33H 2 O sensing materials often have some drawbacks in the eld of gas sensors, such as poor selectivity, low response, high operating temperature, slow response/recovery speed, etc., so that they cannot be well applied to monitor the toxic and harmful gases. For the sake of improving the gas-sensing performance, a variety of morphologies and structures of WO 3 ·0.33H 2 O nanomaterials have been synthesized, including 0-D nanoparticles [20], 1-D nanorods [21], 2-D nanosheets [22], and 3-D microspheres [23]. In addition, it is also an effective way to enhance their gas-sensing performance by adding noble metal elements into the sensing materials. Ganesh et al. [24] reported that some noble metal doping in the nanomaterials could change their morphology and energy band structure, which resulted in the increase of speci c surface area and the formation of more active reaction center of gas molecules on the nanomaterial surface, thereby leading to the enhancement in the gas sensing properties of nanomaterials. Thai et al. [25] demonstrated that some noble metal catalysts such as Pt or Pd possessed strong reactivity to some gaseous molecules, especially some containing-hydrogen atoms molecules, such as NH 3 , ethanol, H 2 , and lique ed petroleum gas. Chen et al. [26] synthesized a Pt/NiO thin lm-based gas sensor for monitoring NH 3 . They indicated that NH 3 molecules could be effectively and directly dissociated into nitrogen and hydrogen molecules by introducing Pt in the sensing materials.
Heretofore, the tungsten sources used for synthesizing functional WO 3 nanomaterials are mainly some analytical reagents or extremely high-purity tungsten metals, such as Na 2 WO 4 [27], H 2 WO 4 [28], (NH 4 ) 10 [30], etc., which not only have high production costs, but also show low plasticity in terms of purity, particle size, etc. Thus, these disadvantages will affect the microstructure and physic-chemical properties of WO 3 nanomaterials prepared later. Therefore, it is of great signi cance to nd a tungsten source material with low price and good plasticity to prepare WO 3 functional nanomaterials.
Scheelite is a kind of tungsten mineral with tetragonal crystal system, which is the main raw material for tungsten smelting. Some noble metal elements are often associated with scheelite, and these noble metal elements may be conducive to improving the gas sensing performance of nanomaterials. In this paper, a tungsten-containing metallurgical raw material was selected as the tungsten source, and the optimized alkali extraction process of tungsten was adopted to obtain a test-grade tungsten-containing leaching solution with low impurity content.
Subsequently, the WO 3 ·0.33H 2 O microshuttles (WMSs) with stable crystal phase, excellent purity and good dispersivity were prepared by a one-step hydrothermal method with the leaching solution as the precursor solution.
For the sake of further investigating the in uence of noble metal elements on the microstructure and gas sensing performance of WMSs, the in-situ doping of Pt was used to simulate the associated noble metal elements in scheelite.
The results showed that Pt doping could signi cantly improve the performance of the WO 3 ·0.33H 2 O sensing materials to NH 3 , and the relevant gas sensing mechanism was also discussed.

Materials
The tungsten-containing metallurgical raw material used in this study was mainly composed of scheelite minerals, which was provided by Gansu Xinzhou Mining Co., Ltd. In all the experiments, the as-used reagents were analytical grade (AR) without further puri cation, and the as-used water was deionized water. In this study, an X-ray Fluorescence spectrometer (XRF) was used to determine the main components of the metallurgical raw material, and the corresponding results are illustrated in Table 1. As observed in Table 1, the main component in the metallurgical raw material is WO 3 (62.36%), and the main gangue compositions are CaO (19.11%), SiO 2 (5.55%), P 2 O 5 (3.99%), MgO (2.82%), and Fe 2 O 3 (1.55%). Subsequently, the sieve water analysis test was also carried out to further clarify the particle size distribution of the metallurgical raw material. As observed in Table 2, the part with the particle size of less than 38 µm accounts for 89.13% of the total mass of the metallurgical raw material. It demonstrates that the particle size of the metallurgical raw material is very ne, which can meet the test requirements and be directly used in the subsequent leaching experiments.

Preparation of the samples
The preparation process of the functional WMSs by adopting a cheap tungsten-containing metallurgical raw material as the tungsten source is different from that of the conventional nanomaterials, which mainly consists of the following two steps.
Firstly, the tungsten-containing leaching solution was prepared from a tungsten-containing metallurgical raw material by a NaOH leaching process. This step plays an important role in the WMSs preparation. It could not only convert solid-phase tungsten into liquid-phase tungsten, but also could remove most of impurities in the tungsten-containing metallurgical raw material. During the NaOH leaching process, most of the alkaline harmful impurities, such as iron, calcium, and magnesium, etc., remained in the leaching residue in the form of precipitates, while most of tungsten introduced into the leaching solution in the form of tungstate, realizing the effective separation of impurities and tungsten. The leaching experiments of the metallurgical raw material by NaOH as the leaching reagent were carried out in a 250 mL micro-high pressure autoclave (Shanghai Yanzheng Co., Ltd) equipped with a temperature sensor and a magnetic rotor. The speci c operation steps are as follows: the metallurgical raw material (5.0 g) and NaOH (3.5 g) were rstly put into the above autoclave in turn. Secondly, according to the pre-set liquid-solid ratio of 1:1 (quality ratio of deionized water to metallurgical raw material), a certain amount of deionized water was added into the above equipment. Then, the leaching experiment was carried out at 180 o C for 2 h, and the stirring speed was set as 400 rpm. After the reaction, the leaching slurry was transferred into a beaker, and some cleaning and ltering operations were carried out to get the tungsten-containing leaching solution. Subsequently, the content of main elements in the leaching solution was examined by an inductively coupled plasma atomic emission spectrometer (ICP-AES), and the relevant results are illustrated in Table 3. It can be observed from Table 3 that the content of tungsten in the leaching solution is relatively high and the contents of other impurities are low, and no other new impurities that are di cult to remove are introduced in the leaching process. The as-obtained tungsten-containing leaching solution could meet the requirements of the experimental-purity standard, so it could be directly used as the precursor solution for the preparation of WO 3 ·0.33H 2 O sensing materials in the future.
Secondly, using the as-obtained tungsten-containing leaching solution as the precursor solution, the WMSs were synthesized by a one-step hydrothermal method with the assistance of Na 2 SO 4 surfactant. The leaching solution contains a large amount of Na + ions and OH ─ ions, and Na + ions always acts as an important component of structure modi er in the synthesis process of WO 3 sensing materials. If its content is too high, it will directly affect the microstructure of the nanomaterials. Therefore, the Na + ions in the leaching solution should be removed rst in the preparation process. The detailed experiment process is as follows. Firstly, 10 mL of leaching solution was placed into a beaker with the capacity of 250 mL, and 3 M of HCl solution with a volume of 24 mL was slowly added into the above beaker with strongly stirring for 2 min to obtain tungstic acid precipitate. The as-obtained precipitation product was washed and ltered for 5 times with deionized water to remove the metal cations from the leaching solution.
Then, NaOH (0.27 g) and deionized water (50 mL) were mixed with the pre-treated tungstic acid precipitate, and continuous stirring was carried out until the as-obtained white tungstic acid was dissolved completely. Subsequently, 0.6 g of Na 2 SO 4 was added into the above mentioned mixture, and kept strongly stirring for 10 min. For the sake of obtaining the WMSs with different Pt concentrations, a certain amount of chloroplatinic acid (H 2 PtCl 6 ) solution based on the molar ratio (0, 0.7%, 1.0%, and 1.3%) of Pt to W was placed into the above mixed solution under continuously stirring for 10 min. The pH value of the mixture was adjusted to 2.2 using 3 M of HCl solution under intensely stirring for 10 min. Then, the as-obtained mixed solution was transferred to a 100 mL hydrothermal autoclave, and reacted at 140 o C for 12 h. When the hydrothermal reaction was completed, the as-obtained precipitation product was collected by washing and ltering with deionized water for 5 times, and then dry treatment was performed in an oven at 60 o C.
Finally, the grayish-white WMSs with stable crystal form were obtained by annealing the resultant product in a tube furnace at 400 o C for 4 h.

Characterization of the samples
An X-ray Fluorescence spectrometer (XRF, Shimadzu XRF-1800) was adopted to examine the elemental compositions of the tungsten-containing metallurgical raw material. An inductively coupled plasma atomic emission spectrometer (ICP-AES, Thermo Fisher Scienti c) was used to investigate the main elemental compositions of the as-prepared tungsten-containing leaching solution. An X-ray diffractometer (XRD, PANalytical X'Pert Pro, Cu-Kα radiation λ=1.5406 Å) was applied to investigate the main component and crystal structure of the WMSs powder. A eld-emission scanning electron microscope (FESEM, ZEISS Ultra Plus) was used to observe the surface morphology of the Ptdoped WMSs with different Pt concentrations. A high-resolution transmission electron microscope (TEM, FEIG 2 -20) was used to characterize the microstructure of the as-prepared nanomaterials. An energy-dispersive spectroscope (EDS) attached to TEM equipment was used to investigate the elemental distribution state of the WMSs. An X-ray photoelectron spectroscope (XPS, ESCALAB 250 Xi, Al Kα radiation) was adopted to characterize the elemental compositions and relevant valence state of the sensing materials. An infrared spectrometer (FTIR, NICOLET380) was used to analyze the functional groups and chemical bonds of the as-synthesized samples.

Fabrication of gas sensor device
The WMSs powder and an appropriate amount of absolute ethanol were placed in a 1.5 mL of centrifuge tube, and dispersed it for 15 min under ultrasonic to obtain a uniformly viscous slurry. Then, the slurry was evenly brushed on the ceramic tube surface, so that a pair of pre-plated annular Au electrodes was completely covered by the slurry to obtain a gas sensing lm. Among them, a pair of Pt leads was connected to each Au electrode, and a Ni-Cr alloy heating coil was traversed through the whole hollow ceramic tube to adjust operating temperature by changing the heating voltage. After the paste coated on the Au electrode was completely dried, the four Pt leads and the two ends of the heating coil were respectively welded to the corresponding terminal on the hexagonal base. Finally, the assembled gas sensor was aged at 300 o C for a week to strengthen the contact between the sensing particles and enhance the surface stability of the gas sensing material.

Gas sensing measurements
In this study, a static gas distribution system (WS-30A, Weisheng Electronics Co., Ltd) was used to investigate the performance of pure and Pt-doped WMSs sensing materials. The measurements were carried out in a fume hood with a relative humidity ranging from 28% to 30%, and the operating temperatures of all sensor samples were in the range of 75−225 o C. The target gas was NH 3 , and the other interference gases mainly included NO 2 , acetone, toluene, methanol, methanal, and SO 2 . For the preparation of NH 3 and other interfering gases, a micro-syringe was rstly adopted to extract the required volume of liquid reagents. Then, the liquid reagents were placed on the evaporation device in the test chamber, and the mixed gases could be obtained by mixing the evaporated gas and the air using a fan. The functional relationship between the needed liquid volume and the as-obtained gas concentration is shown as follows: Here, V x represents the volume of the required liquid reagent (mL), V refers to the volume of the test chamber (mL), C is the concentration of the detected gas, M is the molecular weight of the liquid reagent (g), P is the density of the liquid reagent (g/cm 3 ), d represents the purity of the liquid reagent, T R is the test ambient temperature ( o C), and T B is the temperature in the test chamber ( o C).
The response of the gas sensor to the detected gas is calculated according to the two formulas of S=R a /R g (for reducing gas) and S=R g /R a (for oxidizing gas), in which R a and R g refer to the resistance values of the gas sensor in clean air and detected gas atmosphere, respectively. The response time (T res ) refers to the time required for the resistance change of the gas sensor to achieve 90% of the nal balance value after injecting the detected gas, while the recovery time (T rec ) refers to the time required for the resistance of the gas sensor to recover to 90% of the initial resistance value after releasing the detected gas [31].  The microstructure, crystallinity and elemental distribution of 1.0 mol% Pt-doped WMSs were further measured by TEM and EDS technologies, and the corresponding results are illustrated in Fig.3. Figs. 3(a-c) are the low-resolution TEM images of this sample, which further prove that the as-synthesized sample with good dispersion is composed of WMSs self-assembled by nanorods. The nanorods are closely packed in the same direction, and the nanorods existed at the middle of the microshutters are obviously longer, which is consistent with the observation results of SEM images. Such kind of coarse structure can effectively increase the speci c surface area of the sensing material, which will be bene cial to the chemisorption of gaseous molecules as well as the subsequent reactions with the material surfaces. Fig. 3(d) is the high-resolution TEM image of the WMSs. Some very clear lattice fringes can be found, indicating a high crystallinity of the sample. Among them, the lattice spacing of 0.365 nm can match well with the (110) crystal plane of hexagonal WO 3 ·0.33H 2 O, and the lattice spacing of 0.227 nm can match well with the (020) crystal plane of PtO 2 . Fig. 3(e) illustrates the selected area electron diffraction (SAED) image of the WMSs. As observed in this gure, there are many neat and regular diffraction spots in the entire eld of view, indicating that the microshutters with excellent crystallization are composed of single-crystal nanoparticles. Figs. 3(f-i) are the distribution images of main elements of Pt-doped WMSs. As observed in these gures, the three elements of W, O and Pt are the main elements of the WMSs, and each element is uniformly distributed in the microshutters, indicating that Pt element is successfully doped into the microshutters by the in-situ hydrothermal method in the form of its oxidate.

Results And Discussion
The chemical valence state and elemental compositions of the 1.0 mol% Pt-doped WMSs characterized by XPS technology are presented in Fig. 4. The C 1s peak (284.8 eV) was regarded as a reference to calibrate the binding energies of the Pt-doped sample. The full scanning XPS spectrum of this sample shown in Fig. 4(a) indicates that the surface elements are comprised of primarily C and some W and O. As observed in Fig. 4(b), the peak located at 531.05 eV corresponds to O 1s, con rming the presence of lattice oxygen (O 2─ ), which can be bonded with W [32]. As seen in Fig. 4(c), the W 4f spectra of the WMSs can be deconvoluted into two strong peaks of W 4f 5/2 (37.43 eV) and W 4f 7/2 (35.33 eV), demonstrating that the tungsten in WO 3 ·0.33H 2 O crystal mainly exists in the form of W 6+ state [33]. In addition, a single peak at 40.98 eV can be observed, which originates from a weak emission of W 5P 3/2 [34]. Fig. 4(d) shows the XPS spectrum of Pt 4f. There is an obvious absorption peak located at 79.46 eV, and its binding energy is corresponding to Pt 4f, indicating the presence of PtO 2 in the Pt-doped WMSs [31].
The main functional groups and chemical bonds of pure and Pt-doped WMSs were investigated by FTIR technology, and the results are shown in Fig. 5. It is found that the four samples have similar FTIR spectrum, namely that the positions and intensities of the characteristic peaks are basically same. For each sample, two characteristic peaks appear at 713 and 881 cm -1 , both of which are ascribed to the stretching vibration of the bridging oxygen (W-O-W) [32]. A weak peak can be found at 1389 cm -1 , which can be contributed to the stretching vibration of W-OH [35]. The bending vibration band of O-H can be found at 1619 cm -1 for the sample [36]. The stretching vibration band of -OH appears at 3452 cm -1 [37]. In addition, the characteristic peaks of other impurities are not found, which further indicates that the four samples have higher purity. Simultaneously, there are no characteristic peaks related to Pt in the FTIR spectra, which may be attributed to the low amount of Pt in the samples.  Fig. 6, except for the pure one, the responses of all Pt-doped WMSs show an obvious trend of " rst rising and then descending" with increasing operating temperature. This is mainly ascribed to the chemical activities of the gas sensing materials and gas molecules, and the adsorption-desorption process of the gaseous molecules on the nanomaterial surfaces is greatly affected by the operating temperature [38][39][40][41][42][43]. The responses of all Pt-doped WMSs increase as the temperature increases in the range of 25−175 o C, and then decrease as the temperature value exceeds 175 o C. Therefore, the optimal operating temperature of Pt-doped WMSs towards NH 3 gas is 175 o C. Compared with Pt-doped WMSs, the pure one exhibits higher response at a lower temperature ranging from 25 to 100 o C, and the response reaches the maximum value at 50 o C. However, the response stability of the pure sample is relatively poor, and there is basically no response to NH 3  When Pt concentration is low, the catalyst dispersed on the surface of sensing material can only catalyze a part of NH 3 molecular. As Pt concentration further increases, the nely dispersed PtO 2 nanoparticles will trap more electrons, leading to the improvement in the N-H bond dissociation. Fig. 7 shows the response-recovery curves of the WMSs to 1000 ppm NH 3 gas at their optimal operating temperatures. As seen in Fig. 7, when the four gas sensors are placed in the NH 3 gas atmosphere, the resistance values of the gas sensors drop sharply, and then trend to be stable. When the NH 3 gas is released, the resistance values of all the gas sensors can be completely recovered to their initial values, demonstrating that the WMSs are ntype MOS materials. According to the analysis, the response and recovery times of pure WMSs to 1000 ppm NH 3  sensing material, but also greatly improve the response value, which will be very bene cial to the real-time detection of NH 3 gas. Fig. 8 presents the response-recovery curves of the WMSs to various concentrations (10, 30, 50, 100, 300, 500, and 1000 ppm) of NH 3 gas at optimal operating temperatures. As observed in Fig. 8, the amplitude changes of the resistance values of the four gas sensors obviously show a similar stepwise increasing trend with increasing NH 3 concentration, indicating that the responses also show the same increase trends. Among them, the pure WMSs have almost no response to the low concentration of NH 3 gas. In the case of in-situ Pt doping, the gas sensing materials show a greater response to low-concentration NH 3 gas, indicating that Pt doping can detect lower concentration of NH 3 gas. In addition, the amplitude change of the resistance value of 1.0 mol% Pt-doped WMSs for various concentrations of NH 3 gas are signi cantly higher than those of pure and other Pt-doped ones. Fig. 9 illustrates the responses of the WMSs with different Pt concentrations towards 10−1000 ppm NH 3 gas at the optimal operating temperatures. As observed in this gure, the responses of the four sensing materials are continuously enhanced with increasing NH 3 concentration. Especially, 1.0 mol% Pt-doped WMSs show higher response values to different concentrations of NH 3 gas, while 1.3 mol% Pt-doped WMSs show poor response to NH 3 gas, indicating that the appropriate concentration of Pt doping is helpful to improve the response of gas sensing material. The high concentration of Pt doping may inhibit the gas sensing reactions on the material surface to a certain extent, thus reducing the sensor response. As observed in this gure, 1.0 mol% Pt-doped WMSs does not reach the saturation stage for the chemisorption of NH 3 molecules when the NH 3 concentration is 1000 ppm, indicating that the as-prepared gas sensing material can supply more active sites for the chemisorption of gaseous molecules, and has a wider detection range for NH 3 gas. The responses of 1.0 mol% Pt-doped WMSs to the above different concentrations of NH 3 gas are 1.9, 3.7, 5.0, 8.9, 13.2, 16.8 and 28.2, respectively. Fig. 10(a) shows the reproducibility of 1.0 mol% Pt-doped WMSs to 1000 ppm NH 3 at 175 o C. It can be seen that after the introduction of 1000 ppm NH 3 gas in ve cycles, the Pt-doped WMSs exhibit approximately the same amplitude change of the resistance values, re ecting good reproducibility. In addition, when the NH 3 gas is released, the Pt-doped WMSs can completely recover the initial resistance value, showing excellent detection reversibility. The long-term stability illustrated in Fig. 10(b) demonstrates that the response values of the as-synthesized sensing material to 1000 ppm NH 3 gas has been uctuating around 28 in the whole test period of 30 days, indicating that the present gas sensor has a good long-term stability. Fig. 11 illustrates the selectivity of pure and 1.0 mol% Pt-doped WMSs towards NH 3 gas in different kinds of gas atmospheres, such as methanol, NO 2 , acetone, methylbenzene, SO 2 and methanal. As seen in Fig. 11, the response of Pt-doped WMSs towards 1000 ppm NH 3 at 175 o C is as high as 28.2, while the response values are relatively low and not higher than 3 to 1000 ppm methanol, 5 ppm NO 2 , 100 ppm methylbenzene, 100 ppm methanal, 100 ppm acetone and 100 ppm SO 2 , demonstrating that Pt doping can signi cantly improve the response of the sensing material to NH 3 gas, but it has no obviously enhanced effect on the responses of other interfering gases.

Gas sensing characteristics
In order to further clarify the superiority of the present sensor, the gas sensing properties of the as-synthesized WMSs was compared with those of other WO 3 -based gas sensors, mainly including the peak response, the response/recovery time and the optimal operating temperature, and the corresponding results are shown in Table 4. It can be seen that most of NH 3 gas sensors based on WO 3 sensing materials have the optimal operating temperature in a higher temperature range of 250−350 o C, while the WMSs-based gas sensors prepared in this study show lower operating temperature, especially the Pt-doped WMSs gas sensors. In addition, the as-synthesized Pt-doped WMSs sensing material has lower operating temperature and higher response to the same concentration of NH 3 gas compared with the Pt-decorated WO 3 thin lm in the literature. Although the as-prepared Pt-doped WMSs exhibit a lower response to a higher concentration of NH 3 gas compared with the WO 3 ower-like nanostructures, the operating temperature of the as-prepared sensing material is far lower than that of the reported sensing materials. Therefore, the as-synthesized Pt-doped WMSs show good comprehensive performance.

Gas sensing mechanism
The WMSs belong to a kind of surface-controlled gas sensing material. The gas sensing properties of the WMSs is greatly affected by the type and quantity of the chemisorbed oxygen on the material surface. The gas sensing reaction process of the WMSs mainly includes the following two stages. In the rst stage, oxygen molecules in air adsorb on the surface of the WMSs at a certain operating temperature, capturing electrons from the conduction band On the basis of the gas sensing performance results of the as-synthesized WMSs, the appropriate concentration of Pt doping can signi cantly enhance the gas sensing performance to NH 3 gas. The enhancement in gas sensing properties of the semiconductor materials induced by the noble metal nanoparticles is mainly ascribed to the electronic sensitization and chemical sensitization mechanism of the noble metal nanoparticles, which can play a catalytic role in the electron transport and transfer on the material surface, thereby improving the gas sensing properties to the reducing NH 3 gas [46]. In the process of electron sensitization, noble metal oxide nanoparticles can act as effective electron acceptors to capture free electrons from the surface of the WMSs and form electron depletion layer at the contact interface of the two materials. When the reducing NH 3 molecules contact with the noble metal nanoparticles, the noble metal oxides are reduced and the extra electrons are released and returned to the material surface, which is shown as the enhancement in the response. The chemical sensitization process is mainly due to the catalytic oxidation of noble metal nanoparticles on the material surface. The noble metal nanoparticles can supply active adsorption and reaction sites for NH 3 molecules on the WMSs surface. On the one hand, the introduction of noble metal element will accelerate the formation of chemisorbed oxygen on the material surface. On the other hand, it will make it easier for electrons to transfer from the active adsorption sites to the surfaces of the sensing materials, and react with the oxygen anion on the surfaces of the sensing materials, so as to enhance the gas sensing performance to NH 3 gas.

Conclusions
Pure and Pt-doped WO 3 ·0.33H 2 O microshuttles were prepared from a tungsten-containing metallurgical raw material by combination processes of NaOH leaching for obtaining test-grade leaching solution containing tungsten and onestep hydrothermal method for synthesizing sensing materials. The WMSs with the diameter of approximately 0.8 − 1.9 µm were self-assembled by many one-dimensional nanorods along the same direction. These nanorods with a single hexagonal crystal structure had different lengths, and their diameters were in the range of 17 − 62 nm. A certain concentration of Pt doping not only could improve the gas sensing performance towards NH 3 in terms of the response, reproducibility, and selectivity in the operating temperature range of 25 − 225 o C, but also could signi cantly reduce the detection limit of NH 3 Tables   Table 1 Chemical multi-element analysis results of the tungsten-containing metallurgical raw material.  Response T res /T rec (s) Figure 9 Responses of the WMSs to various concentrations of NH3 gas at optimal operating temperatures.  Selectivity of pure and 1.0 mol% Pt-doped WMSs to different types of gases at optimal operating temperatures.