Acetone is an important industrial raw material as well as an important biomarker in medical diagnosis [1]. Therefore, the detection of acetone has great significance for safety and health. The clinical data showed that the acetone was an important breath biomarker for noninvasive diagnosis of human type-1 diabetes and the concentration for exhaled diabetes exceeded 1.8 ppm, while that for healthy people was 0.3-0.9 ppm [2]. Therefore, high selectivity and low concentration (ppb level) acetone detection will be requested for application. The pure WO₃·H₂O rarely reported for acetone gas detection due to low acetone sensitivity and poor selectivity. There are several methods to improve sensing performances such as the design of micro-structures [3], the regulation of defects [4], the generation of p-n junction [5], and the modification of precious metals [6]. Among them, noble metal modification including electronic or/and chemical sensitization, has become a very important modification strategy [7]. However, the usage of noble metal increases sensor's cost. Therefore, the use of common metals instead of noble metals and the reduction of loading amount are beneficial for the application of gas sensor. Compared with single metal modification, bimetal modification has both electronic and chemical sensitization effects by selecting the optical metals [8]. Pt nanoparticles (NPs) are usually used as a chemical sensitizing material because of its excellent catalytic properties and O₂ spillover effects [9]. As a cheap material, Cu or Cu_xO also shows unique catalytic properties and electronic sensitization effect [10]. Therefore, it is possible to achieve both sensitization by preparing PtCu bimetallic nanocrystals, which can improve their gas sensitivity and selectivity simultaneously. In addition, ultra-low content PtCu modification hardly add to sensor prices.

In this work, the PtCu NPs and WO₃·H₂O hollow sphere (HS) were prepared first. And then, the PtCu/WO₃·H₂O HS was prepared by impregnation method. Their structure, morphology and surface chemical state were characterized by X-ray power diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS), and their sensing properties were measured. In addition, the optical loading amount of PtCu on nanocomposites was further studied. Based on gas sensitivity test results, the sensing mechanism was discussed and the utilization potentiality was illuminated by comparing with recent works.

The preparation schematic illustration of PtCu/WO₃·H₂O HS is displayed in Fig. S1 (Supporting information). The WO₃·H₂O HS was first obtained by a simple hydrothermal method. And then, the bimetallic PtCu NPs were synthesized by liquid phase chemical method [11]. Finally, different amount of PtCu NPs were modified on WO₃·H₂O HS by impregnation method.

The crystalline structure of WO₃·H₂O HS and 0.02% PtCu/ WO₃·H₂O HS was characterized by XRD. As shown in Fig. 1, the diffraction pattern of both the original WO₃·H₂O HS and 0.02% PtCu/WO₃·H₂O HS could be indexed to orthorhombic WO₃·H₂O (JCPDS card No. 43-0679), and no other impurities peaks could be found, proving it to be of high purity [12]. We cannot observe diffraction peaks of Cu or Pt for the 0.02% PtCu/WO₃·H₂O HS due to ultralow content of PtCu NPs in the nanocomposite.

SEM and TEM characterizations were used for studying the morphology of sensitive materials. Obviously, the uniform sphere self-assembled by nanosheets are observed in the Fig. 2a. In Fig. 2b, the spherical WO₃·H₂O has a bright and dark contrast between central and edges, which further proves the hollow spherical structure, and the result is well consistent with SEM image. HRTEM image of WO₃·H₂O HS was displayed in the inset of Fig. 2b. The interplanar distances are 0.26 nm and 0.35 nm, which correspond to the (200) and (111) crystal planes of orthorhombic WO₃·H₂O, respectively [13]. The TEM and HR-TEM images of PtCu nanoparticles are shown in Fig. 2c. Nanoparticles with evenly distribution can be observed in the TEM image. In addition, the (200) plane of Cu and (111) crystal plane of Pt can be noticed in the high resolution transmission electron microscopy (HR-TEM) image [14]. The characterization results show that bimetallic PtCu nanocrystals were prepared successfully. As shown in Figs. 2d and e, the hollow spherical structure still maintain after modification with bimetallic PtCu nanocrystals. The co-existence of (200) plane of Cu, (111) crystal plane of Pt and (111) plane of WO₃·H₂O in the Fig. 2f indicate that PtCu/WO₃·H₂O HS was prepared successfully [15].

The operating temperature has an important effect on the adsorption oxygens on the surface of sensitive materials, and thus affects their sensitivity [16]. In addition, the loading amounts of PtCu also has a great influence on the gas sensitivity. The low PtCu content leads to insufficient active sites on the gas sensitive material. However, the excessive amount of catalysts enlarge the depletion layer, which was formed around the catalyst, can be overlapped, leading to the deterioration of catalytic effect [17, 18]. Responses of sensors based on pure WO₃·H₂O HS and PtCu/WO₃·H₂O HS with different PtCu loading amounts toward 50 ppm acetone at different operating temperatures were displayed in Fig. 3a. Clearly, all sensors based on PtCu/ WO₃·H₂O HS with different PtCu loading amounts show higher response value than WO₃·H₂O HS. In addition, the response value of WO₃·H₂O HS sensor increased from 21.6–88.3 for 0.01% PtCu/WO₃·H₂O HS, 180.9 for 0.015% PtCu/WO₃·H₂O HS, 204.9 for 0.02% PtCu/WO₃·H₂O HS, 116.1 for 0.025% PtCu/WO₃·H₂O HS, and 81.8 for 0.03% PtCu/WO₃·H₂O HS, respectively. The maximum response value of 0.02% PtCu/WO₃·H₂O HS sensor (204.9) is 9.5 times than that of original WO₃·H₂O HS sensor (21.6). Fig. 3b shows that all sensors have fast response/recovery speed. The response/recovery times of 0.02% PtCu/WO₃·H₂O HS and the original WO₃·H₂O HS sensor are 3.4/7.5 s and 3.6/5.7 s. The slight increase of recovery time is due to the decrease in operating temperature from 300 ℃ to 280 ℃. Besides, the response values of 0.02% PtCu/WO₃·H₂O HS sensor remained almost same after 7 cycles of testing (Fig. 3c), which proves good stability and repeatability of the sensor. Selectivity is also an important indicator for measuring the anti-interference ability of sensors, and we tested responses of eight different volatile organic chemicals (VOCs) to evaluate the selectivity of sensors. As shown in Fig. 3d, the response value of 0.02% PtCu/WO₃·H₂O HS sensor to acetone is much higher than other gases, which proves that our sensor has excellent acetone selectivity. The dynamic response of the gas sensors based on WO₃·H₂O HS and 0.02% PtCu/WO₃·H₂O HS at 280 ℃ was shown in Fig. 3e. The dynamic responses show a gradual increase with increasing acetone concentration from 0.01 ppm to 100 ppm. The response of WO₃·H₂O HS sensor to different concentration of acetone was compared with 0.02% PtCu/WO₃·H₂O HS sensor. As shown in Fig. S2 (Supporting information), the response of WO₃·H₂O HS sensor was improved obviously by modification 0.02% PtCu nanocrystal. In addition, the limit of detection was improved from 0.5 ppm to 0.01 ppm, which is a significant performance improvement. As the breath biomarker of human type-1 diabetes, the ppb level detection to acetone is essential for noninvasive diagnosis. Therefore, the dynamic response and the relationship curves of 0.02% PtCu/WO₃·H₂O HS to low acetone concentration (10-1000 ppb) was further studied. A clear stepwise increasing trend from 10 ppb to 500 ppb was shown in the inset of Fig. 3e. For better application, the relationship between the reaction value and the acetone concentration was studied. As shown in Fig. 3f, the relationship between the S-1 and acetone concentrations (C_g) was fitted in the double logarithm coordinate, which showed a good linear relationship (0.01-100 ppm) [19]. Furthermore, we can speculate the adsorption oxygen species on the surface of the material through this linear relationship fitting equations:

where α is the prefactor and β is the surface species charge parameter. According to reports [20], when β is 0.5, the type of adsorbed oxygen ions on the surface of the material is O^2-, but when β is 1, the type of adsorbed oxygen ions change to O^- [20]. The β values of 0.020% PtCu/WO₃·H₂O HS is 0.62. Therefore, the main type of adsorbed oxygen ions may be O^2-. The good linear relationship between response and acetone concentration illustrated that 0.02% PtCu/WO₃·H₂O HS sensor could be used in both ppm and ppb levels acetone detection.

Generally, the sensing performance of metal oxide semiconductor (MOS) mainly depends on the chemical reaction between gas molecules and their surface adsorbed oxygen, which causes the change of charge depletion layer and further results in the change of resistance value [21]. The adsorbed oxygen (O^2-, O^- and O^2-) is generated by chemical adsorption of oxygen on the surface of MOS, and the main types of oxygen ions on the surface of the material are related to the operating temperature [22]. In this work, the optimal operating temperature of 0.02% PtCu/WO₃·H₂O HS sensor is 280 ℃. Combined with the calculation result of β value, we can infer that the form of oxygen ions on the materials surface may be O^2- and a small amount of O^- [19]. The W 4f and O 1s XPS of WO₃·H₂O HS and 0.02% PtCu/WO₃·H₂O HS are shown in Fig. S3 (Supporting information). Both the WO₃·H₂O HS and 0.02% PtCu/WO₃·H₂O HS have the similar W 4f XPS spectra. The peaks with binding energy at 530.6 eV and 532.7 eV are correspond to lattice oxygen (O_lat) and adsorbed oxygen (O_ads) [23], respectively. By calculation, the percentage of adsorbed oxygen in the original WO₃·H₂O HS was 27.6%, but this proportion rose to 31.9% after modification with PtCu NPs. The enhanced adsorbed oxygen percentage may be due to the spillover effect of O₂ on PtCu NPs. In addition, Pt and Cu can react with O₂ in the air to form p-type metal oxides [24, 25]. The pn junction will form between PtCu NPs and n-type WO₃·H₂O HS, which will hinder electron transport and further increase the resistance of the material [26-28]. Fig. 4 shows reaction mechanism of 0.02% PtCu/WO₃·H₂O HS after acetone gas are injected into the air chamber. Acetone molecules will react with O^- or O^2-, and the released electrons will back to the sensitive material. The reaction equation can be described as follows [29]:

which will reduce the electron depletion layer and reduce the resistance value. Besides, PtO_x and Cu_xO will be reduced to PtCu NPs, making the material resistance value is further decrease, and the synergy between Pt and Cu may be another important reason affecting the gas-sensing performance.

In conclusion, the PtCu NPs and WO₃·H₂O HS were first prepared by liquid phase chemistry and hydrothermal methods. Then, PtCu/WO₃·H₂O HS were prepared by impregnating method. The optimum loading amount of PtCu was as low as 0.02%. The gas sensitivity test results show that the 0.02% PtCu/WO₃·H₂O HS sensor has high sensitivity (204.9 tower 50 ppm CH₃COCH₃), good selectivity and fast response/recovery speeds (3.4 s/7.5 s), low detection limit (0.01 ppm) and good stability. The enhanced sensing performance was attributed to the synergy between PtCu and WO₃·H₂O, including the unique spillover effect of PtCu, the regulated depletion layer by p-type PtO_x and Cu_xO to n-type WO₃·H₂O, and their selective catalysis to acetone. This research provides an important basis for the acetone biomarker detection, and provides important reference for the study of ultra-low content modification of materials.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

We greatly appreciate the financial supports from the National Natural Science Foundation of China (Nos. 51702212, 51802195, 31701678, 61671284), Science and Technology Commission of Shanghai Municipality (Nos. 18511110600, 19ZR1435200), Innovation Program of Shanghai Municipal Education Commission (No. 2019-01-07-00-07-E00015), Program of Shanghai Academic Research Leader (No. 19XD1422900).

Supplementary material related to this article can be found, inthe online version, at doi:https://doi.org/10.1016/j.cclet.2020.04.033.