Ligand-free and size-controlled synthesis of oxygen vacancy-rich WO_3−x quantum dots for efficient room-temperature formaldehyde gas sensing

Abstract

A facile solvothermal decomposition route was developed for the size-adjustable synthesis of nearly monodispersed, monoclinic-phase WO_3−x QDs with sizes ranging from 1.3–4.5 nm. These QDs, with a high concentration of oxygen vacancies, have strong quantum confinement effects and can strongly absorb near-infrared (NIR) light. A solid–liquid–solid growth mechanism was proposed to explain the formation of the QDs. Experiments demonstrated that the rich oxygen vacancies and ultrathin grain size imparted these WO_3−x QDs with excellent room-temperature formaldehyde sensing properties.

---

Tungsten oxide nanostructures with high concentrations of oxygen vacancies have been widely used in photocatalysts, electrochromic windows, photoelectrochemical devices, surface-enhanced Raman spectroscopy (SERS), photodynamic therapy, gas sensors, etc.^1–11 As a transition metal oxide, tungsten oxides have different phases and compositions and can be divided into stoichiometric (WO₃ and WO₂) and non-stoichiometric (W₁₈O₄₉, W₅O₁₄, W₂₄O₆₈, W₂₀O₅₈, etc.) species.¹² Even the most common compound WO₃ also has at least four phases related to the formation temperature from low to high (−17–700 °C), in addition to triclinic, monoclinic, orthorhombic, and tetragonal phases.¹³ Among the tungsten oxides, the monoclinic phases of W₁₈O₄₉ and WO₃ are the most intensively studied.¹¹ For monoclinic W₁₈O₄₉, all reported morphologies have been one-dimensional nanostructures, such as nanorods, nanobelts, and nanowires, because of its obvious anisotropic structure.¹⁴ For monoclinic WO₃ (m-WO₃), also due to the anisotropy of its structural features, most of the products prepared have been one- and two-dimensional structures, such as nanowires, nanosheets, and nanobelts.¹⁵ However, zero-dimensional tungsten oxide nanostructures, i.e., quantum dots (QDs), have rarely been reported in the literature due to their difficult synthesis, although tetragonal-phase t-WO₃ QDs with diameters of 1.6 nm have been successfully synthesized using a surfactant-assisted method.^3a Compared with t-WO₃, which has higher crystal symmetry, the QD form of the more stable and most widely used form, m-WO₃ with lower crystal symmetry, has only been synthesized using zeolite (hard template) to the best of our knowledge.¹⁶ Due to their difficult separation, the templates are serious impediments to research on the inherent chemical and physical properties of tungsten oxide QDs and increase the difficulty of constructing various QD-based devices. Moreover, the size-controlled synthesis of tungsten oxide QDs remains a significant challenge.

Gas sensors play important roles in our daily lives and industrial production.^17–20 Among the many types of gas sensors, resistive gas sensors are the most attractive because of their simplicity of operation, lower manufacturing costs, and convenient assembly.²¹ Since Wickens and Hatmanwas created the first commercial gas sensor in 1964,²² scientists have been working hard to improve the sensitivity, stability, and response and recovery rates of these sensors. In general, the key unit of a resistive gas sensor is an active sensitive layer, the resistance of which must be highly sensitive to different environmental atmospheres. The sensitivity of the gas sensor and the grain size of the sensing layer are highly correlated: generally, the smaller the grain size, the higher the sensitivity.²³ The most common gas sensors based on metal oxides often require an effective activation temperature of 100–300 °C, which likely reduces the sensitivity of the sensor due to the crystal growth resulting from the high-temperature operation process. At present, one of the major challenges in the field of sensor research is how to develop an efficient gas sensor that can operate at room temperature.²⁴ Due to their ultrasmall particle size, extremely large surface-to-volume ratio, and large proportion of highly active surface atoms, QDs are promising candidates for preparation of the sensitive layer of room-temperature gas sensors.²⁵

Herein, we report the size-controlled synthesis of nearly monodispersed m-WO_3−x QDs with a high concentration of oxygen vacancies. No templates or ligands were used in the present method, which ensured the operability of the synthetic method and that the surfaces of the QDs were very clean. By adjusting the reaction temperature and corresponding precursor decomposition rate, the size of the m-WO_3−x QDs can be precisely controlled from 1.3 ± 0.2 nm to 4.5 ± 0.5 nm. A solid–liquid–solid (SLS) growth mechanism based on the decomposition of the precursor is proposed to explain the formation of the QDs. The present work provides an effective route for the synthesis of oxygen vacancy-rich WO_3−x QDs with lower crystal symmetry. More importantly, the as-synthesized WO_3−x QDs display high formaldehyde sensitivity with a detection limit of 1.5 ppm at room temperature, which is 22-foldmore sensitive than commercial tungsten oxide. Furthermore, m-WO_3−x QDs with smaller crystal sizes and higher oxygen-vacancy concentrations can achieve much greater gas sensitivity.

The m-WO_3−x QDs used herein were prepared by simple solvothermal decomposition. In brief, moderate amounts of anhydrous tungsten chloride (WCl₆) and hydrazine hydrate (N₂H₄·H₂O, hazardous chemical) were mixed in anhydrous ethanol in a nitrogen-filled glove box, and the mixed solution was transferred to a Teflon-lined autoclave. The autoclave was tightly sealed and placed in an electric heating furnace at 180 °C for 11 h. Finally, the formed blue colloidal products were collected by high-speed centrifugation (for detailed synthetic procedures, please see the experimental section in ESI†).

Firstly, the crystal structure and phase purity of the prepared blue sample were detected by powder X-ray diffraction (XRD). As shown in Fig. S1,† all the diffraction peaks of the sample can be easily indexed as monoclinic phase WO_3−x (PCPDF no. 89-4476), indicating the high purity of the prepared sample. Notably, all of the diffraction peaks of the sample were obviously widened, which strongly suggests a very small crystal grain size. According to the full width at half maximum (FWHM) of the (020) diffraction peak, the average grain size of m-WO_3−x was only approximately 2.3 nm. In addition, the Fourier transform infrared (FTIR) spectrum revealed that the surface of the sample was clean (Fig. S2†). Taken together, the XRD and FTIR results demonstrate the synthesis of the m-WO_3−x colloidal nanocrystals by the present method.

The prepared sample could be dispersed in water or ethanol to reconstruct a stable colloidal solution that presented evident Tyndall scattering (Fig. 1a). The morphology of the sample was observed by transmission electron microscopy (TEM). As shown in Fig. 1b, the sample was composed of a large quantity of monodispersed m-WO_3−x QDs with uniform diameters of approximately 2.2 nm. The sharp diffraction rings obtained from selected-area electron diffraction (SAED) reveal the m-WO_3−x QDs were highly crystalline (inset in Fig. 1b). Furthermore, high-resolution transmission electron microscopy (HRTEM) images further demonstrated that the m-WO₃ QDs were highly crystalline and that there was no amorphous layer around the QD surface (Fig. 1c and S3†). The lattice fringes with a spacing of 0.385 nm can be indexed as the (020) crystal plane of m-WO₃ (Fig. 1d). The Raman spectrum shows that the as-synthesized sample was typical m-WO₃ (Fig. 1e). The sample components were further confirmed by the energy-dispersive X-ray (EDS) spectrum, which clearly shows only the production of tungsten and oxygen signals by the sample (Fig. 1f). In addition, we randomly measured the sizes of 200 particles and obtained a size distribution of the QDs of 2.2 ± 0.3 nm (Fig. 1g), which highly agrees with the XRD characterizations.

Fig. 1 (a) Tyndall scattering obtained from the aqueous solution of 2.2 ± 0.3 nm m-WO_3−x QDs. (b–g) TEM images, HRTEM image, Raman spectrum, EDS spectrum, and size distribution of the prepared 2.2 ± 0.3 nm m-WO_3−x QDs. (h) TEM image, HRTEM image (inset), and size distribution (inset) of the prepared 1.3 ± 0.2 nm m-WO_3−x QDs. (i) TEM image, HRTEM image (inset), and size distribution (inset) of the prepared 4.5 ± 0.5 nm m-WO_3−x QDs.

Importantly, the size of the m-WO_3−x QDs can be precisely controlled by adjusting the reaction temperature. As shown in Fig. 1h, m-WO₃ QDs with sizes of 1.3 ± 0.2 nm were synthesized when the reaction temperature was reduced to 160 °C, whereas 4.5 ± 0.5 nm QDs were obtained when the reaction temperature was increased to 200 °C (Fig. 1i). Importantly, the 1.3 ± 0.2 nm QDs obtained at the lower temperature were still highly crystalline, which was confirmed by their HRTEM image (inset in Fig. 1h).

As shown in the UV/Vis absorption spectrum, the 2.2 ± 0.3 nm WO_3−x QDs exhibited interesting optical absorption properties (Fig. 2). Their absorption band (3.2 eV) was obviously blue shifted compared with the typical absorption band of bulk WO₃ (2.6–2.7 eV). Considering that the reported exciton Bohr radius of tungsten oxide is 3 nm,²⁶ this obvious band-edge blue shift should be a reflection of its strong quantum confinement effect. More interestingly, in addition to their band-edge absorption, these WO_3−x QDs can also strongly absorb visible and near infrared (NIR) light. Especially in the NIR region, a strong upward tilt of the absorption curve extended to the limit of measurement, 800 nm, which can be considered to be closely related with the free electrons induced by oxygen vacancies,^1–5 suggesting the existence of a large number of oxygen vacancies in the WO_3−x QDs.

Fig. 2 UV/Vis absorption spectrum of the 2.2 ± 0.3 nm WO_3−x QDs.

Subsequently, we systematically investigated the formation mechanism of the m-WO_3−x QDs by analyzing the structure and morphology of the intermediates at different reaction steps. The 2.2 ± 0.3 nm m-WO₃ QDs are cited to illustrate this mechanism. After the initial 1 h solvothermal reaction, many white products were obtained that had deposited at the bottom of the autoclave. The TEM image revealed that the white products were composed of approximately spherical solid particles with an average diameter of 110 nm (Fig. 3a). The XRD pattern (Fig. 3e) of the initial white product can be indexed as ammonium tungsten oxide hydrate (5(NH₄)₂O·12WO₃·11H₂O, PCPDF no. 18-0127, ATOH for short). In the presence of hydrazine hydrate, WCl₆ hydrolysis produces the ATOH complex. As the reaction time increased to 4 h, the original solid particles became slightly fluffy, and many ultrathin nanocrystals appeared (Fig. 3b). Compared with the XRD pattern of ATOH, XRD diffraction peaks belonging to tungsten oxide appeared, which can be obviously observed in the angle range of 20–35 (marked with a dashed box in Fig. 3f). At the same time, the strength of the diffraction peaks belonging to ATOH became relatively weak. Therefore, these tungsten oxide QDs were produced from the decomposition of ATOH. As the reaction time increased to 8 hours, more QDs were obtained, and only a small amount of ATOH had not been decomposed (Fig. 3c). The XRD pattern also proved that the main components of these samples were monoclinic-phase tungsten oxide (Fig. 3g). Finally, as the reaction time extended to 11 h, all the ATOH particles were completely decomposed into m-WO₃ QDs (Fig. 3d), and the XRD pattern of the sample reflected that of pure m-WO₃ (Fig. 3h). By analysing these experimental data, the time-based XRD patterns were in agreement with the morphological evolution shown in Fig. 3a–d. Taken together, the TEM and XRD results demonstrate that the m-WO₃ QDs came from the decomposition of the white ATOH particles under the present solvothermal conditions. These ATOH particles provide not only a reaction precursor for this decomposition reaction but also a large quantity of nucleation sites for m-WO₃ nucleation (Fig. 3i). Although there was no protection by additional added surfactants or hard templates during the QD formation process, the precursor itself acted as a self-template in the formation of the QDs. On the surface of the ATOH precursor, specifically in its interior, steric hindrance greatly limited the growth of the m-WO₃ QDs. In contrast, holding the other reaction conditions constant, comparative experiments showed that only urchin-like W₁₈O₄₉ nanowire aggregates were prepared by directly hydrolysing WCl₆ in ethanol without N₂H₄·H₂O (Fig. S4†), and no ATOH was found in the reaction process. These results suggest that the present ATOH precursor decomposition route is an effective strategy for the synthesis of high quality m-WO_3−x QDs and can also be used to prepare other ultrathin transition metal oxide nanocrystals.

Fig. 3 (a–d) TEM images of the reaction intermediates from 1–11 h. (e–h) XRD patterns of the four intermediates. (i) Schematic diagram of the formation of the WO_3−x QDs.

Generally, the stoichiometric m-WO₃ is bright yellow; however, the present m-WO_3−x QDs were a vivid blue colour (left, Fig. 4a), which strongly suggests a high concentration of oxygen vacancies contained in the m-WO_3−x QDs, as in the blue TiO_2−x and MoO_3−x nanocrystals reported by other groups.²⁷ To confirm this hypothesis, m-WO₃ QDs with a low concentration of oxygen vacancies were prepared by oxidizing the blue m-WO_3−x QDs (2.2 ± 0.3 nm) with a dilute aqueous H₂O₂ solution (1 M) at room temperature (the oxidized sample was named W–H₂O₂-2.2). After 30 minutes of oxidation, the colour of the sample gradually changed from the original blue to yellow, as shown in Fig. 4a (right). Although the XRD patterns (Fig. 4b) and TEM images (Fig. S5†) showed no obvious change in the crystalline phase or morphology before and after this oxidation process, the UV/Vis absorbance spectra were significantly different. As shown in Fig. 4c, the blue m-WO_3−x QDs displayed strong absorption (400–800 nm) from the visible to near-infrared (NIR) regions, whereas no NIR absorption was detected for the yellow m-WO₃ QDs obtained by H₂O₂ oxidation. Furthermore, the X-ray photoelectron spectrum (XPS) revealed a high concentration of W⁵⁺ ions contained in the blue m-WO_3−x QDs (Fig. 4d). In comparison, the XPS characterization revealed that only W⁶⁺ ions were present in the yellow m-WO₃ QDs, which confirmed that the large quantity of oxygen vacancies contained in the blue m-WO_3−x QDs were removed by H₂O₂ oxidation. Based on the comprehensive analysis of the above-mentioned XRD, TEM, UV/Vis, and XPS results, this change in colour was indeed caused by the removal of oxygen vacancies, which, in turn, proved that a number of oxygen vacancies was contained in the as-prepared blue m-WO_3−x QDs. Due to the generation of a large number of active sites resulting from the oxygen vacancies and the shortening of the charge transport distance resulting from the QDs, the ultrafine size and large number of crystal defects are very advantageous to the improvement of the gas sensing properties of materials.²⁸ Therefore, the oxygen vacancy-rich m-WO_3−x QDs should possess excellent gas detection performance.

Fig. 4 (a) Colour conversion, (b) XRD patterns, (c) UV/Vis absorption spectra, and (d) XPS spectra of the 2.2 ± 0.3 nm m-WO_3−x QDs before (red line) and after oxidation (black line) by H₂O₂ (1 M).

To confirm this hypothesis, we first investigated the ability of the 2.2 ± 0.3 nm m-WO_3−x QDs to detect formaldehyde (HCHO, one of the most common gaseous indoor pollutants) at room temperature (25 °C). The detection sensitivity (S) can be defined as follows:

S = R₀/R

where R₀ is the resistance of the m-WO_3−x QDs in the atmosphere and R is the resistance of the QDs in a gas mixture of HCHO–air (relative humidity of approximately 20%). These m-WO_3−x QDs had high sensitivity for the HCHO gas. When the HCHO concentration was 100, 200, 500, and 1000 ppm, the corresponding sensitivity was 1.6, 2.3, 3.4, and 4.6, respectively (Fig. 5a). The lowest detectable limit was 5 ppm for the 2.2 ± 0.3 nm m-WO_3−x QDs. Furthermore, the QD-based sensor displayed extremely rapid response and recovery processes. Only 4 seconds were required to reach the maximum response of 90%, and 3 seconds of recovery time were observed when HCHO was removed by blasting. In evaluating the performance of gas sensors, stability is another important indicator of their practical applications. The results of stability experiments showed that the sensitivity of these QDs for HCHO gas (100 ppm) had a good repeatability (Fig. 5b).

Fig. 5 (a) Sensitivity responses of the 2.2 ± 0.3 nm sized m-WO_3−x QDs to various concentrations of HCHO at room temperature. (b) Stability test of the 2.2 ± 0.3 nm sized m-WO_3−x QDs to 100 ppm HCHO. (c) Detection limits of the samples with different sizes and oxygen vacancy concentrations.

Furthermore, the detection limit of formaldehyde was closely connected with the size of the QDs. As shown in Fig. 5c, when the 1.3 ± 0.2 nm m-WO_3−x QDs were used as the sensitive layer, the lowest detectable limit of HCHO was further improved to 1.5 ppm. On the contrary, when the 4.5 ± 0.5 nm m-WO_3−x QDs were used as the sensitive layer, the lowest detectable limit of HCHO was reduced to 12 ppm. At the same time, oxygen vacancies had a positive effect on the improvement of the HCHO gas sensing performance. Using sample W–H₂O₂-2.2 as the sensitive layer, the detection limit for HCHO decreased from the original 5 ppm to 21 ppm. To better understand the high sensitivity of these m-WO₃ QDs, the HCHO gas-sensing properties of commercial m-WO₃ nanoparticles (50–100 nm in diameter, named C-WO₃, Fig. S6†) were also investigated for comparison. The detectable limit of HCHO was 33 ppm (Fig. 5c) for the commercial product. Obviously, the HCHO sensitivity of the 1.3 ± 0.2 nm WO_3−x QDs was 22-fold higher than that of C-WO₃.

Based on the above experimental results, we believe that the high sensitivity of the prepared QDs can be reasonably attributed to two important factors. One factor is the very small grain size of the QDs. Early studies suggested that gas-sensing properties would be greatly improved when the size of the semiconductor was below the double thickness of the corresponding space-charge layer (SCL),²⁹ i.e., the smaller the particle size, the better the gas sensing performance. Herein, the size of the m-WO_3−x QDs used was only 1.3–4.5 nm, which is close to the size of the unit cell of m-WO₃. As a result, the as-synthesized m-WO_3−x QDs displayed very high sensitivity to HCHO gas. Another important factor is the rich oxygen vacancies contained in the QDs. The presence of these oxygen defects provides more active sites for gas adsorption.³⁰

Conclusions

In summary, a facile solvothermal decomposition route was developed for the size-adjustable synthesis of defect-rich m-WO_3−x QDs. By adjusting the reaction temperature, the size distribution of the m-WO_3−x QDs could be precisely controlled from 1.3 to 4.5 nm. A precursor confinement mechanism was proposed for the formation of the QDs. Due to the extremely small size and rich oxygen vacancy defects, the m-WO₃ QDs possessed high gas sensitivity and stability as an HCHO probe, and a lowest detectable limit of 1.5 ppm was obtained at room temperature. In addition to their use as an efficient gas-sensing material, these m-WO₃ QDs can be used in the fabrication of electrochromic devices and in photocatalysis.

Acknowledgements

This work received financial support from the Dean Fund of the Chinese Academy of Inspection and Quarantine (2016JK025), the Science Foundation of AQSIQ (2015IK308), and the National Natural Science Foundation of China (51472226).

Notes and references

1. (a) Z. F. Huang, J. Song, L. Pan, X. Zhang, L. Wang and J. J. Zou, Adv. Mater., 2015, 27, 5309–5327; (b) J. J. Song, Z. F. Huang, L. Pan, J. J. Zou, X. W Zhang and L. Wang, ACS Catal., 2015, 5, 6594–6599.
2. J. Yan, T. Wang, G. Wu, W. Dai, N. Guan, L. Li and J. Gong, Adv. Mater., 2015, 27, 1580–1586.
3. (a) S. Cong, Y. Tian, Q. Li, Z. Zhao and F. Geng, Adv. Mater., 2014, 26, 4260–4267; (b) S. Cong, Y. Yuan, Z. Chen, J. Hou, M. Yang, Y. Su, Y. Zhang, L. Li, Q. Li, F. Geng and Z. Zhao, Nat. Commun., 2015, 6, 7800; (c) Y. Y. Tian, S. Cong, W. M. Su, H. Y. Chen, Q. W. Li, F. X. Geng and Z. G. Zhao, Nano Lett., 2014, 14, 2150–2156.
4. K. Manthiram and A. P. Alivisatos, J. Am. Chem. Soc., 2012, 134, 3995–3998.
5. J. W. Liu, J. Zheng, J. L. Wang, J. Xu, H. H. Li and S. H. Yu, Nano Lett., 2013, 13, 3589–3593.
6. (a) W. Cheng, F. Rechbergera and M. Niederberger, Nanoscale, 2016, 8, 14074–14077; (b) H. Bai, W. C. Yi, J. Y. Liu, Q. Lv, Q. Zhang, Q. Ma, H. F. Yang and G. C. Xi, Nanoscale, 2016, 8, 13545–13551; (c) I. G. Gonzalez-Martinez, A. Bachmatiuk, V. Bezugly, J. Kunstmann, T. Gemming, Z. Liu, G. Cuniberti and M. H. Rümmeli, Nanoscale, 2016, 8, 11340–11362; (d) G. L. Li, S. H. Zhang, C. S. Guo and S. Q. Liu, Nanoscale, 2016, 8, 9861–9868.
7. G. Xi, S. Ouyang, P. Li, J. Ye, Q. Ma, N. Su, H. Bai and C. Wang, Angew. Chem., Int. Ed., 2012, 51, 2395–2399.
8. B. Ni and X. Wang, Chem. Sci., 2016, 7, 3978–3991.
9. P. He, B. Xu, X. Xu, L. Song and X. Wang, Chem. Sci., 2016, 7, 1011–1015.
10. P. Kalluru, R. Vankayala, C.-S. Chiang and K. C. Hwang, Angew. Chem., Int. Ed., 2013, 52, 12332–12336.
11. W. Cheng, Y. Ju, P. Payamyar, D. Primc, J. Rao, C. Willa, D. Koziej and M. Niederberger, Angew. Chem., Int. Ed., 2014, 54, 340–344.
12. H. Zheng, J. Z. Ou, M. S. Strano, R. B. Kaner, A. Mitchell and K. Kalantar-zadeh, Adv. Funct. Mater., 2011, 21, 2175–2196.
13. Z. G. Zhao and M. Miyauchi, Angew. Chem., Int. Ed., 2008, 120, 7159–7163.
14. (a) K. Lee, W. S. Seo and J. T. Park, J. Am. Chem. Soc., 2003, 125, 3408–3409; (b) J. Polleux, N. Pinna, M. Antonietti and M. Niederberger, J. Am. Chem. Soc., 2005, 127, 15595–15601; (c) J. Polleux, A. Gurlo, N. Barsan, U. Weimar, M. Antonietti and M. Niederberger, Angew. Chem., Int. Ed., 2006, 118, 267–271; (d) A. Yella, M. N. Tahir, S. Meuer, R. Zentel, R. d. Berger, M. Panthöfer and W. Tremel, J. Am. Chem. Soc., 2009, 131, 17566–17575; (e) B. Moshofsky and T. Mokari, Chem. Mater., 2013, 25, 1384–1391; (f) J. Liu, O. Margeat, W. Dachraoui, X. Liu, M. Fahlman and J. Ackermann, Adv. Funct. Mater., 2014, 24, 6029–6037.
15. (a) D. Chen and J. Ye, Adv. Funct. Mater., 2008, 18, 1922–1928; (b) M. Shibuya and M. Miyauchi, Adv. Mater., 2009, 21, 1373–1376; (c) X. Chen, Y. Zhou, Q. Liu, Z. Li, J. Liu and Z. Zou, ACS Appl. Mater. Interfaces, 2012, 4, 3372–3377; (d) G. Xi, J. Ye, Q. Ma, N. Su, H. Bai and C. Wang, J. Am. Chem. Soc., 2012, 134, 6508–6511; (e) W. Li, P. Da, Y. Zhang, Y. Wang, X. Lin, X. Gong and G. Zheng, ACS Nano, 2014, 8, 11770–11777; (f) P. Yang, P. Sun, Z. Chai, L. Huang, X. Cai, S. Tan, J. Song and W. Mai, Angew. Chem., Int. Ed., 2014, 53, 11935–11939; (g) Z. He, Q. Liu, H. Hou, F. Gao, B. Tang and W. Yang, ACS Appl. Mater. Interfaces, 2015, 7, 10878–10885.
16. (a) H. Watanabe, K. Fujikata, Y. Oaki and H. Imai, Chem. Commun., 2013, 49, 8477–8479; (b) M. Epifani, E. Comini, R. Díaz, T. Andreu, A. Genç, J. Arbiol, P. Siciliano, G. Faglia and J. R. Morante, ACS Appl. Mater. Interfaces, 2014, 6, 16808–16816.
17. (a) J. Zhang, X. Liu, G. Neri and N. Pinna, Adv. Mater., 2015, 28, 795–831; (b) T. Wagner, S. Haffer, C. Weinberger, D. Klaus and M. Tiemann, Chem. Soc. Rev., 2013, 42, 4036–4053.
18. (a) S. Li, N. Wang, Y. Yue, G. Wang, Z. Zu and Y. Zhang, Chem. Sci., 2015, 6, 2495–2500; (b) Y. Liu, S. Antwi-Boampong, J. J. BelBruno, M. A. Crane and S. E. Tanski, Nicotine Tob. Res., 2013, 15, 1511–1518; (c) M. Righettoni, A. Amann and S. E. Pratsinis, Mater. Today, 2015, 18, 163–171.
19. A. Mallick, B. Garai, M. A. Addicoat, P. S. Petkov, T. Heine and R. Banerjee, Chem. Sci., 2015, 6, 1420–1425.
20. S. Zhang, K. Dokko and M. Watanabe, Chem. Sci., 2015, 6, 3684–3691.
21. (a) N. Barsan, D. Koziej and U. Weimar, Sens. Actuators, B, 2007, 121, 18–35; (b) M. E. Franke, T. J. Koplin and U. Simon, Small, 2006, 2, 36–50.
22. K. Ihokura and J. Watson, The Stannic Oxide Gas Sensor Principles and Applications, CRC Press, Boca Raton, FL, USA, 1994.
23. (a) Y. Hu, J. Zhou, P.-H. Yeh, Z. Li, T.-Y. Wei and Z. L. Wang, Adv. Mater., 2010, 22, 3327–3332; (b) X. Pan, X. Liu, A. Bermak and Z. Fan, ACS Nano, 2013, 7, 9318–9324.
24. (a) H. Li, Z. Yin, Q. He, H. Li, X. Huang, G. Lu, D. W. H. Fam, A. I. Y. Tok, Q. Zhang and H. Zhang, Small, 2011, 8, 63–67; (b) P. A. Russo, N. Donato, S. G. Leonardi, S. Baek, D. E. Conte, G. Neri and N. Pinna, Angew. Chem., Int. Ed., 2012, 51, 11053–11057; (c) D.-J. Yang, I. Kamienchick, D. Y. Youn, A. Rothschild and I.-D. Kim, Adv. Funct. Mater., 2010, 20, 4258–4264; (d) K. Lee, R. Gatensby, N. McEvoy, T. Hallam and G. S. Duesberg, Adv. Mater, 2013, 25, 6699–6702.
25. (a) I. L. Medintz, H. T. Uyeda, E. R. Goldman and H. Mattoussi, Nat. Mater., 2005, 4, 435–446; (b) S. Zhu, Q. Meng, L. Wang, J. Zhang, Y. Song, H. Jin, K. Zhang, H. Sun, H. Wang and B. Yang, Angew. Chem., Int. Ed., 2013, 125, 4045–4049.
26. R. A. May, L. Kondrachova, B. P. Hahn and K. J. Stevenson, J. Phys. Chem. C, 2007, 111, 18251–18257.
27. (a) T. R. Gordon, M. Cargnello, T. Paik, F. Mangolini, R. T. Weber, P. Fornasiero and C. B. Murray, J. Am. Chem. Soc., 2012, 134, 6751–6761; (b) Q. Huang, S. Hu, J. Zhuang and X. Wang, Chem.–Eur. J., 2012, 18, 15283–15287.
28. (a) H. Liu, M. Li, O. Voznyy, L. Hu, Q. Fu, D. Zhou, Z. Xia, E. H. Sargent and J. Tang, Adv. Mater., 2014, 26, 2718–2724; (b) Y. Liu, X. Dong and P. Chen, Chem. Soc. Rev., 2012, 41, 2283–2307.
29. F. Gyger, M. Hübner, C. Feldmann, N. Barsan and U. Weimar, Chem. Mater., 2010, 22, 4821–4827.
30. X. Lai, J. Li, B. A. Korgel, Z. Dong, Z. Li, F. Su, J. Du and D. Wang, Angew. Chem., Int. Ed., 2011, 50, 2738–2741.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedure, XRD pattern, FTIR spectrum, and HRTEM image of the WO_3−x QDs, SEM, TEM, HRTEM images, and XRD pattern of the W₁₈O₄₉ nanowires, commercial WO₃ powder, structure of the WO_3−x QD-based gas sensor, structure of the gas-sensing measurement system. See DOI: 10.1039/c6ra20531c

---