English

• Nitrogen dioxide (NO₂) is a significant toxic air pollutant emitted from combustion processes and vehicle exhaust. The emission of NO₂ can lead to ozone formation and acid rain [1]. Prolonged exposure to NO₂, even at levels below a few parts per billion (ppb), can induce various respiratory issues, including irritation, emphysema, and chronic bronchitis [2]. Consequently, the European Union set a NO₂ limit of 200 µg/m³ to protect public health and the environment [3]. The negative impacts of NO₂ highlight the significance of gas sensing technology, especially the use of metal oxide semiconductor (MOS)-based chemiresistive gas sensors such as TiO₂, SnO₂, and ZnO [4, 5]. However, these gas sensors usually require high operating temperatures or the loading of precious metals such as Pt, Au, and Pd to reduce the operating temperature [6, 7]. The high cost of energy and precious metals has limited the economic advantage of these sensors, which has hindered their wider application as portable and wearable devices in environmental pollutant monitoring. To address these challenges, the construction of heterojunctions by combining MOS with other functional materials has been demonstrated as an effective strategy to gain high sensing performance at room temperature (RT) [8-10].

  In recent years, lead halide perovskites have attracted significant attention in fields such as photovoltaics and light-emitting diodes owing to their extensive chemical tunability and optoelectronic properties [11, 12]. In optical applications, defects in perovskites are generally viewed as detrimental due to their ability to capture carriers and consequently reduce device performance [13]. Conversely, in gas sensing applications, these defects may be regarded as an advantage [14]. Defects in perovskites can act as electron traps and active sites, enhancing the interaction between the material and gas molecules. Thus, the presence of defects improves the sensitivity and selectivity of sensors [15, 16]. Compared to the toxic lead halide perovskite materials [17], lead-free perovskite materials retain the exceptional chemical tunability of perovskite materials while also offering advantages such as environmental friendliness, stability, and high surface-active site content. Although some lead-free perovskite materials have applied for NO₂ sensing [18-20], the gas-sensing-related reactions are sluggish under room-environmental light (REL) at RT. Therefore, the development of sensitivity NO₂ sensors with rapid response/recovery ability at REL and RT conditions is still largely desirable.

  In this work, with the goal of designing a sensitive and long-term stable NO₂ chemiresistive gas sensor at RT, a heterojunction was prepared by combining a lead-free perovskite material, Cs₂AgInCl₆ with chlorine vacancies (V_Cl-CAIC), with TiO₂ nanotube arrays (NTs). The 1D nanotube array structure serves as an excellent platform, providing a highly efficient electron transport pathway that enhances the performance of sensor sensing layers. The NO₂ detection performance of the V_Cl-CAIC/TiO₂ heterojunction was investigated in detail. The defects in the perovskite act as electron traps and active sites, enhancing the interaction between the sensor material and gas molecules. The sensor exhibits remarkable sensitivity, selectivity, repeatability, and stability. These findings advance the field of lead-free perovskite-based gas sensors and also highlight the imperative environmental and health challenges posed by NO₂ emissions.

  Fig. 1a illustrates the steps used for the preparation of gas sensing materials. Scanning electron microscopy (SEM) images in Figs. 1b and c show TiO₂ NTs prepared by electrochemical anodization have a diameter of 150nm and a length of ~9 µm. These NTs are uniformly arranged. Surface charge analysis reveals that V_Cl-CAIC and TiO₂ NTs possess opposite charges (Fig. S2 in Supporting information). As demonstrated in Fig. 1d, the dip-coating method was used to successfully modify TiO₂ NTs with V_Cl-CAIC nanocrystals. With increasing CAIC solution concentration (X=3, 6, 10, and 20mmol/L), the loading of V_Cl-CAICx increases (Fig. S3 in Supporting information). The V_Cl-CAIC-decorated TiO₂ NTs prepared with a concentration of 10mmol/L (denoted as V_Cl-CT-10, Fig. 1d) show a highly dispersed and uniform distribution of CAIC nanocrystals on the surface and walls of the NTs. Moreover, EDX elemental mapping (Fig. S4 in Supporting information) shows the uniform distribution of Cs, Ag, In, and Cl elements on the TiO₂ NTs. With a further increase in the CAIC precursor concentration, the CAIC particles tend to aggregate and block the openings of the nanotubes (Fig. S3), which may impede gas permeation. Figs. 1e and f show the transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) images of V_Cl-CT-10, respectively. Besides the well-distinguished (220) plane of CAIC and the (101) plane of TiO₂ [21, 22], a certain amount of lattice disorders and vacancy defects were found for the (220) lattice plane of CAIC (Fig. 1g). The X-ray diffraction (XRD) patterns (Fig. 2a and Fig. S5 in Supporting information) show diffraction peaks ascribed to anatase crystalline TiO₂ (JCPDS No. 22-1012) [22] and Ti substrate. The peaks at 24.1° and 34.3° are attributed to cubic Cs₂AgInCl₆ (JCPDS No. 154-6187) [21]. Notably, with increasing CAIC content, the intensities of the characteristic peaks of V_Cl-CAIC at 24.1° and 34.3° are significantly enhanced (Fig. S6 in Supporting information).

  Figure 1

  

  Figure 1.  (a) Schematic diagram of gas sensing device synthesis. (b, c) SEM images of TiO₂ NTs. (d) SEM, (e) TEM, and (f) HRTEM with fast Fourier transformed (FFT) analysis of V_Cl-CT-10. (g) HRTEM image of defective V_Cl-CAIC.

  DownLoad: Full-Size Img PowerPoint

  Figure 2

  

  Figure 2.  (a) XRD patterns of samples. (b) EPR spectra of V_Cl-CT-10 and CT-10. (c) In 3d XPS spectra of V_Cl-CAIC and V_Cl-CT-10.

  DownLoad: Full-Size Img PowerPoint

  Fig. 2b shows the electron paramagnetic resonance (EPR) signals of V_Cl-CT-10 and CT-10. Clearly, V_Cl-CT-10 shows a more pronounced EPR signal (g=2.01), which indicates the enrichment of unpaired electrons in V_Cl-CT-10. These electrons originate from those trapped around the Cl^- vacancies [20, 23-25]. The elemental composition of the V_Cl-CAIC on the TiO₂ NTs was determined through energy-dispersive X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS) analyses (Tables S1 and S2 in Supporting information). According to XPS, V_Cl-CAIC has a Cs: Ag: In: Cl atomic ratio of 2.01:1:1.01:5.41 (Fig. S7 in Supporting information). The slight deviation from the Cs₂AgInCl₆ stoichiometry can be ascribed to the formation of chlorine vacancies. This is further confirmed by the detailed In 3d XPS spectra shown in Fig. 2c. The two main peaks located at 451.2eV and 444.6eV correspond to In 3d_3/2 and In 3d_5/2 of the In-Cl bond in the (InCl₆)³⁻ units in V_Cl-CAIC [26, 27]. In the In 3d spectrum of V_Cl-CT-10, these peaks exhibit a positive shift to 451.5eV and 444.9eV. The two minor peaks located at 451.6eV and 444.0eV are associated with a lower oxidation state +(3−x), which has been reported to be related to the presence of a halide vacancy [23, 26]. Moreover, compared to pure V_Cl-CAIC, the Cs 3d, Ag 3d, and Cl 2p of XPS signals of V_Cl-CT-10 are also shifted towards higher binding energy values (Fig. S7). These results indicate the possible presence of a strong built-in electric field at the interface between V_Cl-CAIC and TiO₂ NTs, which would facilitate interfacial electron transportation [23].

  To achieve the optimal NO₂ sensing performance, the influence of V_Cl-CAIC loading amount on the TiO₂ NTs was investigated at RT (Fig. S8 in Supporting information). As plotted in Fig. 3a, the as-prepared V_Cl-CT-x samples always exhibit a significantly enhanced response (R_g/R_a) toward NO₂ compared with TiO₂ NTs, and the highest response is achieved on V_Cl-CT-10. This sensing response enhancement can be attributed to the heterojunction formed between TiO₂ and V_Cl-CAIC [28, 29]. As shown in Fig. 3b, the response of V_Cl-CT-10 to 1ppm NO₂ reaches 7.26, with response and recovery times of 38s and 59s, respectively. Compared with the other V_Cl-CT-x samples, V_Cl-CT-10 possesses the highest sensitivity and rapid response/recovery speed (Fig. S9 in Supporting information). Therefore, subsequent testing was performed by using V_Cl-CT-10 sample as the sensing chip. To elucidate the origin of the good sensing performance of V_Cl-CT-10, CAIC-10 was directly grown on TiO₂ NTs (CT-10, Fig. S3d). The response of V_Cl-CT-10 is 2.07 times that of CT-10 for detecting 1ppm NO₂ (Fig. S10 in Supporting information). This is because the defective sites on the V_Cl-CT serve as electronic traps and active reaction sites. In addition, these defective sites can lower the energy barrier for charge transfer, thus accelerating the charge transfer with NO₂ molecules [3, 4].

  Figure 3

  

  Figure 3.  (a) Gas responses of TiO₂ NTs and V_Cl-CT-x. (b) Response and recovery times of V_Cl-CT-10 toward 1ppm of NO₂. (c) Dynamic responses and (d) linear relationship curve of V_Cl-CT-10 toward 0.02–100ppm of NO₂. (e) Repeatability of V_Cl-CT-10 toward 1ppm of NO₂. (f) Radar graph analysis of key sensing parameters.

  DownLoad: Full-Size Img PowerPoint

  In Fig. 3c, the sensing performance of V_Cl-CT-10 toward a series of NO₂ concentrations was investigated. The calibration curves of the response values show linear responses over the NO₂ concentration ranges of 0.02–1 ppm and 5–100 ppm (Fig. 3d). Interestingly, a larger slope is observed in the low-concentration range compared to the high-concentration range [30]. The lower detection limit (LOD) of V_Cl-CT-10 is calculated as 1 ppb (Figs. S11 and S12 in Supporting information), which falls below the exposure level mandated by the National Institute for Occupational Safety and Health [3]. The repeatability of V_Cl-CT-10 was confirmed by seven consecutive cycles of exposure to 1ppm NO₂. The response is almost the same in each cycle (Fig. 3e). Furthermore, V_Cl-CT-10 also shows a repeatable response at low NO₂ concentrations (Fig. S13 in Supporting information). A radar chart comparing the key parameters of the V_Cl-CT-x-based NO₂ gas sensors is shown in Fig. 3f. V_Cl-CT-10 exhibits the most optimal sensitivity, response/recovery speed, and LOD value.

  Selectivity, long-term stability, and humidity resistance (RH) are also important criteria for the practical application of gas sensors [31, 32]. Fig. 4a shows the sensing responses of the V_Cl-CT-10 and TiO₂ NTs samples toward NO₂ and some typical co-existing interference gases (the interference concentrations are 5–15 times the NO₂ concentration). The as-formed TiO₂ NTs exhibit a poor response to NO₂ and other gases. In contrast, V_Cl-CT-10 exhibits a remarkably high response toward NO₂ gas and negligible responses to the interference gases. The outstanding selectivity of V_Cl-CT-10 can be attributed to the affinity of NO₂ molecules to V_Cl-CT-10, which is induced by the abundant Cl vacancies on this sample [33]. Fig. 4b presents the real-time response of V_Cl-CT-10 to the sequential injection and exhaustion of 300ppm interference gases and 20ppm NO₂. A high response is only observed when V_Cl-CT-10 is exposed to NO₂. Fig. 4c displays the anti-interference ability of this sensor. When mixtures of NO₂ and other interference gas are continuously injected into the test chamber, the response of V_Cl-CT-10 is almost the same as that toward NO₂ gas. Furthermore, the sensor exhibited similar sensing performance under REL and in dark (Fig. S14 in Supporting information). These results indicate the good selectivity and reliability of V_Cl-CT-10, implying its promising application prospects for monitoring NO₂ in complex environments [34, 35].

  Figure 4

  

  Figure 4.  (a) Selectivity of V_Cl-CT-10 and pure TiO₂ NTs toward NO₂ and different interference gases. (b, c) Mixed gas detection using V_Cl-CT-10. (d) Real-time continuous NO₂ sensing curve. (e) Long-term stability measurement of 1 ppm NO₂ over 50 days. (f) Response of V_Cl-CT-10 under different RH.

  DownLoad: Full-Size Img PowerPoint

  The real-time sensing performance of V_Cl-CT-10 was also measured by the continuous addition of NO₂, as shown in Fig. 4d. The V_Cl-CT-10-based sensing chip provides the corresponding response when successive exposure to different concentrations of NO₂. Importantly, this sensor shows a good in-time response and rapid recovery after exposure to fresh air. Furthermore, the long-term sensing response for 0.2, 0.35, 0.5, and 0.8ppm of NO₂ were investigated over 50 days. The responses at all concentrations show no significant decay (Fig. 4e and Fig. S15 in Supporting information), indicating the excellent long-term stability of V_Cl-CT-10 based sensing chips. Meanwhile, the XRD patterns of V_Cl-CT-10 samples were analyzed after storage in the environment for up to 3 months (Fig. S16 in Supporting information). No distinct changes in the diffraction peaks are found. To confirm the uniformity of V_Cl-CT-10 performance, responses from ten V_Cl-CT-10 samples were obtained, showing satisfactory uniformity with a relative standard difference (RSD) of 2.1% (Fig. S17 in Supporting information). Considering the variable environment humidity in real applications, the sensing performance of the V_Cl-CT-10-based sensing chip to 1 ppm NO₂ was measured under different relative humidity conditions (Fig. 4f and Fig. S18 in Supporting information). The sample shows a slightly decreasing response when the ambient RH is raised. This phenomenon can be explained by the competitive adsorption of H₂O and NO₂ molecules [36]. Despite this, the sensing chip still retains 81% and 77% of the initial response (RH=20%) under high humidities of 75% and 87%, respectively. Additionally, the sensor performance only showed slight fluctuations with the temperature variations (Fig. S19 in Supporting information). A comparison of the sensing performance of previously reported perovskite NO₂ sensors with that of V_Cl-CT-10 is shown in Table S3 (Supporting information). The sensor reported in this work exhibits outstanding advantages, with a low detection limit (LOD=0.001ppm) and high sensitivity.

  Fig. 5a shows the energy band diagrams of TiO₂ NTs and V_Cl-CAIC before contact. Both V_Cl-CAIC and TiO₂ NTs are n-type semiconductors, where electrons are the main carriers [37]. The energy band gaps of V_Cl-CAIC and TiO₂ NTs are 3.25eV and 3.34eV, respectively (Fig. S20 in Supporting information). After the heterojunction is in contact (Fig. 5b), electrons transfer from the conduction band of V_Cl-CAIC to the conduction band of TiO₂ NTs until the Fermi level reaches equilibrium. Mott–Schottky curves further confirm the band diagram of the heterojunction between n-type V_Cl-CAIC and n-type TiO₂ NTs in air (Fig. S21 in Supporting information). When the system is exposed to NO₂ (Fig. 5c), the concentration of carriers in the heterojunction is further depleted by the consumption of electrons. Thus, the depletion layer at the interface widens, which increases the resistance of the sensor.

  Figure 5

  

  Figure 5.  Gas sensing mechanism. (a-c) Change in the energy band of V_Cl-CAIC/TiO₂ NTs heterojunctions (before contact, in air, and after exposure to NO₂). Charge transfer process of V_Cl-CAIC/TiO₂ NTs heterojunctions (d) in air and (e) after exposure to NO₂.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 5d, the oxygen molecules (O₂) in air adsorbed on V_Cl-CT-10 (O_{2 (ads)}) capture electrons from the sensor materials to form oxygen ions (O^–_2(ads), Eq. 1) [37, 38].

  $ \mathrm{O}_{2 \text { (ads) }}+\mathrm{e}^{-} \rightarrow \mathrm{O}_2^{-} \text {(ads) } $

  (1)

  When the gas sensor is exposed to the NO₂ atmosphere (Fig. 5e), the highly electronegative NO₂ molecules extract electrons from the sensing layer (Eq. 2) and also react with the O₂^– ions, consuming electrons to form NO₂^– (Eq. 3) [39].

  $ \mathrm{NO}_{2 \text { (gas) }}+\mathrm{e}^{-} \rightarrow \mathrm{NO}_2^{-} \text {(ads) } $

  (2)

  $ \mathrm{NO}_{2 \text { (ads) }}+\mathrm{O}_2^{-}+2 \mathrm{e}^{-} \rightarrow \mathrm{NO}_2^{-}+2 \mathrm{O}^{-} $

  (3)

  These reactions lead to a further reduction in the number of primary carriers (i.e., electrons (e^–)), thereby increasing the resistance of the gas sensor. When the environment reverts to air, electrons previously captured by NO₂ are released back into the conduction band of the sensing materials. Consequently, the resistance returns to its original state. As a result, excellent NO₂ sensing performance is achieved at RT.

  Compared to TiO₂ NTs, the superior sensing sensitivity of the V_Cl-CT-10 sensor can be attributed to two factors: First, the 1D nanotube array structure facilitates the electron transfer, and the nanotubes provides plenty of active sites. Second, the plenty V_Cl also plays a crucial role in enhancing the gas detection capacity of the sensor. In perovskite nanocrystals, the CB consists of spin-orbit coupling, while the VB has an antibonding feature [13-15]. These contrasting differences in the orbital characteristics induce the formation of shallow trap states [15, 40], which likely improve the sensitivity of the sensor. In this case, NO₂, acting as an electron acceptor, interacts with the defect sites, thus inducing obvious changes in the material resistance.

  High-sensitivity flexible gas sensors have garnered widespread attention for use in wearable and portable electronic devices, particularly in the fields of personal environmental monitoring and healthcare [41]. As shown in Fig. 6a, the as-proposed sensor shows good flexibility. The dynamic response-recovery characteristics and response values of this sensor show minimal variation at bending angles of 0°, 30°, 70°, and 90° (Fig. 6b), and the sensing performance shows negligibly changes after being bent 100 and 500 times (Fig. 6c). These experimental outcomes demonstrate the promising potential of V_Cl-CT-10 in the domain of flexible sensing applications.

  Figure 6

  

  Figure 6.  Flexibility and application. (a) Optical images of flexible sensor. (b) Response recovery curve for 1ppm NO₂ at bending angles of 0°, 30°, 70°, and 90°. (c) Sensitivity after 100 and 500 bends. (d) Equipment images for remote control vehicle loading tests. (e, f) Wireless sensor systems for gas leak detection or hazard alarms.

  DownLoad: Full-Size Img PowerPoint

  To assess the in situ practical application value of the prepared sensor in real environments, a wireless measuring system was developed to monitor gas leaking and issue hazard warnings in autonomous vehicles under simulated cruising conditions (Fig. 6d). This system contains a V_Cl-CT-10 sensor, an LED flashing alarm, a 18,650 lithium-ion battery, and portable testing circuits (Fig. S22 in Supporting information). The remote-controlled car loading with the measuring system performs the gas monitoring task (Fig. 6e). The sensor provides a real-time sharp alarm (red LED lighting) as soon as the sensor detects the leaking of NO₂ (Fig. 6f). Furthermore, this remote-controlled car also demonstrates excellent application ability for sensing the NO₂ released by a fuel car (see Video S1 in Supporting information). These application tests imply that the wireless NO₂ sensing system possesses satisfactory application value for gas industry applications and pollutant discharge monitoring.

  In conclusion, a NO₂ gas sensor was designed by coating Cl vacancy-rich Cs₂AgInCl₆ onto TiO₂ NTs. The optimized sensor shows high NO₂ sensitivity, with rapid response/recovery speeds (38/59 (s/s) at 1 ppm) at room temperature. Furthermore, this sensor also exhibits good moisture resistance, flexibility, stability, miniaturization. These advantages suggest the excellent promise of V_Cl-CT-10 for environmental monitoring. The combination of lead-free perovskite nanocrystals with MOS offers a new route for constructing sensitive interfaces for gas sensing applications.

  Declaration of competing interest

  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.

  CRediT authorship contribution statement

  Keke Li: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Haiquan Wang: Writing – original draft, Data curation. Zhen-Kun He: Methodology, Formal analysis. Yan-Yan Song: Writing – review & editing, Methodology, Funding acquisition. Zhida Gao: Writing – review & editing, Methodology. Chenxi Zhao: Writing – review & editing, Formal analysis.

  Acknowledgments

  This work was supported by the National Natural Science Foundation of China (No. 22374015), the Fundamental Research Funds for the Central Universities (N2424020), Liaoning Province Foundation for Distinguished Young Scholars (No. 1727146584490, to Y.-Y. Song), and Liaoning Binhai laboratory (No. LBLG-2024-02).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110610.

  
  
• 1. [1]

     D. Zhang, Z. Liu, C. Li, et al., Nano Lett. 4 (2004) 1919–1924.

  2. [2]

     H. Long, A. Harley-Trochimczyk, T. Pham, et al., Adv. Funct. Mater. 26 (2016) 5158–5165. doi: 10.1002/adfm.201601562

  3. [3]

     K. Kuklinska, L. Wolska, J. Namiesnik, et al., Atmos. Pollut. Res. 1 (2015) 129–137.

  4. [4]

     J. Xu, Y. Xu, J. Li, et al., ACS Sens. 8 (2023) 3487–3497. doi: 10.1021/acssensors.3c01061

  5. [5]

     Q.L. Hong, Y. Zhao, X. Ai, et al., Chem. Eng. J. 482 (2024) 148960.

  6. [6]

     L. Yang, J. Feng, J.N. Wang, et al., Chin. Chem. Lett. 33 (2022) 5169–5173.

  7. [7]

     Q. Xue, H.Y. Sun, Y.N. Li, et al., Chem. Eng. J. 421 (2021) 129760.

  8. [8]

     J. Xu, C. Liang, Z. Gao, et al., Chin. Chem. Lett. 34 (2023) 107863.

  9. [9]

     L. Huang, G. Cai, R. Zeng, et al., Anal. Chem. 94 (2022) 9487–9495. doi: 10.1021/acs.analchem.2c02010

  10. [10]

      J. Shu, Z. Qiu, S. Lv, et al., Anal. Chem. 89 (2017) 11135–11142. doi: 10.1021/acs.analchem.7b03491

  11. [11]

      T. Liu, Y. Liu, X. Gao, et al., Chin. Chem. Lett. 34 (2023) 107883.

  12. [12]

      H. Lin, S. Zhang, D. Zhao, et al., Chin. Chem. Lett. (2024), https://doi.org/10.1016/j.cclet.2024.109795. doi: 10.1016/j.cclet.2024.109795

  13. [13]

      X. Li, X. Gao, X. Zhang, et al., Adv. Sci. 8 (2021) 2003334.

  14. [14]

      J. Casanova-Chafer, R. Garcia-Aboal, P. Atienzar, et al., ACS Sens. 7 (2022) 3753–3763. doi: 10.1021/acssensors.2c01581

  15. [15]

      H. Huang, M. Hao, Y. Song, et al., Small 16 (2020) 1904462.

  16. [16]

      W. Xuan, L. Zheng, L. Cao, et al., ACS Sens. 8 (2023) 1252–1260. doi: 10.1021/acssensors.2c02656

  17. [17]

      K. Ho, M. Wei, E.H. Sargent, et al., ACS Energy Lett. 6 (2021) 934–940. doi: 10.1021/acsenergylett.0c02247

  18. [18]

      W. Ye, L. Zhao, H.Z. Lin, et al., Nat. Commun. 14 (2023) 2133.

  19. [19]

      Z.K. Chen, W. Ye, H.Z. Lin, et al., ACS Sens. 6 (2021) 3800–3807. doi: 10.1021/acssensors.1c01791

  20. [20]

      W. Ye, J.H. He, Q. Cao, et al., Adv. Mater. 33 (2021) 2100674.

  21. [21]

      K. Li, S. Li, W. Zhang, et al., J. Colloid Interface Sci. 596 (2021) 376–383.

  22. [22]

      R. Kou, H. He, Y. Lu, et al., Chem. Eng. J. 476 (2023) 146546.

  23. [23]

      H. Lv, H. Yin, N. Jiao, et al., Small Methods 7 (2023) 2201365.

  24. [24]

      Z. He, Q. Tang, X. Liu, et al., Energy Fuels 35 (2021) 15005–15009. doi: 10.1021/acs.energyfuels.1c01482

  25. [25]

      J. Pi, X. Jia, Z. Long, et al., Adv. Energy Mater. 12 (2022) 2202074.

  26. [26]

      A. Khalid, A.M. Alanazi, S.A. Alderhami, et al., Mat. Sci. Semicon. Proc. 165 (2023) 107678.

  27. [27]

      Z. Lin, Q. Wu, P. Wang, et al., Phys. Scr. 98 (2023) 115960. doi: 10.1088/1402-4896/acfc80

  28. [28]

      L. Tang, P. Xu, M. Li, et al., Chin. Chem. Lett. 31 (2020) 2155–2158.

  29. [29]

      M.E. Franke, T.J. Koplin, U. Simon, Small 2 (2006) 36–50. doi: 10.1002/smll.200500261

  30. [30]

      Z. Chu, M. Xiao, Q. Dong, et al., Chin. Chem. Lett. 34 (2023) 107197.

  31. [31]

      Y. Gao, J. Tang, Q. Zhou, et al., Anal. Chem. 96 (2024) 5014–5021. doi: 10.1021/acs.analchem.4c00324

  32. [32]

      Y. Gao, Z. Yu, L. Huang, et al., Anal. Chem. 95 (2023) 9130–9137. doi: 10.1021/acs.analchem.3c01690

  33. [33]

      X. Guo, P. Wang, T. Wu, et al., Angew. Chem. Int. Ed. 63 (2024) e202318792.

  34. [34]

      A.A. Parfenov, O.R. Yamilova, L.G. Gutsev, et al., J. Mater. Chem. C 9 (2021) 2561–2568. doi: 10.1039/d0tc03559a

  35. [35]

      M.A. Stoeckel, M. Gobbi, S. Bonacchi, et al., Adv. Mater. 29 (2017) 1702469.

  36. [36]

      Y. Yang, J. Mao, D. Yin, et al., J. Hazard. Mater. 455 (2023) 131591.

  37. [37]

      Q. Xu, B. Zhu, B. Cheng, et al., Appl. Catal. B: Environ. 255 (2019) 117770.

  38. [38]

      Z. Song, W. Tang, Z. Chen, et al., Small 18 (2022) 2203212.

  39. [39]

      Y. Yang, D. Zhang, D. Wang, et al., J. Mater. Chem. A 9 (2021) 14495–14506. doi: 10.1039/d1ta03739k

  40. [40]

      X. Jiao, Z. Chen, X. Li, et al., J. Am. Chem. Soc. 139 (2017) 7586–7594. doi: 10.1021/jacs.7b02290

  41. [41]

      J. Zhao, H. Wang, Y. Cai, et al., ACS Sens. 9 (2024) 1644–1655. doi: 10.1021/acssensors.4c00137

• 

  Figure 1  (a) Schematic diagram of gas sensing device synthesis. (b, c) SEM images of TiO₂ NTs. (d) SEM, (e) TEM, and (f) HRTEM with fast Fourier transformed (FFT) analysis of V_Cl-CT-10. (g) HRTEM image of defective V_Cl-CAIC.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) XRD patterns of samples. (b) EPR spectra of V_Cl-CT-10 and CT-10. (c) In 3d XPS spectra of V_Cl-CAIC and V_Cl-CT-10.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Gas responses of TiO₂ NTs and V_Cl-CT-x. (b) Response and recovery times of V_Cl-CT-10 toward 1ppm of NO₂. (c) Dynamic responses and (d) linear relationship curve of V_Cl-CT-10 toward 0.02–100ppm of NO₂. (e) Repeatability of V_Cl-CT-10 toward 1ppm of NO₂. (f) Radar graph analysis of key sensing parameters.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Selectivity of V_Cl-CT-10 and pure TiO₂ NTs toward NO₂ and different interference gases. (b, c) Mixed gas detection using V_Cl-CT-10. (d) Real-time continuous NO₂ sensing curve. (e) Long-term stability measurement of 1 ppm NO₂ over 50 days. (f) Response of V_Cl-CT-10 under different RH.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Gas sensing mechanism. (a-c) Change in the energy band of V_Cl-CAIC/TiO₂ NTs heterojunctions (before contact, in air, and after exposure to NO₂). Charge transfer process of V_Cl-CAIC/TiO₂ NTs heterojunctions (d) in air and (e) after exposure to NO₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Flexibility and application. (a) Optical images of flexible sensor. (b) Response recovery curve for 1ppm NO₂ at bending angles of 0°, 30°, 70°, and 90°. (c) Sensitivity after 100 and 500 bends. (d) Equipment images for remote control vehicle loading tests. (e, f) Wireless sensor systems for gas leak detection or hazard alarms.

  下载: 全尺寸图片 幻灯片
•