English

• Nitrogen dioxide (NO₂) with toxicity and corrosiveness may cause photochemical smog and acid rain, seriously damaging the environment and threatening human health [1–3]. Specifically, NO₂ harms human respiratory tract, leading to chest tightness, dyspnea and even death [4,5]. Accordingly, the time-weighted average (TWA) exposure limit of NO₂ is 1 ppm and the threshold limit value is 3 ppm [6], in which selective detection and remote monitoring of low-concentration NO₂ are critically required.

  The chemiresistive gas sensors are extensively utilized for detecting NO₂ due to their excellent portability [7] and sensing materials as a core component draw increasing attention. Among various NO₂ sensing materials, semiconducting metal oxides (SMOs) such as ZnO nanowalls [8], porous In₂O₃ nanosheets [9], Ni-doped SnO₂ nanofiber array [10], Fe-doped WO₃ films [11] and Fe₂O₃/BiVO₄ composites [12] have been investigated by exploring nanostructure, doping and constructing heterojunctions. Recently, two-dimensional transition metal dichalcogenide (TMDs) such as WS₂ [13], MoS₂ [14] and WSe₂/WS₂ [15] have been developed for room-temperature NO₂ sensing. Although great progress has been made, the power consumption of SMOs-built sensors and the response speed of TMDs need further improving. Especially, the above-reported strategies on boosting selectivity will simultaneously improve the responses of both target gas and interfering ones, which could lead to a decrease in interfering selectivity.

  Carbon materials such as graphene and its composites show great application potential in room-temperature NO₂ sensing owing to their highly electrical conductivity and large specific surface areas [16,17]. However, the absorption of gas molecules on pristine carbon materials is determined to their weak van der Waals interaction, limiting their usage in highly selective and low-detection-limit gas sensing [18]. Actually, functionalized carbon materials with sensing dopants or nanostructures have been demonstrated to improve adsorption capabilities [19,20]. Nevertheless, it needs further exploration for a NO₂ sensing that could simultaneously possess high selectivity, low detection limit and fast response.

  Ideally, the carbon materials with a large specific surface area are decorated by polymers accompanied with functional groups, which may be devoted to specific identification of target gas such as NO₂ while suppressing other interfering gases, however, few has been reported. In this study, a three-dimensional (3D) carbon framework functionalized with polyethylenimine (PEI/C framework, Figs. 1a and b) has been developed for selective NO₂ sensing. Moreover, the detection limit of PEI/C framework was found as low as 100 ppb NO₂. Here, PEI/C framework and their-built sensor prototypes will be taken as examples in the following description.

  Figure 1

  

  Figure 1.  (a, b) The schematic diagram on the synthesis of PEI/C framework. The SEM images of (a₁, a₂) C framework and (b₁, b₂) PEI/C framework by a closed magnification. (c) The XPS survey spectra, (d) high-resolution N 1s spectra and (e) FT-IR spectra of C framework and PEI/C framework.

  DownLoad: Full-Size Img PowerPoint

  In Figs. 1a₁ and a₂, the morphology of C framework was observed by scanning electron microscopy (SEM) and show porous 3D structure with smooth surface. Comparatively, the irregular coating was seen over the surface of 3D framework after introducing PEI in Figs. 1b₁ and b₂ and Fig. S1 (Supporting information). The various PEI/C frameworks are obtained by feeding various concentration of PEI in the precursor, however, it should be illustrated that the PEI content on C framework cannot be tuned. To further confirm the doping of PEI, X-ray photoelectron spectroscopy (XPS) was conducted. In Fig. 1c, the survey spectra of C framework show the peaks of C and O elements, while that of PEI/C framework present additional N element, which is the representative element of PEI and indicate the decoration of PEI on the C framework. Meanwhile, the high-resolution N 1s and C 1s spectra in Fig. 1d and Fig. S2a (Supporting information) is verified as well. In Fig. S2b (Supporting information), the high-resolution O 1s spectrum is deconvoluted into adsorbed oxygen (O_ads) and hydroxyl groups, in which O_ads of PEI/C framework possess higher proportion than that of C framework, contributing to the improved NO₂ sensing performance [21]. Moreover, the Fourier transform infrared spectroscopy (FT-IR) was performed in Fig. 1e, in which the FT-IR spectrum of PEI/C framework exhibits additional peaks at 2928 and 2837 cm⁻¹ originating from PEI compared with that of C framework [22]. As a result, all the above characterizations reveal that the PEI is doped into C framework.

  The PEI content-dependent NO₂ sensing was firstly evaluated at room temperature, the sensing resistance curves of pristine C framework and C framework functionalized with various content of PEI were shown in Fig. 2a, in which the pristine C framework presents poor response and recovery especially at higher concentration of 10 ppm NO₂. Remarkably, the response and recovery could be improved once introducing PEI, and the response value depend on the content of PEI on the C framework, especially the 10 mmol/L PEI/C framework exhibit the highest response to NO₂ in Fig. 2b and is thus taken as sample in the following demonstration. In Fig. 2c, the lower concentration of NO₂ was evaluated and the lowest detected content can reach ~100 ppb. Further, the response/recovery time of PEI/C framework to 5 ppm and 1 ppm NO₂ are evaluated of ~ 16.8 s/98.8 s and 33.5 s/208.7 s in Fig. 2d, respectively. Comparatively, the NO₂ sensing performances of various sensing materials at room temperature were summarized in Table S1 (Supporting information) [21,23–29], the PEI/C framework simultaneously presents advantages in low detection limit and fast response and recovery (Fig. 2e). In addition, the repetitive and long-term stability are crucial parameters and were thus investigated. In Fig. 2f, the PEI/C framework presents better repeatability than that of C framework by cyclically testing 1 ppm NO₂, and the almost identical responses of PEI/C framework were summarized in Fig. 2g. Also, the 1 ppm NO₂ sensing performance of the same PEI/C framework sensor prototype during 40 days were evaluated and compared in Fig. 2h, which suggests excellent repeatability and reproducibility. It should be illustrated that the sluggish sensing response and recovery after 40 days may be attributed to the humidity and the interference from other molecules to PEI/C framework [30].

  Figure 2

  

  Figure 2.  (a, b) The comparison on NO₂ sensing performance between C framework and the one functionalized with various PEI. The real-time resistance curves of PEI/C framework to (c) 0.1–1 ppm NO₂. (d) The response and recovery time to 5 ppm and 1 ppm NO₂, respectively. (e) Comparison on low detection limit, response and recovery time with those of previous publications. (f) The real-time resistance curves of PEI/C framework and C framework to cycling 1 ppm NO₂. (g) The responses of PEI/C framework during 11 cycles. (h) The evaluation on long-term stability.

  DownLoad: Full-Size Img PowerPoint

  Selectivity is a vital parameter in sensing performance, which was evaluated by exposing the C framework and PEI/C framework to various gases with 100 ppm of ethanol, acetone, H₂ and CO, 10 ppm NH₃, 1 ppm of H₂S and NO₂ in Fig. S3 (Supporting information), respectively. The sensing response values were summarized in Fig. 3a, in which the pristine C framework possesses indistinguishable response to NH₃, H₂S and NO₂. Comparatively, PEI/C framework selectively increases the response of target NO₂ and decreases those of interfering gases such as NH₃ and H₂S. Further, we elaborately performed the NO₂ sensing evaluation of PEI/C framework after exposing it to the main interfering gases of NH₃ and H₂S in Fig. 3b, showing excellent anti-interference ability. Meanwhile, the principal component analysis (PCA) was carried out to distinguish the target NO₂ from interfering NH₃ and H₂S (Fig. 3c), which further verifies the excellent selective identification of PEI/C framework to NO₂. The detailed data of the above PCA are analyzed in Tables S2 and S3 (Supporting information). Additionally, the sensing performance of PEI/C framework to 1 ppm NO₂ at various humidity was studied in Fig. S4 (Supporting information) and the responses were compared in Fig. 3d, and one can see that the high environment humidity may affect NO₂ sensing. To further evaluate the reproducibility of the PEI/C framework sensor prototype in practical application, the NO₂ sensor device was constructed in Fig. S5 (Supporting information). When sensor module is powered by laptop, the acquired NO₂ sensing curve not only can be visualized on the smartphone but also on the laptop (Fig. 4a). Correspondingly, the cycling 1 ppm NO₂ sensing data was recorded on the laptop (Fig. 4a₁) and smartphone (Fig. 4a₂), respectively, showing excellent reliability of our simulated sensing device.

  Figure 3

  

  Figure 3.  (a) The comparison on selectivity between C framework and PEI/C framework. (b) The real-time resistance curve and (c) PCA analysis of PEI/C framework to target NO₂, interfering NH₃ and H₂S. (d) The comparison on sensing response to 1 ppm NO₂ at various relative humidity.

  DownLoad: Full-Size Img PowerPoint

  Figure 4

  

  Figure 4.  (a) The photograph of sensing-response curve simultaneously displayed on the laptop and smartphone. The records of sensing voltage to cycling 1 ppm NO₂ on (a₁) laptop and (a₂) smartphone, respectively. The sensing mechanism schematic of (b) C framework and (c) PEI/C framework.

  DownLoad: Full-Size Img PowerPoint

  To understand the sensing mechanism of the PEI/C framework, the role of the pristine C framework was first investigated. In Fig. 2a, the C framework shows decreased sensing resistance when it was exposed to oxidizing NO₂ gas, which indicates the C framework possesses p-type semiconductor characteristic. Such phenomenon is verified by the Mott–Schottky (M–S) plot with a negative slope in Fig. S6 (Supporting information), depicting a typical p-type semiconductor nature (hole as the main carrier). Accordingly, the NO₂ sensing mechanism of C framework is diagramed in Fig. 4b. In air, the O₂ capture electrons from C framework to form adsorbed O₂⁻ (Eq. 1), increasing the surface hole concentration of C framework thus decreasing the resistance. In NO₂, the NO₂ will extract electrons from adsorbed oxygen ions on the surface of C framework (Eq. 2). Also, NO₂ directly capture more electrons from C framework (Eq. 3) to form adsorbed NO₂⁻ due to the stronger electron affinity of NO₂ than that of O₂ [27,31]. Both the above reactions generate more holes on p-type C framework, the NO₂ sensing with decreased resistance was observed.

  (1)

  (2)

  (3)

  Similarly, the NO₂ sensing mechanism of PEI/C framework is shown in Fig. 4c. It should be noted that the response of C framework to NO₂ increases while those of interfering gases decrease after being functionalized with PEI (Fig. 3a). Such phenomenon is entitled as "PEI serves as an active layer for target NO₂ while a passivated one for interfering gases", suggesting a boosted selectivity. Theoretically, PEI possess rich amine groups with electron-donating property serving as n-dopants to the p-type C framework [32,33], which is confirmed by the increased initial resistance of C framework after doping PEI (Fig. 2a). Therefore, the electrons transfer from the PEI/C framework surface to the NO₂ molecules due to the stronger electron affinity of NO₂ [27]. In this case, the PEI polymer acts as an active intermediate layer [34], promoting NO₂ sensing with fast response and recovery. On the contrary, if the PEI/C framework was exposed to the interfering gases such as H₂S and NH₃, which will react with pre-adsorbed oxygen species (O₂⁻) to release electrons into PEI/C framework. However, the C framework has pre-received rich electrons from the donor of PEI, which may cause insensitivity to the extra electrons and thus decreased response to the interfering gases. As can be interpreted that PEI acts as a passivated layer for the interfering gases. Accordingly, the PEI/C framework selectively enhances the sensing response to target NO₂, while decreasing the ones to those of interfering gases.

  The NO₂ sensing was simulated by building the sensing device integrated PEI/C framework, communicating with a smartphone. Specifically, the PEI/C framework sensor prototype and temperature-humidity (temp.-hum.) sensor are connected to the microcontroller NodeMCU (ESP8266), which conducts Wireless Fidelity (Wi-Fi) communication between the sensors and smartphone, and the corresponding block diagram is schematized in Fig. 5a. Meanwhile, the NodeMCU performs the program to process the sensing data from PEI/C framework sensor prototype and temp.-hum. sensor, and to control the NO₂ alarming, the flow chart of the main program is shown in Fig. 5b. Notably, the NO₂ sensor prototype on the sensor module and the displayed information on the smartphone are marked in Fig. 5c. Experimentally, the alarming was simulated in the supplementary video, when 3 ppm NO₂ was injected and the sensing voltage reach alarm threshold, the smartphone shows red "AlARMING!" word (Fig. 5d). When the sensor module is placed in the air, the green "Monitoring" display on smartphone (Fig. 5c). Remarkably, as above presented sensor module powered by battery, the environmental temp.-hum. and NO₂ sensing data can be remotely monitored and updated in real-time by the smartphone.

  Figure 5

  

  Figure 5.  (a) The block diagram of NO₂ monitoring and alarming device and (b) the corresponding flow diagram of alarm program. (c) The photograph of NO₂ monitoring and alarming device. (d) The real-time sensing data and alarming display on the smartphone.

  DownLoad: Full-Size Img PowerPoint

  To summarize, the enhanced selective NO₂ sensing has been developed with PEI/C framework via combined template synthesis and subsequent doping. Experimentally, the 3D PEI/C framework with porous morphology and irregular coating are observed. Theoretically, the PEI serves as an active layer for target NO₂ while a passivated layer for interfering gases, and the PEI/C framework possesses a large specific surface ratio (637 m²/g, Fig. S7a in Supporting information). Beneficially, the sensor prototypes with PEI/C framework show excellent selectivity to NO₂, in which the response of C framework to NO₂ increases while those of interfering gases decrease after being functionalized with PEI. Moreover, the sensor prototypes present low detection limit of ~100 ppb NO₂ and a fast response time of 16.8 s to 5 ppm NO₂. As an example of application, the PEI/C framework has been integrated into a sensing device that communicates with a smartphone to simulate monitoring and alarming NO₂, which is potential in future intelligent sensing of Internet of Things.

  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.

  Acknowledgments

  This work was financially supported by the National Natural Science Foundation of China (No. 52072184) and Tianjin Research Innovation Project for Postgraduate Students (General Project, No. 2022BKY035).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.109230.

  
  
• 1. [1]

     M. Woellner, S. Hausdorf, N. Klein, et al., Adv. Mater. 30 (2018) 1704679. doi: 10.1002/adma.201704679

  2. [2]

     Z. Yang, H. Zou, Y. Zhang, et al., Adv. Funct. Mater. 32 (2022) 2108959. doi: 10.1002/adfm.202108959

  3. [3]

     Z. Chu, M. Xiao, Q. Dong, et al., Chin. Chem. Lett. 34 (2023) 107197. doi: 10.1016/j.cclet.2022.02.003

  4. [4]

     L.B. Kreuzer, C.K.N. Patel, Science 173 (1971) 45–47. doi: 10.1126/science.173.3991.45

  5. [5]

     H.T. Huang, W.L. Zhang, X.D. Zhang, X. Guo, Sens. Actuator. B: Chem. 265 (2018) 443–451. doi: 10.1016/j.snb.2018.03.073

  6. [6]

     K. Tian, X.X. Wang, H.Y. Li, et al., Sens. Actuator. B: Chem. 227 (2016) 554–560. doi: 10.1016/j.snb.2015.12.104

  7. [7]

     M.E. Franke, T.J. Koplin, U. Simon, Small 2 (2006) 301. doi: 10.1002/smll.200690010

  8. [8]

     L. Yu, J. Wei, Y. Luo, et al., Sens. Actuator. B: Chem. 204 (2014) 96–101. doi: 10.1016/j.snb.2014.07.071

  9. [9]

     L. Gao, Z. Cheng, Q. Xiang, et al., Sens. Actuator. B: Chem. 208 (2015) 436–443. doi: 10.1016/j.snb.2014.11.053

  10. [10]

      W.T. Li, X.D. Zhang, X. Guo, Sens. Actuator. B: Chem. 244 (2017) 509–521. doi: 10.1016/j.snb.2017.01.022

  11. [11]

      T. Tesfamichael, C. Piloto, M. Arita, J. Bell, Sens. Actuator. B: Chem. 221 (2015) 393–400. doi: 10.1016/j.snb.2015.06.090

  12. [12]

      S. Bai, K. Tian, H. Fu, et al., Sens. Actuator. B: Chem. 268 (2018) 136–143. doi: 10.1016/j.snb.2018.03.173

  13. [13]

      K.Y. Ko, J.G. Song, Y. Kim, et al., ACS Nano 10 (2016) 9287–9296. doi: 10.1021/acsnano.6b03631

  14. [14]

      H.H. Hau, T.T.H. Duong, N.K. Man, et al., Sens. Actuator. A: Phys. 332 (2021) 113137. doi: 10.1016/j.sna.2021.113137

  15. [15]

      Y. Kim, S. Lee, J.G. Song, et al., Adv. Funct. Mater. 30 (2020) 2003360. doi: 10.1002/adfm.202003360

  16. [16]

      J. Wu, Z. Li, X. Xie, et al., J. Mater. Chem. A 6 (2018) 478–488. doi: 10.1039/C7TA08775F

  17. [17]

      T. Wang, X. Li, C. Li, et al., Sens. Actuator. B: Chem. 382 (2023) 133507. doi: 10.1016/j.snb.2023.133507

  18. [18]

      Y.R. Choi, Y.G. Yoon, K.S. Choi, et al., Carbon 91 (2015) 178–187. doi: 10.1016/j.carbon.2015.04.082

  19. [19]

      T.T. Tung, M.J. Nine, M. Krebsz, et al., Adv. Funct. Mater. 27 (2017) 1702891. doi: 10.1002/adfm.201702891

  20. [20]

      W. Li, R. Chen, W. Qi, et al., ACS Sens. 4 (2019) 2809–2818. doi: 10.1021/acssensors.9b01509

  21. [21]

      J.L. Fan, X.F. Hu, W.W. Qin, et al., J. Mater. Chem. C 11 (2023) 2364–2374. doi: 10.1039/D2TC04680F

  22. [22]

      D.K. Singh, V. Kumar, V.K. Singh, S.H. Hasan, RSC Adv. 6 (2016) 56684–56697.

  23. [23]

      M. Reddeppa, N.T. KimPhung, G.V. Murali, et al., Sens. Actuator. B: Chem. (2020) 129175.

  24. [24]

      X. Chen, S. Wang, C. Su, et al., Sens. Actuator. B: Chem. 305 (2020) 127393. doi: 10.1016/j.snb.2019.127393

  25. [25]

      Y. Yang, D. Zhang, D. Wang, Z. Xu, J. Zhang, J. Mater. Chem. A 9 (2021) 14495–14506. doi: 10.1039/D1TA03739K

  26. [26]

      D. Wang, D. Zhang, J. Guo, et al., Nano Energy 89 (2021) 106410. doi: 10.1016/j.nanoen.2021.106410

  27. [27]

      W. Zheng, Y. Xu, L. Zheng, et al., Adv. Funct. Mater. 30 (2020) 2000435. doi: 10.1002/adfm.202000435

  28. [28]

      Z. Yang, D. Zhang, H. Chen, Sens. Actuator. B: Chem. 300 (2019) 127037. doi: 10.1016/j.snb.2019.127037

  29. [29]

      W. Yang, P. Wan, X. Zhou, et al., ACS Appl. Mater. Interfaces 6 (2014) 21093–21100. doi: 10.1021/am505949a

  30. [30]

      X. Xing, Z. Li, X. Chen, et al., ACS Appl. Mater. Interfaces 14 (2022) 17911–17919. doi: 10.1021/acsami.1c24190

  31. [31]

      Y. Su, G. Xie, H. Tai, et al., Nano Energy 47 (2018) 316–324. doi: 10.1016/j.nanoen.2018.02.031

  32. [32]

      Y. Zhou, C. Fuentes-Hernandez, J. Shim, et al., Science 336 (2012) 327–332. doi: 10.1126/science.1218829

  33. [33]

      S.-Y. Cho, K.M. Cho, S. Chong, et al., ACS Sens. 3 (2018) 1329–1337. doi: 10.1021/acssensors.8b00247

  34. [34]

      S. Kumar, V. Pavelyev, P. Mishra, N. Tripathi, Bull. Mater. Sci. 43 (2020) 61. doi: 10.1007/s12034-020-2043-6

• 

  Figure 1  (a, b) The schematic diagram on the synthesis of PEI/C framework. The SEM images of (a₁, a₂) C framework and (b₁, b₂) PEI/C framework by a closed magnification. (c) The XPS survey spectra, (d) high-resolution N 1s spectra and (e) FT-IR spectra of C framework and PEI/C framework.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a, b) The comparison on NO₂ sensing performance between C framework and the one functionalized with various PEI. The real-time resistance curves of PEI/C framework to (c) 0.1–1 ppm NO₂. (d) The response and recovery time to 5 ppm and 1 ppm NO₂, respectively. (e) Comparison on low detection limit, response and recovery time with those of previous publications. (f) The real-time resistance curves of PEI/C framework and C framework to cycling 1 ppm NO₂. (g) The responses of PEI/C framework during 11 cycles. (h) The evaluation on long-term stability.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) The comparison on selectivity between C framework and PEI/C framework. (b) The real-time resistance curve and (c) PCA analysis of PEI/C framework to target NO₂, interfering NH₃ and H₂S. (d) The comparison on sensing response to 1 ppm NO₂ at various relative humidity.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) The photograph of sensing-response curve simultaneously displayed on the laptop and smartphone. The records of sensing voltage to cycling 1 ppm NO₂ on (a₁) laptop and (a₂) smartphone, respectively. The sensing mechanism schematic of (b) C framework and (c) PEI/C framework.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) The block diagram of NO₂ monitoring and alarming device and (b) the corresponding flow diagram of alarm program. (c) The photograph of NO₂ monitoring and alarming device. (d) The real-time sensing data and alarming display on the smartphone.

  下载: 全尺寸图片 幻灯片
•