English

• Hydrazine (N₂H₄) is a carcinogenic substance that induces pathological alterations in cells [1]. Moreover, its N₂H₄ vapor or liquid can be absorbed into the body through the respiratory tract, gastrointestinal tract, or skin, leading to central nervous system disorders, dermatitis, and permanent eye damage [2,3]. Despite its adverse effects, it remains an indispensable constituent in the aerospace and chemical sectors [4]. With the development of industry, the widespread use of N₂H₄ has posed potential risks to human health and environmental safety [5]. The United States Environmental Protection Agency (EPA) mandates that its concentration in water should not exceed 10 parts per billion (ppb) [6]. Indeed, a range of detection methods has been developed for trace analysis of N₂H₄. These methods primarily include fluorescence analysis [7,8], electrochemical analysis [9,10], colorimetry [11], and so on. Although the above–mentioned methods demonstrate satisfactory detection performance, there are still limitations in the detection of gaseous N₂H₄.

  Raman spectroscopy is widely recognized as a highly powerful detection technique, distinguished by its non–destructive and non–invasive nature, as well as its relatively straightforward operational procedures [12-16]. When a beam of light illuminates a sample, a small portion of the incident light undergoes wavelength shifts due to the sample's structural characteristics [17]. This phenomenon is known as Raman scattering effect. However, Raman scattering, inherently possessing weak scattering characteristics, cannot be utilized for trace analysis during its initial discovery until the emergence of Surface–enhanced Raman spectroscopy (SERS) [18-22]. In recent years, SERS spectroscopy has been widely used for the detection of many small molecular substances [23-27]. There are also some applications in the detection of N₂H₄, which exhibit excellent performance [28-30]. However, in most instances, noble metals are incorporated into Raman substrates utilized for SERS detection. Indeed, this specific substrate not only comes with a significant cost but also demonstrates a permanent decrease in surface–enhanced Raman scattering (SERS) activity when its surface structure undergoes alterations. As a consequence, the cost of detection experiences a significant surge [31,32]. Currently, the investigation into recyclable non–noble metal Raman substrates for the detection of low–concentration N₂H₄ poses a significant challenge in the field of research.

  Polyoxometalates (POMs) are typically metal oxide clusters [33-41]. POMs contain a significant number of metal–oxygen bonds [42-53], and the stretching vibration of these metal–oxygen bonds exhibits pronounced Raman signals [54-56]. Moreover, POMs possess a high surface charge, which can attract cations from the solution and form new compounds that precipitate in the form of crystals [57]. These well–defined crystalline POMs materials further amplify the Raman signals of the metal–oxygen bonds, resulting in strong Raman signals. In addition, since the transition metals in POMs are in their highest oxidation states, they can be easily reduced by reducing agents, leading to the formation of metal centers with single electrons [58]. These metal centers may induce lattice defects, thereby weakening the Raman signals [59]. In light of the inherent reversibility of the oxidation––reduction process of POMs, the reduced form of POMs can be efficiently regenerated to its oxidized state in the presence of an oxidizing agent [60]. Utilizing the oxidation-reduction mechanism of crystalline POMs holds the potential for achieving repeated utilization of substrate materials. Furthermore, research has demonstrated that certain stable, solvated metal cations (such as Pb²⁺, Bi³⁺, and Ln³⁺, among others) have been found to efficiently form crystalline materials with polyoxometalates through electrostatic attraction [61,62]. These materials exhibit excellent stability and facilitate the reduction of metal centers in the polyoxometalate framework [63]. However, to date, only the photochromic properties of such materials have been extensively studied, while their detection capabilities and potential applications in the field of Raman sensing remain unexplored.

  In this study, a single crystal compound of [Bi(DMF)₈]PMo₁₂O₄₀ (Bi–PMo₁₂) is designed and synthesized by simple conventional solution method. Through the utilization of Bi–PMo₁₂ as a Raman substrate and the presence of N₂H₄ gas, the Mo^VI ions within the POMs anions can be effectively reduced to Mo^V centers. Simultaneously, the intensity of Raman signals weakens with increasing concentrations of N₂H₄, allowing for the detection of N₂H₄ vapor. The detection range spans from 4.5 × 10¹ ppm to 4.5 × 10⁻⁹ ppm. In addition, during the detection process, it is also observed for the first time that the surface of the Bi-PMo₁₂ crystal changes from smooth to rough upon exposure to N₂H₄ vapor. This surface etching also contributes to the quenching of the Raman signals, further assisting in the detection process. Moreover, following the detection process, Bi-PMo₁₂ can undergo recrystallization in a solution of N, N-dimethylformamide (DMF) containing H₂O₂, thereby enabling the cyclic utilization of the substrate materials.

  In previous studies, DMF has been found to significantly enhance the reduction ability of POMs, resulting in an amplified signal response for light intensity detection [64]. In this work, to enhance the redox capacity of POMs and achieve the goal of N₂H₄ detection, it is desirable to incorporate a greater number of DMF molecules into the system, by utilizing Bi³⁺, a metal ion with a high coordination number. During the experimental process, a straightforward and conventional approach involving room temperature stirring and diffusion-induced crystallization was employed. By utilizing this method, a solvent-assisted POMs, Bi-PMo₁₂, was successfully synthesized with an approximate yield of 30% (based on Mo). Comparatively, the synthesis method for Bi-PMo₁₂ is simpler and more cost-effective than that of other noble metal substrates.

  The X-ray single crystal structure analysis reveals that Bi-PMo₁₂ belongs to the triclinic crystal system with the space group (Table S1 in Supporting information). It is composed of [PMo₁₂O₄₀]³⁻ and [Bi(DMF)₈]³⁺ units (Fig. 1a). [PMo₁₂O₄₀]³⁻ is a typical Keggin-type POMs [65]. In [Bi(DMF)₈]³⁺, eight solvent molecules DMF act as monodentate ligands, coordinating with the central Bi³⁺ and forming a commonly observed eight-coordinate square antiprismatic geometry. The [Bi(DMF)₈]³⁺ cation readily forms crystals with the [PMo₁₂O₄₀]³⁻ anion in solution, exhibiting a 1:1 ratio (Fig. 1b). Fig. S1 (Supporting information) shows the morphology of dried Bi-PMo₁₂, indicating the relative stability of the crystalline structure after drying. The experimental X-ray powder diffraction (XRD) data of Bi-PMo₁₂ is consistent with the fitting data of its single crystal phase (Fig. 1c).

  Figure 1

  

  Figure 1.  (a) The crystal structure diagram and (b) single unit cell illustration of Bi–PMo₁₂. (c) The XRD patterns of Bi–PMo₁₂. (d) FTIR and (e) Raman spectra of Bi–PMo₁₂ and H₃PMo₁₂ respectively. (f) SEM and (g) the element mapping images of P, Mo, Bi, and N of Bi–PMo₁₂.

  DownLoad: Full-Size Img PowerPoint

  Analysis of the Fourier–transform infrared spectroscopy (FTIR) of Bi–PMo₁₂ reveals characteristic peaks at 1060 cm⁻¹ to 798 cm⁻¹ (Fig. 1d), corresponding to H₃PMo₁₂O₄₀·4H₂O (H₃PMo₁₂), and a peak at 1651 cm⁻¹ and 1350 cm⁻¹ corresponding to the C=O stretching vibration of DMF [64,66]. Fig. 1e displays the Raman spectra of Bi–PMo₁₂ and H₃PMo₁₂ under 532 nm laser wavelength. The characteristic absorption peaks at 887 cm⁻¹, and 987 cm⁻¹ correspond to the symmetric and antisymmetric stretching vibrations of the Mo=O bonds [67]. As shown in Fig. 1f, the scanning electron microscopy (SEM) image of Bi–PMo₁₂ reveals that even after grinding, the crystal facets remain intact. Although some surface cracks are visible, within the laser spot area (1 µm), the surface remains smooth and intact, ensuring the stability of the output Raman signals. The clear element distribution of P, Mo, Bi, and N in Bi–PMo₁₂ can be observed in Fig. 1g. The uniform distribution of these elements further confirms the successful synthesis of Bi–PMo₁₂.

  We explore the parameters that may affect the detection results of N₂H₄ in order to obtain the optimal testing conditions for Bi–PMo₁₂. As shown in Figs. S2 and S3 (Supporting information), stable Raman signals can be obtained at a laser intensity of 25% and a sampling time of 4 s. Moreover, by heating the sample, it can be observed from the Raman spectra that the Raman signal of Bi–PMo₁₂ remains unchanged until 70 ℃ and undergoes quenching at 80 ℃ (Fig. S4 in Supporting information). This phenomenon can be attributed to the reduction of Bi–PMo₁₂ induced by a small quantity of volatile DMF (Fig. S5 in Supporting information). As illustrated in Fig. S6 (Supporting information), the relative humidity has no impact on the Raman signal intensity of Bi–PMo₁₂, regardless of the presence of N₂H₄ at a concentration of 4.5 × 10⁻⁹ ppm (Figs. S6a and b) or its absence (Figs. S6c and d). Subsequently, ten random points on Bi–PMo₁₂ substrate were selected for Raman testing, and it was observed that the Raman signals exhibited a uniform distribution across these points (Fig. S7a in Supporting information). Fig. S7b (Supporting information) showcases the histogram of Raman signal intensities at 987 cm⁻¹ (I₉₈₇), revealing an impressively low relative standard deviation (RSD) of merely 1.06%. This strongly indicates the remarkable stability of the Bi–PMo₁₂ substrate.

  Based on these findings, it can be concluded that the quenching effect of the Raman signal is optimized at a reaction temperature of 70 ℃ and a reaction time of 30 min (Fig. 2a). It is evident from the obtained spectra that the intensity of Raman signals exhibits a quenching trend with the increasing concentration of N₂H₄ (Fig. 2b). Furthermore, it can be observed from the blue line in Fig. 2c that there is a trend in the change of I₉₈₇ with the logarithm of N₂H₄ concentration (lg(C_N2H4)). The linear correlation coefficient (R²) is 0.9978. The logarithm of the N₂H₄ concentration is inversely proportional to I₉₈₇ within the concentration range of 4.5 × 10¹ ppm to 4.5 × 10⁻⁹ ppm. The equation representing this relationship is: y = 206.62206 – 101.27295x (y represents the intensity of I₉₈₇, and x represents lg(C_N2H4)). In the detection of N₂H₄, Bi–PMo₁₂ can realize the detection at least 4.5 × 10⁻⁹ ppm. As shown in Table S2 (Supporting information), the results obtained in this study are superior to other methods for detecting N₂H₄. The results of the new method surpass the effectiveness of SERS detection of N₂H₄, providing a novel approach for the detection of low–concentration reducing gases. Furthermore, when the concentration of N₂H₄ is 4.5 × 10⁻⁹ ppm and 4.5 × 10⁻¹ ppm, the Raman signal intensity at 987 cm⁻¹ is similar for the three different samples, exhibiting good reproducibility (Fig. S8 in Supporting information). Furthermore, in the presence of 4.5 ppm N₂H₄, the observed variation in I₉₈₇ values at 10 randomly selected points is minimal, with a relatively low standard RSD value of 1.71% (Fig. S9 in Supporting information). In addition, it was discovered that H₃PMo₁₂ also exhibited Raman quenching behavior under N₂H₄ atmosphere (Fig. 2c and Fig. S10 in Supporting information). The Raman quenching effect of H₃PMo₁₂ was optimized under the conditions of 70 ℃ and 30 min (Fig. S11 in Supporting information). However, the detection concentration range only attains at 4.5 × 10¹–4.5 × 10⁻³ ppm. This finding suggests that the presence of [Bi(DMF)₈]³⁺ cations significantly enhances the detection performance of Bi–PMo₁₂, thus playing a crucial role in achieving effective detection at low concentrations.

  Figure 2

  

  Figure 2.  (a) The histogram of Raman signal intensities at 987 cm⁻¹ of Bi–PMo₁₂ in 4.5 × 10⁻⁹ ppm N₂H₄ at different temperatures and times. (b) Raman spectra of Bi–PMo₁₂ after reaction with different concentrations of N₂H₄ (4.5 × 10¹–4.5 × 10⁻⁹ ppm). (c) The correlation curve between the logarithm of N₂H₄ concentration (lg(C_N2H4), C_N2H4 represents the concentration of the N₂H₄ vapor) of Bi–PMo₁₂ and H₃PMo₁₂ at 987 cm⁻¹ strength. (d) The histogram of Raman signal intensity of Bi–PMo₁₂ at 987 cm⁻¹ in the presence of N₂H₄ (4.5 × 10⁻⁴ ppm) and other reducing compounds (4.5 × 10⁻² ppm), respectively. The error bar is obtained at least three replicate trials.

  DownLoad: Full-Size Img PowerPoint

  To showcase the interference resistance capability of Bi–PMo₁₂, we conducted parallel tests on several amine compounds (4.5 × 10⁻² ppm), which have the potential to cause interference. The purpose of these tests was to evaluate the selectivity of Bi–PMo₁₂ in detecting N₂H₄ (4.5 × 10⁻⁴ ppm). As depicted in Fig. 2d, although a quenching effect was observed in the presence of triethylamine (TA), N, N–dimethylformamide (DMF), ammonium hydroxide (NH₄OH), trolamine (TEA), and ethylenediamine (EDA) atmospheres, it was more pronounced in the presence of N₂H₄. Compared to other amine compounds, the I₉₈₇ value decreases by over 60.56% in the presence of N₂H₄. And compared to the blank, other amine compounds only show a decrease of approximately 6.32% in the I₉₈₇ value. Therefore, Bi–PMo₁₂ exhibits excellent selectivity for N₂H₄ detection. This exceptional selectivity can be attributed to the limited ability of other low–concentration amine compounds to effectively reduce Bi–PMo₁₂.

  The FTIR and Raman spectra of Bi–PMo₁₂ before and after detection indicate that there is no change in the molecular structure of Bi–PMo₁₂ (Figs. 3a and b). It is consistent with the conclusion that the structure of POMs remains unchanged after receiving electrons [68]. As shown in Fig. 3c, the X–ray photoelectron spectroscopy (XPS) characteristic peaks for Mo 3d are observed at 236.15, 232.98, 234.93, and 231.77 eV, indicating the presence of Mo^VI 3d_3/2, Mo^VI 3d_5/2, Mo^V 3d_3/2, and Mo^V 3d_5/2. After treatment with N₂H₄ at concentrations of 4.5 × 10⁻⁴ ppm and 4.5 ppm respectively, the peak area corresponding to Mo^V increased from 23.43% to 31.22% and 38.32%, respectively. This can be attributed to that Mo^VI in Bi–PMo₁₂ undergoes partial reduction. It should be noted that the appearance of Mo^V in Bi–PMo₁₂ prior to detection is mainly due to the influence of X–ray in XPS testing, while weaker laser and lower frequency light do not have this effect. Furthermore, there was no change observed in the Bi, O, P, and N elements after the N₂H₄ treatment (Fig. 3d and Fig. S12 in Supporting information).

  Figure 3

  

  Figure 3.  (a) Comparison of FTIR and (b) Raman spectra of Bi–PMo₁₂ before and after detection of N₂H₄. XPS analysis of (c) Mo 3d and (d) Bi 4f spectra of Bi–PMo₁₂ before and after reduction by 4.5 × 10⁻⁴ ppm and 4.5 ppm N₂H₄. (e) Nyquist plots of Bi–PMo₁₂ in the presence of N₂H₄. (f) UV–vis diffuse reflectance spectra (left) and colour changes (right) of Bi–PMo₁₂ with different concentrations of N₂H₄. The surface analysis of Bi–PMo₁₂ (g) before and (h, i) after reduction with N₂H₄ (4.5 × 10⁻⁴ ppm and 4.5 ppm).

  DownLoad: Full-Size Img PowerPoint

  Previous studies have demonstrated that the reduction of Mo^VI generates Mo^V species with hole–lone electron pairs, resulting in increased conductivity and decreased impedance [64]. As shown in Fig. 3e, after the N₂H₄ treatment, the impedance values of Bi–PMo₁₂ gradually decrease, indicating the generation of Mo^V species. Moreover, as depicted in Fig. 3f, the color of Bi–PMo₁₂ gradually deepens with increasing concentrations of N₂H₄ treatment, accompanied by a corresponding increase in the absorbance values within the range of 530–800 nm in the solid–state diffuse reflectance spectra (UV–vis). Specifically, the absorbance values at 710 nm exhibit an increase from 0.0234 (without N₂H₄ treatment) to 0.7729 (with 4.5 ppm N₂H₄ treatment). This analysis indicates that the single–electron transitions occurring between different d orbitals on the Mo^V centers of the reduced Bi–PMo₁₂ can efficiently absorb the Raman scattering irradiation, thereby resulting in the quenching of the Raman signals. As such, we propose to denominate this a new effect as Raman resonance energy transfer (RRET). By harnessing this novel RRET effect, more stable and highly active Raman sensing substrates can be skillfully engineered, enabling their efficacious detection of diverse reducible small molecules.

  Fig. 3g demonstrates the surface morphology of Bi–PMo₁₂ prior to N₂H₄ reduction, revealing a smooth surface with well–defined crystal planes within the laser spot area (1 µm) and no discernible defects. The surface characterization of Bi–PMo₁₂ subsequent to treatment with N₂H₄ (4.5 × 10⁻⁴ and 4.5 ppm) was examined (Figs. 3h and i), revealing that an escalation in N₂H₄ concentration led to enhanced surface etching of Bi–PMo₁₂ and increased depth of surface pores. Moreover, the surface pores of Bi–PMo₁₂ after N₂H₄ treatment exhibit remarkable uniformity, which is a crucial aspect for maintaining stable Raman signals. It can be speculated that this surface modification may represent an additional factor contributing to the attenuation of the Raman signals. To substantiate this proposition, an analysis of the morphology of H₃PMo₁₂ after the detection of equimolar concentration of N₂H₄ was also conducted. The results indicate that the surface of H₃PMo₁₂ is smooth (Fig. S13a in Supporting information) and similar to Bi–PMo₁₂, showing no apparent defects within the laser spot range before N₂H₄ treatment. After treatment with 4.5 ppm N₂H₄ (Fig. S13b in Supporting information), H₃PMo₁₂ exhibited some degree of surface weathering, but did not show uniform pore formation like Bi–PMo₁₂. This change will not increase the diffuse reflection of light. This aligns with the experimental findings, confirming that the surface alteration of Bi–PMo₁₂ is also contributable to the attenuation of Raman signals.

  Based on the aforementioned analyses, we can propose the following reasonable detection mechanism for Bi–PMo₁₂ as depicted in Fig. 4. During the reaction, Mo^VI in POMs is reduced to Mo^V, filling unsTable 4d orbitals. These d–d orbital transitions absorb energy from Raman scattering, resulting in a substantial reduction in signal output. On the other hand, in the presence of a N₂H₄ atmosphere, DMF efficiently captures gaseous N₂H₄ molecules from the surrounding environment, facilitating close contact between gaseous N₂H₄ and the crystal facets of Bi–PMo₁₂. The adsorbed N₂H₄ induces uniform pore defects on the surface of Bi–PMo₁₂, enhancing the diffuse reflection of light. The aforementioned synergistic effects of these two actions ultimately led to the quenching behavior of Raman scattering signals, thereby enabling highly sensitive detection of N₂H₄.

  Figure 4

  

  Figure 4.  Proposed mechanism for the Raman detection of N₂H₄ by the Bi–PMo₁₂ substrate.

  DownLoad: Full-Size Img PowerPoint

  During the experimental process, we observed that Bi–PMo₁₂ crystals could be recrystallized in DMF (Fig. 5). By introducing a small quantity of H₂O₂ into DMF solution of Bi–PMo₁₂, the Bi–PMo₁₂ compound undergoes reoxidation. Subsequently, this solution is subjected to diffusion crystallization in ethanol, resulting in the crystallization of recrystallized Bi–PMo₁₂ (r–Bi–PMo₁₂). The recovery rate reaches 51%.

  Figure 5

  

  Figure 5.  Schematic diagram of recycling process for Bi–PMo₁₂.

  DownLoad: Full-Size Img PowerPoint

  FTIR spectra and XRD patterns show that the molecular structure of r–Bi–PMo₁₂ remains unchanged, with characteristic peaks of [PMo₁₂O₄₀]³⁻ and DMF appearing at 1060–798 cm⁻¹ and 1651 cm⁻¹, respectively. XRD patterns demonstrate no alteration in the characteristic diffraction peaks of [PMo₁₂O₄₀]³⁻ at 8.28°, 8.90°, 9.10°, and 27.9° (Figs. 6a and b). Further confirmation was obtained that Bi–PMo₁₂ and r–Bi–PMo₁₂ have the same crystal structure. As shown in Fig. 6c, the XPS spectrum of Mo 3d reveals a noticeable decrease in the peak area of Mo^V from 38.32% to 18.41%, demonstrating the successful oxidation of Mo^V to Mo^VI by H₂O₂. Furthermore, SEM testing was conducted on the surface of the reprecipitated r–Bi–PMo₁₂ crystals from DMF (Fig. 6d). On the resulted surface, the r–Bi–PMo₁₂ surface is restored to a smooth state, and the influence of N₂H₄ on its surface disappears (illustration of Fig. 6d). From Figs. 6e–h, it can be observed that r–Bi–PMo₁₂ maintains its Raman signal quenching properties during the four cycles. Furthermore, it can be observed from Fig. 6i that the Raman signal intensity at 987 cm⁻¹ still exhibits a linear relationship with the lg(CN₂H₄) with similar slopes. However, it can be observed that due to the oxidizing effect of H₂O₂, the amount of Mo^VI increases, resulting in a stronger initial Raman signal of r–Bi–PMo₁₂. The fourth attempt to measure the concentration of N₂H₄ only reached a limit of 4.5 × 10⁻⁸ ppm. Nevertheless, even after four cycles, the detection capability of Bi–PMo₁₂ remains comparable to the reported SERS technique [28-30]. Therefore, the Bi–PMo₁₂ crystalline substrate exhibits exceptional recyclability and serves as an excellent substrate for Raman detection.

  Figure 6

  

  Figure 6.  (a) Comparative FTIR and (b) Raman spectra of Bi–PMo₁₂ with different cycles. (c) XPS analysis of r–Bi–PMo₁₂ for Mo 3d spectra. (d) SEM image characterization of r–Bi–PMo₁₂. Raman spectra of r–Bi–PMo₁₂ after reaction with different concentrations of N₂H₄ (4.5 × 10¹–4.5 × 10⁻⁹ ppm): (e) First cycle; (f) second cycle, (g) third cycle and (h) fourth cycle. (i) The correlation curve between the lg(C_N2H4) of r–Bi–PMo₁₂ and the strength at 987 cm⁻¹.

  DownLoad: Full-Size Img PowerPoint

  In this study, a novel crystalline Raman substrate, Bi–PMo₁₂, free of noble metals, was successfully designed and synthesized. It demonstrates exceptional detection capabilities for N₂H₄, with a detection limit as low as 4.5 × 10⁻⁹ ppm, surpassing even the performance of the SERS technique. Through the exploration of detection mechanism, a novel RRET effect based on the absorption of Raman scattered light by reduced POMs was discovered for the first time. Simultaneously, solvation of cations during the detection process played a crucial role in the adsorption of the target molecules, inducing quenching of the substrate's Raman signals. The discovery of these two effects provides important guidance for the design and synthesis of non–noble metal Raman detection substrates, enabling the development of more Raman detection substrates based on crystalline POMs materials. This study will significantly accelerate the rapid progress of highly active non–noble metal Raman detection substrates in the domain of toxic gas detection.

  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

  The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (No. 21971085) and the Natural Science Foundation of Shandong Province (No. ZR2021MB008).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109719.

  
  
• 1. [1]

     X.D. Jin, C.Z. Liu, X.M. Wang, et al., Sens. Actuator. B: Chem. 216 (2015) 141–149. doi: 10.1016/j.snb.2015.03.088

  2. [2]

     S. Erdemir, S. Malkondu, M. Oguz, Chem. Eng. J. 468 (2023) 143767. doi: 10.1016/j.cej.2023.143767

  3. [3]

     M.G. Teng, Z.X. Zhou, Y.M. Qin, et al., Sens. Actuator. B: Chem. 311 (2020) 127914.

  4. [4]

     P.C. Wang, Y.C. Liao, Y.L. Lai, et al., Carbon 50 (2012) 1650–1658. doi: 10.1016/j.carbon.2011.12.003

  5. [5]

     B. Garg, T. Bisht, Y.C. Ling, Spectrochim. Acta A Mol. Biomol. Spectrosc. 251 (2021) 119456. doi: 10.1016/j.saa.2021.119456

  6. [6]

     P. Ji, J.Y. Li, W.S. Wang, et al., Spectrochim. Acta A Mol. Biomol. Spectrosc. 293 (2023) 122492. doi: 10.1016/j.saa.2023.122492

  7. [7]

     B.B. Wang, R.J. Yang, W. Zhao, J. Hazard. Mater. 406 (2021) 124589. doi: 10.1016/j.jhazmat.2020.124589

  8. [8]

     P. Liu, W.N. Wu, Y. Wang, Y.C. Fan, Z.H. Xu, J. Hazard. Mater. 440 (2022) 129713. doi: 10.1016/j.jhazmat.2022.129713

  9. [9]

     J.P. Liu, Y.Y. Li, J. Jiang, X.T. Huang, Dalton Trans. 39 (2010) 8693–8697. doi: 10.1039/c0dt00258e

  10. [10]

      I.V. Bagal, A. Ejaz, A. Waseem, et al., ACS Appl. Nano Mater. 3 (2020) 4394–4406. doi: 10.1021/acsanm.0c00501

  11. [11]

      K. Zuo, J. Zhang, L.T. Zeng, Spectrochim. Acta A Mol. Biomol. Spectrosc. 283 (2022) 121765. doi: 10.1016/j.saa.2022.121765

  12. [12]

      Y.Y. Lu, X. Zhang, L.Y. Zhao, et al., Nat. Commun. 14 (2023) 5860. doi: 10.1038/s41467-023-41563-5

  13. [13]

      C. Hanske, E.H. Hill, D. Vila–Liarte, et al., ACS Appl. Mater. Interfaces 11 (2019) 11763–11771. doi: 10.1021/acsami.9b00334

  14. [14]

      L.Y. Bi, H.R.C. Zhang, W.C. Hu, et al., Biosens. Bioelectron. 237 (2023) 115519. doi: 10.1016/j.bios.2023.115519

  15. [15]

      H.H. Li, M.M. Hassan, J.J. Wang, et al., Food Chem. 339 (2021) 127843. doi: 10.1016/j.foodchem.2020.127843

  16. [16]

      Y.H. Wang, S.N. Li, R.Y. Zhou, et al., Nat. Protoc. 18 (2023) 883–901. doi: 10.1038/s41596-022-00782-8

  17. [17]

      H.L. Liu, K. Zhan, K. Wang, X.H. Xia, Trends Anal. Chem. 159 (2023) 116939. doi: 10.1016/j.trac.2023.116939

  18. [18]

      S.E.J. Bell, G. Charron, E. Cortés, et al., Angew. Chem. Int. Ed. 59 (2020) 5454–5462. doi: 10.1002/anie.201908154

  19. [19]

      T. Itoh, M. Procházka, Z.C. Dong, et al., Chem. Rev. 123 (2023) 1552–1634. doi: 10.1021/acs.chemrev.2c00316

  20. [20]

      S. Cong, Y.Y. Yuan, Z.G. Chen, et al., Nat. Commun. 6 (2015) 7800. doi: 10.1038/ncomms8800

  21. [21]

      J.Q. Zhao, J.H. Wang, Y.W. Liu, et al., Biosens. Bioelectron. 216 (2022) 114660. doi: 10.1016/j.bios.2022.114660

  22. [22]

      Y.W. Liu, H. Ma, X.X. Han, B. Zhao, Mater. Horizons 8 (2021) 370–382. doi: 10.1039/D0MH01356K

  23. [23]

      S. Cong, Z. Wang, W.B. Gong, et al., Nat. Commun. 10 (2019) 678. doi: 10.1038/s41467-019-08656-6

  24. [24]

      S. Schlücker, Angew. Chem. Int. Ed. 53 (2014) 4756–4795. doi: 10.1002/anie.201205748

  25. [25]

      J. Wang, W.J. Zhu, J.Y. Zhang, et al., Sens. Actuator. B: Chem. 381 (2023) 133450. doi: 10.1016/j.snb.2023.133450

  26. [26]

      C. Peng, L.L. Jin, F. Wang, H.F. Yang, H.Y. He, Biosens. Bioelectron. 225 (2023) 115079. doi: 10.1016/j.bios.2023.115079

  27. [27]

      S.E.J. Bell, N.M.S. Sirimuthu, J. Am. Chem. Soc. 128 (2006) 15580–15581. doi: 10.1021/ja066263w

  28. [28]

      G.D. Xu, N. Guo, Q.J. Zhang, et al., J. Hazard. Mater. 424 (2022) 127303. doi: 10.1016/j.jhazmat.2021.127303

  29. [29]

      J. Wang, J.Y. Zhang, W.J. Zhu, et al., J. Mater. Chem. C 11 (2023) 3050–3058. doi: 10.1039/D2TC05148F

  30. [30]

      X. Gu, J.P. Camden, Anal. Chem. 87 (2015) 6460–6464. doi: 10.1021/acs.analchem.5b01566

  31. [31]

      M. Shafi, P.Y. Duan, W.Y. Liu, et al., Sens. Actuator. B: Chem. 380 (2023) 133410. doi: 10.1016/j.snb.2023.133410

  32. [32]

      H. Fang, C.X. Zhang, L. Liu, Y.M. Zhao, H.J. Xu, Biosens. Bioelectron. 64 (2015) 434–441. doi: 10.1016/j.bios.2014.09.053

  33. [33]

      J. Song, Y.Y. Jiang, Y.Z. Lu, et al., Small 19 (2023) 2301740. doi: 10.1002/smll.202301740

  34. [34]

      K.J. Qin, D.J. Zang, Y.G. Wei, Chin. Chem. Lett. 34 (2023) 107999. doi: 10.1016/j.cclet.2022.107999

  35. [35]

      L. Guo, L. He, Q.H. Zhuang, et al., Small 19 (2023) 2207315. doi: 10.1002/smll.202207315

  36. [36]

      S.M. Zhang, W.X. Shi, X. Wang, Science 377 (2022) 100–104. doi: 10.1126/science.abm7574

  37. [37]

      J.X. Liu, X.B. Zhang, Y.L. Li, S.L. Huang, G.Y. Yang, Coord. Chem. Rev. 414 (2020) 213260. doi: 10.1016/j.ccr.2020.213260

  38. [38]

      Y. Guan, H.P. Xiao, X.X. Li, S.T. Zheng, Polyoxometalates 2 (2023) 9140023. doi: 10.26599/pom.2023.9140023

  39. [39]

      X.X. Li, C.H. Li, M.J. Hou, et al., Nat. Commun. 14 (2023) 5025. doi: 10.1038/s41467-023-40685-0

  40. [40]

      J. Liu, N. Jiang, J.M. Lin, et al., Angew. Chem. Int. Ed. 62 (2023) e202304728. doi: 10.1002/anie.202304728

  41. [41]

      B. Li, W. Li, H.L. Li, L.X. Wu, Acc. Chem. Res. 50 (2017) 1391–1399. doi: 10.1021/acs.accounts.7b00055

  42. [42]

      C. Wang, Z.X. Gao, H.Y. Zang, T.W. Dong, Z.M. Su, Inorg. Chem. Front. 10 (2023) 3641–3647. doi: 10.1039/d3qi00451a

  43. [43]

      H.Y. Zhang, D.Y. Cui, T.Y. Shen, et al., ACS Appl. Mater. Interfaces 15 (2023) 32281–32290. doi: 10.1021/acsami.3c03161

  44. [44]

      F.R. Li, H.R. Liu, W.C. Chen, et al., Inorg. Chem. Front. 9 (2022) 3828–3838. doi: 10.1039/d2qi01003h

  45. [45]

      Q.D. Liu, H.D. Yu, Q.H. Zhang, D. Wang, X. Wang, Adv. Funct. Mater. 31 (2021) 2103561. doi: 10.1002/adfm.202103561

  46. [46]

      T.T. Zhao, N.L. Bell, G. Chisholm, et al., Energ. Environ. Sci. 16 (2023) 2603–2610. doi: 10.1039/d3ee00569k

  47. [47]

      Y. Yang, Y.M. Fu, Y.R. Tian, et al., Inorg. Chem. Front. 10 (2023) 6221–6228. doi: 10.1039/d3qi01501g

  48. [48]

      Y.F. Liu, G.D. Zeng, Y.T. Cheng, et al., Chin. Chem. Lett. 35 (2024) 108480. doi: 10.1016/j.cclet.2023.108480

  49. [49]

      K.T. Zheng, B.X. Niu, C.M. Lin, et al., Chin. Chem. Lett. 34 (2023) 107238. doi: 10.1016/j.cclet.2022.02.043

  50. [50]

      H.Q. Yin, L.L. Yang, H. Sun, et al., Chin. Chem. Lett. 34 (2023) 107337. doi: 10.1016/j.cclet.2022.03.060

  51. [51]

      G.P. Yang, K. Li, C.W. Hu, Inorg. Chem. Front. 9 (2022) 5408–5433. doi: 10.1039/d2qi01834a

  52. [52]

      L.L. Fan, M.L. Wang, X.Y. Dong, et al., Chem. Eng. J. 449 (2022)137819. doi: 10.1016/j.cej.2022.137819

  53. [53]

      J.C. Liu, J.W. Zhao, C. Streb, Y.F. Song, Coord. Chem. Rev. 471 (2022) 214734. doi: 10.1016/j.ccr.2022.214734

  54. [54]

      M. Lu, M. Zhang, J. Liu, et al., J. Am. Chem. Soc. 144 (2022) 1861–1871. doi: 10.1021/jacs.1c11987

  55. [55]

      M. Lafuente, I. Pellejero, A. Clemente, et al., ACS Appl. Mater. Interfaces 12 (2020) 36458–36467. doi: 10.1021/acsami.0c06725

  56. [56]

      A.J. Bridgeman, Chem. Eur. J. 10 (2004) 2935–2941. doi: 10.1002/chem.200305781

  57. [57]

      Y.H. Han, J.J. Lan, K. Li, et al., Chem. Asian J. 17 (2022) e202200950. doi: 10.1002/asia.202200950

  58. [58]

      J.W. Jordan, J.M. Cameron, G.A. Lowe, et al., Angew. Chem. Int. Ed. 61 (2022) e202115619. doi: 10.1002/anie.202115619

  59. [59]

      X.Z. Zhang, W.J. Zhu, Z.X. Yang, et al., J. Mater. Chem. C 10 (2022) 15451–15457. doi: 10.1039/d2tc03053e

  60. [60]

      D. Zhou, B.H. Han, Adv. Funct. Mater. 20 (2010) 2717–2722. doi: 10.1002/adfm.200902323

  61. [61]

      J.C. Liu, J.F. Wang, Q. Han, et al., Angew. Chem. Int. Ed. 60 (2021) 11153–11157. doi: 10.1002/anie.202017318

  62. [62]

      X.Q. Huang, W.L. Cui, S. Liu, et al., Chin. Chem. Lett. 34 (2023) 107692. doi: 10.1016/j.cclet.2022.07.035

  63. [63]

      L.G. Gong, X.X. Qi, K. Yu, et al., J. Mater. Chem. A 8 (2020) 5709–5720. doi: 10.1039/c9ta14103k

  64. [64]

      J.Y. Zhang, X.Y. Wang, G. Wang, et al., Chin. Chem. Lett. 34 (2023) 107231. doi: 10.1016/j.cclet.2022.02.036

  65. [65]

      G.G. Wang, T.C. Chen, C.J. Gómez–García, et al., Small 16 (2020) 2001626. doi: 10.1002/smll.202001626

  66. [66]

      Z.B. Zhang, Z.F. Li, Z.M. Dong, et al., Chin. Chem. Lett. 33 (2022) 3577–3580. doi: 10.1016/j.cclet.2022.01.062

  67. [67]

      J. Hu, F.F. Jia, Y.F. Song, Chem. Eng. J. 326 (2017) 273–280. doi: 10.1016/j.cej.2017.05.153

  68. [68]

      M. Jiang, D.D. Zhu, H.Y. Zhang, X.B. Zhao, New J. Chem. 38 (2014) 3354–3357. doi: 10.1039/C4NJ00241E

• 

  Figure 1  (a) The crystal structure diagram and (b) single unit cell illustration of Bi–PMo₁₂. (c) The XRD patterns of Bi–PMo₁₂. (d) FTIR and (e) Raman spectra of Bi–PMo₁₂ and H₃PMo₁₂ respectively. (f) SEM and (g) the element mapping images of P, Mo, Bi, and N of Bi–PMo₁₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) The histogram of Raman signal intensities at 987 cm⁻¹ of Bi–PMo₁₂ in 4.5 × 10⁻⁹ ppm N₂H₄ at different temperatures and times. (b) Raman spectra of Bi–PMo₁₂ after reaction with different concentrations of N₂H₄ (4.5 × 10¹–4.5 × 10⁻⁹ ppm). (c) The correlation curve between the logarithm of N₂H₄ concentration (lg(C_N2H4), C_N2H4 represents the concentration of the N₂H₄ vapor) of Bi–PMo₁₂ and H₃PMo₁₂ at 987 cm⁻¹ strength. (d) The histogram of Raman signal intensity of Bi–PMo₁₂ at 987 cm⁻¹ in the presence of N₂H₄ (4.5 × 10⁻⁴ ppm) and other reducing compounds (4.5 × 10⁻² ppm), respectively. The error bar is obtained at least three replicate trials.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Comparison of FTIR and (b) Raman spectra of Bi–PMo₁₂ before and after detection of N₂H₄. XPS analysis of (c) Mo 3d and (d) Bi 4f spectra of Bi–PMo₁₂ before and after reduction by 4.5 × 10⁻⁴ ppm and 4.5 ppm N₂H₄. (e) Nyquist plots of Bi–PMo₁₂ in the presence of N₂H₄. (f) UV–vis diffuse reflectance spectra (left) and colour changes (right) of Bi–PMo₁₂ with different concentrations of N₂H₄. The surface analysis of Bi–PMo₁₂ (g) before and (h, i) after reduction with N₂H₄ (4.5 × 10⁻⁴ ppm and 4.5 ppm).

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Proposed mechanism for the Raman detection of N₂H₄ by the Bi–PMo₁₂ substrate.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Schematic diagram of recycling process for Bi–PMo₁₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  (a) Comparative FTIR and (b) Raman spectra of Bi–PMo₁₂ with different cycles. (c) XPS analysis of r–Bi–PMo₁₂ for Mo 3d spectra. (d) SEM image characterization of r–Bi–PMo₁₂. Raman spectra of r–Bi–PMo₁₂ after reaction with different concentrations of N₂H₄ (4.5 × 10¹–4.5 × 10⁻⁹ ppm): (e) First cycle; (f) second cycle, (g) third cycle and (h) fourth cycle. (i) The correlation curve between the lg(C_N2H4) of r–Bi–PMo₁₂ and the strength at 987 cm⁻¹.

  下载: 全尺寸图片 幻灯片
•