Improving the surface-enhanced Raman scattering activity of carbon nitride by two-step calcining

Abstract

We demonstrated a novel two-step calcining method for growing highly efficient surface-enhanced Raman scattering (SERS) substrates, modified g-C₃N₄. Their morphologies, microstructures and optical properties were successfully controlled by altering the calcining time. Moreover, the modified g-C₃N₄-2 h + Ag substrate exhibited superior SERS activity, suggesting its potential applications in SERS sensing.

---

Graphitic carbon nitride (g-C₃N₄) has attracted widespread attention due to its outstanding and unique chemical, catalytic and optical properties. Along with its low cost, easy availability, and remarkably high stability to oxidation, g-C₃N₄ is a very attractive material for water splitting,^1,2 photocatalytic,^3,4 organic catalytic^5–7 and electrocatalytic^8,9 applications. Commonly for inorganic nanocatalysts, the size and crystal facets could significantly affect their catalytic performance. However, for carbon and g-C₃N₄ based catalysts, the morphology, microstructure and structural defects usually dominate their properties and catalytic activities. Therefore, to manipulate g-C₃N₄ catalytic and electronic performance, modifications of the g-C₃N₄ synthesis have been intensively studied.

Recently, Antonietti's group reported that the use of organic molecules, such as barbituric acid or molecules, which lead to integration of heteroatoms within the g-C₃N₄ structure, can arouse significant changes in electronic and catalytic properties.¹⁰ Using a cyanuric acid–melamine complex in ethanol as a precursor, Shalom and co-workers prepared pancake-like hollow g-C₃N₄ structures, which exhibited superior photocatalytic activity compared to the bulk material.³ Nevertheless, the obtained samples were disorganized textures with small grain sizes or the preparation process was less controlled. Moreover, in our previous study,¹¹ g-C₃N₄ was directly prepared by polymerization of melamine and the g-C₃N₄ + Ag hybrid substrate exhibited a strong SERS response, but the g-C₃N₄ + Ag hybrid substrate itself had a strong Raman signal that could interfere with the SERS detection of probe molecules. Therefore, it is essential to find new and simple synthetic pathways to prepare optimal g-C₃N₄ structures with improved SERS activity.

In this communication, the modified g-C₃N₄ substrates were simply prepared by a two-step calcining method. The morphologies, microstructures and optical properties of the modified g-C₃N₄ were comprehensively characterized. Moreover, the modified g-C₃N₄-2 h + Ag substrates showed superior SERS activity compared to g-C₃N₄ + Ag substrates.

First, the g-C₃N₄ was prepared by heating melamine to 520 °C at a rate of 10 K min⁻¹ for 4 h under an air atmosphere in a muffle furnace.¹² Then, the as-prepared yellow powder was further calcined in a chemical vapor deposition (CVD) furnace. Pure N₂ was introduced into the quartz tube at a flow rate of 50 sccm and the pressure inside the tube was held constant at 1 atm. The furnace was gradually heated to 600 °C at a rate of 10 K min⁻¹ and held for different time periods. After cooling to room temperature, a fluffy powder was obtained. The products with different calcining times in the CVD furnace were referred to as two-step calcining samples, g-C₃N₄-x h (x = 1, 2 and 3).

The samples with different calcining times were characterized by SEM. It can be clearly observed that the as-prepared g-C₃N₄ displayed a micrometer-size sheet-like structure (Fig. 1a). The two-step calcining samples, g-C₃N₄-x h (x = 1, 2 and 3) also showed sheet-like shapes, as shown in Fig. 1b–d. However, it was noteworthy that the average lateral size of thin sheets gradually widened with increased calcining time.

Fig. 1 SEM images of g-C₃N₄ (a), g-C₃N₄-1 h (b), g-C₃N₄-2 h (c) and g-C₃N₄-3 h (d).

Moreover, it was found that the XRD patterns of all g-C₃N₄-x h samples showed nearly identical peaks (Fig. 2a). The (002) peak θ = 27.3° reflected the characteristic interlayer stacking structure, whereas (100) diffraction at 13.1° indicated interplanar structural packing.¹² To investigate the slight changes in the intensity of (002) diffraction, eight different batches of samples with the same mass were measured under the same test conditions. For the same type of samples, the obtained XRD patterns from eight different batches overlapped extremely well. The relative standard deviations (RSD%) of the intensity of (002) diffraction in the XRD patterns was less than 0.9%. By comparing the XRD patterns of these four samples, it was found that the intensity of (002) diffraction related to the interlayer stacking gradually decreased with prolonged calcining time (inset of Fig. 2a), suggesting that the layer thickness became thinner. The results of SEM and XRD indicated that when the calcining time was prolonged, the average size of g-C₃N₄-x h samples became larger in the x–y plane and thinner in the longitudinal direction. Fig. 2b shows the fluorescence spectra of all samples. The fluorescence spectra of the modified g-C₃N₄-x h samples were normalized by setting the peak intensity to one unit at the intrinsic emission peaks. The intrinsic emission peak displayed a red-shift from 444 nm to 471, 476 and 480 nm, correspondingly. These spectral changes implied that the prolonged calcining time increased the conjugation length in the CN structure, resulting in the widening of average lateral size of the thin sheets, well in accordance with the SEM results (Fig. 1).

Fig. 2 XRD patterns (a) and fluorescence spectra (b) of different samples. The inset of (a) shows the relationship between the peak intensity at 27.3° and calcining time.

FT-IR spectra of all samples are shown in Fig. 3. The absorption bands located at ∼1637, 1415 and 809 cm⁻¹ originated from the C=N stretching vibration, the out-of-phase stretching vibration of N–C (sp³) and the breathing vibration of the tri-s-triazine ring,¹² respectively (Fig. 3a). In addition, there was no peaks around 2200 cm⁻¹ (–C≡N nitrile function) (Fig. 3b), which confirmed the planarity of these materials.^13–15 By comparing the FT-IR spectra of the g-C₃N₄ and two-step calcining samples, the peaks at 1573, 1327 and 1249 cm⁻¹ were shifted by 5, 6 and 8 cm⁻¹, respectively, toward low frequencies after the longer calcining process. Based on the first principles calculations, it was found that the bands near 1573, 1327 and 1249 cm⁻¹ belonged to N–C–N symmetrical stretching, asymmetrical C=N stretching and =C(sp²) bending vibration modes, respectively (Fig. 3c). It seems possible that the shifts might be caused by the electrophilic effect of the thin layer g-C₃N₄, resulting in the weakened strength of C–N covalent bonds and the lower frequencies.¹⁶ Moreover, in comparison with g-C₃N₄, the two-step calcining sample exhibits a gradually decreased absorption peak around 3172 cm⁻¹ that was indicative of secondary and primary amines.¹ This observation indicated that the further calcining process does not significantly disturb the local order of the polymer-like melamine chains, but just speaks to the formation of a more condensed polymeric network of g-C₃N₄-x h with fewer –NH₂ and –NH– motifs.

Fig. 3 FT-IR spectra (a and b) of different samples. Assignments of three vibrational modes of g-C₃N₄ within CASTEP calculations (c). Red and gray spheres represent C and N atoms, and green arrows show dipole derivative unit vectors.

To explore the SERS activity of these samples, we compared the normal Raman spectra of g-C₃N₄-x h + Ag hybrids (Fig. 4a). Many strong characteristic peaks of g-C₃N₄ were observed in the spectrum of g-C₃N₄ + Ag, and two broad peaks appeared around 1500 and 400 cm⁻¹ in the spectrum of g-C₃N₄-3 h + Ag. If they were used as the SERS substrates, their own strong signal could interfere with the SERS signals of the probe molecules. However, no obvious peak was observed in the normal Raman spectra of g-C₃N₄-1 h + Ag and g-C₃N₄-2 h + Ag, suggesting they may be ideal SERS substrates.

Fig. 4 Raman (a) and SERS (b) spectra on different substrates, SERS spectra of CV at different concentrations (c) and the relationship between CV concentrations and Raman intensity at 1619 cm⁻¹ (d) on g-C₃N₄-2 h + Ag substrate.

Moreover, Fig. 4b compares the SERS spectra of CV (2.5 × 10⁻⁶ mol L⁻¹) on different substrates. Strong SERS responses and plentiful CV peaks were observed on all substrates. In general, both electromagnetic enhancement and chemical enhancement contribute to the total SERS enhancement. Herein, in these hybrid substrates, the Ag nanoparticles immobilized on the surface and edges of modified g-C₃N₄ sheets could produce a greatly enhanced electromagnetic field (‘hot spots’), leading to a dramatically enhanced Raman response to the probe molecules near the hot spots. Moreover, there was a chemical enhancement from the charge transfer between Raman probe molecules and Ag nanoparticles, as reported in our previous study.^11,17

Although we have investigated the SERS response of the g-C₃N₄ + Ag substrate and it provided an enhancing factor as high as 4.6 × 10⁸ for CV in our previous work,¹¹ the SERS effect of g-C₃N₄ + Ag was relatively poor among these four substrates in this work. Moreover, we note that the g-C₃N₄-2 h + Ag substrate provided an enhancing factor as high as 3.0 × 10⁹, much greater than that provided by the other three substrates. This result demonstrated the strongest Raman enhancement effect on the g-C₃N₄-2 h + Ag substrate, probably due to the different morphology and formation of a more appropriate condensed polymeric network of g-C₃N₄-2 h in the two-step calcining process.

Furthermore, we investigated the SERS spectra of CV at different concentrations on g-C₃N₄-2 h + Ag (Fig. 4c) and the relationship between SERS intensity (at 1619 cm⁻¹) and the concentration of CV (Fig. 4d), which indicates that there was a wide linear response range from 2.5 × 10⁻⁸ to 2.5 × 10⁻⁶ mol L⁻¹. Therefore, the semiquantitative analysis of organic pollutants with g-C₃N₄-2 h + Ag as SERS substrates is feasible.

To gain further insight into the essential reason for the difference among the SERS activity of these four SERS substrates, the zeta potential and thermogravimetric analysis were investigated. It was found that the aqueous dispersions of g-C₃N₄ exhibited a highly negative surface charge of −20.1 mV, which was in good agreement with that reported in previous studies.¹² In addition, for the modified g-C₃N₄-x h (x = 1, 2 and 3), the zeta potential of the corresponding aqueous dispersions were −22.5, −33.6 and −24.3 mV, respectively. As far as we know, the probe molecules (aromatic molecules) were strongly adsorbed on the surfaces of g-C₃N₄ sheets and the surrounding Ag nanoparticles, due to the π–π interaction and electrostatic interaction between aromatic molecules and g-C₃N₄.¹¹ Therefore, the modified g-C₃N₄-2 h with the highest zeta potential value can absorb more probe molecules around the surface and edges, which was beneficial for increasing the concentration of probe molecules in the vicinity of the hot spots, ultimately resulting in the highest SERS activity among the different substrates. In addition, Fig. 5 shows the thermogravimetric analysis (TGA) results of melamine and the modified g-C₃N₄-x h (x = 1, 2 and 3). The melamine was relatively stable up to 290 °C, after which it started to decompose and lose mass. The TGA curve of g-C₃N₄ revealed that it was significantly robust and nonvolatile up to 600 °C, even under air. A strong endothermal peak appeared at 630 °C (Fig. 5a), paralleled by consecutive complete weight loss, indicating thermal decomposition and complete vaporization of the fragments. Furthermore, the TGA curves of the modified g-C₃N₄-x h showed some features similar to that of g-C₃N₄. Moreover, the strong endothermal peak displayed a red-shift with prolonged calcining time (Fig. 5b). Note that the thermal stability of CN materials somewhat differs between different preparation methods, which may be due to different degrees of condensation and different packing motifs.¹⁸ Therefore, the TGA results directly confirmed that the modified g-C₃N₄-x h with different calcining times possessed different degrees of polymerization, well consistent with the FT-IR results (Fig. 3). This indicated that the minute changes in morphology and degree of polymerization could change the microstructure of modified g-C₃N₄, leading to the g-C₃N₄-2 h + Ag substrate with an almost invisible Raman response. For SERS sensing, a SERS substrate that does not produce any interfering peaks would be expected to possess a lower limit of detection for the probe molecules. Therefore, by comparing the g-C₃N₄ and g-C₃N₄-3 h + Ag with strong Raman signals, the g-C₃N₄-2 h + Ag substrate showed a related low limit of detection and a wide linear response range, as shown in Fig. 4c and d. In a certain sense, this result also indicated the g-C₃N₄-2 h + Ag substrate possessed higher SERS activity. Therefore, the highest SERS activity of the g-C₃N₄-2 h + Ag substrate was partly attributed to the modified g-C₃N₄-2 h with the highest zeta potential, which may result in absorption of more probe molecules around the hot spots. The other possible reason for the highest SERS activity could be due to minute changes in morphology and a more appropriate condensed polymeric network of g-C₃N₄-2 h, which prevented the interference of SERS substrate itself in SERS sensing.

Fig. 5 Thermogravimetric analysis (TGA) results for melamine and g-C₃N₄ (a) and different modified g-C₃N₄ substrates (b).

In conclusion, we showed a path to prepare modified g-C₃N₄ materials by a two-step calcining method. It was found that the average lateral size and the degree of polymerisation of the modified g-C₃N₄ sheets were gradually widened and larger with increased calcining time. Moreover, compared to the reported g-C₃N₄ + Ag substrates, the modified g-C₃N₄-2 h + Ag substrates exhibited superior SERS activity. The utilization of a two-step calcining method, therefore, opens new opportunities for significant improvement of carbon nitride synthesis, microstructure, and SERS activity.

Preparation of modified g-C₃N₄ and SERS substrates

In brief, the g-C₃N₄ was prepared by heating melamine to 520 °C for 4 h under an air atmosphere in a muffle furnace. Then, the as-prepared yellow powder was further calcined in a CVD furnace and held this at temperature for different time periods. After the furnace cooled to room temperature, a more fluffy powder was obtained. Herein, Ag nanoparticles were prepared by the NH₂OH·HCl reduction method.¹⁷ A total of 40 mL of NaOH solution (7.5 mmol L⁻¹) was added to 50 mL of a hydroxylamine hydrochloride solution (3.0 mmol L⁻¹). Then, 10 mL of AgNO₃ aqueous solution (10 mmol L⁻¹) was rapidly added to the abovementioned mixture under ultrasound irradiation with a power of 140 W. After 5 min, a milky gray color colloid was obtained and stored in a refrigerator at 4 °C for further use. Finally, g-C₃N₄ + Ag hybrids were prepared. To prepare the hybrids, the modified g-C₃N₄ powders (0.02 g) were added to 20 mL of Ag nanoparticles colloid under vigorous stirring at room temperature. After stirring for 1 h, aliquots of the mixture were transferred into 5 mL microcentrifuge tubes and centrifuged at 5000 rpm for 5 min. The precipitates from each tube were redispersed in 5 mL of Milli-Q water. The resultant dispersion was stored at ambient temperature prior to use.¹¹ All the chemicals were of analytical grade and were used as received. Milli-Q water (18.2 MΩ cm) provided by a Milli-Q Labo apparatus (Thermo Fischer Scientific) was used in all experiments.

Characterization

The obtained samples were analyzed by scanning electron microscope (SEM, JSM-5510LVA, Japan), X-ray diffraction (XRD, D8 Advance TXS, Germany), FT-IR spectrometer (Nicolet 6700, Thermo Fischer), and fluorescence spectrometer (FP-6200, Jasco). Raman and SERS spectra were obtained on a confocal Raman spectrometer (DXR, Thermo Fischer) equipped with a diode laser of excitation of 532 nm. Measurements were taken with a laser power of 1 mW; and the samples were exposed twice and collection time was 0.2 s. Zeta potentials were measured on a Nano-ZS90 zetasizer instrument (Malvern). Thermogravimetric analyses (TGA) were performed at a heating rate of 10 K min⁻¹ under nitrogen flow (Perkin-Elmer, Pyris Diamond TGA/DTA). The geometry and IR analysis of g-C₃N₄ were performed using the plane-wave ultrasoft (PWUS) pseudo-potential method as implemented in the Cambridge Sequential Total Energy Package (CASTEP) code.¹⁹ The theoretical study was also performed using the PWUS pseudo-potential method with the generalized gradient approximation (GGA) and correlation in the Perdew–Wang 91 (PW91),²⁰ and with the local density approximation (LDA) functional of Ceperley and Alder as parameterized by Perdew and Zunger (CAPZ).²¹

Notes and references

1. X. Li, J. Zhang, X. Chen, A. Fischer, A. Thomas, M. Antonietti and X. Wang, Chem. Mater., 2011, 23, 4344–4348.
2. M. Shalom, S. Gimenez, F. Schipper, I. Herraiz-Cardona, J. Bisquert and M. Antonietti, Angew. Chem., Int. Ed., 2014, 126, 3728–3732.
3. M. Shalom, S. Inal, C. Fettkenhauer, D. Neher and M. Antonietti, J. Am. Chem. Soc., 2013, 135, 7118–7121.
4. J. Oh, S. Lee, K. Zhang, J. O. Hwang, J. Han, G. Park, S. O. Kim, J. H. Park and S. Park, Carbon, 2014, 66, 119–125.
5. X. Li, J. Chen, X. Wang, J. Sun and M. Antonietti, J. Am. Chem. Soc., 2011, 133, 8074–8077.
6. X. Chen, J. Zhang, X. Fu, M. Antonietti and X. Wang, J. Am. Chem. Soc., 2009, 131, 11658–11659.
7. X. Li, X. Wang and M. Antonietti, ACS Catal., 2012, 2, 2082–2086.
8. D. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo and J. Nakamura, Science, 2016, 351, 361–365.
9. W. Gong, J. Zou, S. Zhang, X. Zhou and J. Jiang, Electroanalysis, 2016, 28, 227–234.
10. J. Zhang, X. Chen, K. Takanabe, K. Maeda, K. Domen, J. D. Epping, X. Fu, M. Antonietti and X. Wang, Angew. Chem., Int. Ed., 2010, 49, 441–444.
11. J. Jiang, L. Zhu, J. Zou, L. Ou-yang, A. Zheng and H. Tang, Carbon, 2015, 87, 193–205.
12. J. Jiang, L. Ou-yang, L. Zhu, A. Zheng, J. Zou, X. Yi and H. Tang, Carbon, 2014, 80, 213–221.
13. Y. Zhao, Z. Liu, W. Chu, L. Song, Z. Zhang, D. Yu, Y. Tian, S. Xie and L. Sun, Adv. Mater., 2008, 20, 1777–1781.
14. A. Pfitzmann, E. Fliedner and M. Fedtke, Polym. Bull., 1994, 32, 311–317.
15. C. Wörner and R. Mülhaupt, Angew. Chem., Int. Ed., 1993, 32, 1306–1308.
16. J. Xu, L. Zhang, R. Shi and Y. Zhu, J. Mater. Chem. A, 2013, 1, 14766–14772.
17. J. Jiang, L. Ou-yang, L. Zhu, J. Zou and H. Tang, Sci. Rep., 2014, 4, 3942.
18. X. Wang, S. Blechert and M. Antonietti, ACS Catal., 2012, 2, 1596–1606.
19. M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717–2744.
20. J. P. Perdew and Y. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 45, 13244–13249.
21. D. M. Ceperley and B. J. Alder, Phys. Rev. Lett., 1980, 45, 566–569.

---