HRTEM images of N-CQDs (representative sample of N-0.25) show that the synthesized N-CQDs are spherical (Fig. 1a,c) with a particle size ranging from 1.78 to 6.50 n.m. The Gaussian distribution (Fig. 1b) of a sample of 150 particles shows an average particle size of 4.60 ± 0.87 nm. In addition, the N-CQDs possess a crystalline structure as indicated by the graphite lattice d-spacing of 0.22 nm (Figure 1d). Similar features were also observed for the other g-CQD samples, N-1, N-5, and N-10 analyzed by TEM (Fig. S3).
figure 1
HRTEM images of N-CQD at different magnifications and scale: (a) 20 nm, (c) 10 nm, (b) Gaussian particle size distribution histogram, (d) graphics core lattice. N-CQDs typically have a particle size of 4.60 ± 0.87 nm.
To determine the nature of the functionalization, the synthesized N-CQDs were investigated by Fourier transform infrared (FTIR) spectroscopy. The samples were classified into two groups: (i) N-CQD with lower ammonia concentration (from N-0.25 to N-1) and (ii) with higher ammonia concentration (from N -2.5 to N-10). FTIR spectra (Fig. 2) showed that all N-CQDs have hydrophilic groups on their surface such as O–H (hydroxyl) corresponding to the peak at 3389 cm−1 and N–H (3263 cm−1), confirming thus its good solubility in water. In addition, C–H (2950 cm-1), C=O (1581 cm-1), C–N (1435 cm-1) and C–O (1080 cm-1) bond vibrations were also observed in each. sample 13,14,15. Comparison of the FTIR spectra (Fig. S4) of the samples showed that increasing N doping (ammonia concentration, from N-0.25 to N-1) showed a decrease in vibration for the bond CO at 1080 cm-1. While the group of samples with a higher concentration of ammonia (N-2.5 to N-10) showed a strong vibration of the CN bond at 1435 cm-1.
Figure 2
FTIR spectra of N-CQD with lower ammonia concentration (from N-0.25) and higher ammonia concentration (N-10).
To achieve a deeper understanding of the surface characterization of N-CQDs and also to investigate the chemical composition of N-CQDs, X-ray photoelectron spectroscopy (XPS) was used. The resulting XPS spectra shown in Figure 3 were deconvoluted using Voigt functions (Lorentzian and Gaussian width) with a different inelastic background for each component16. A minimum number of components are used to obtain a comfortable fit. The binding energy scale was calibrated to the standard C 1s value of 284.6 eV. The atomic composition has been determined using the integral areas provided by the deconvolution procedure normalized to the atomic sensitivity factor (Table S1). The XPS spectrum of N-CQDs shows three typical peaks C1s (285.0 eV), N1s (399.0 eV) and O1s (531.0 eV). The fitted C1s spectrum was deconvoluted into four components, corresponding to carbon in the form of bonds C=C/CC (~ 284.4 eV), CO/CN (~ 285.8 eV), C=O (~ 287 ,3) and O=C–OH (~ 288.4 eV)17. Whereas, the N1s band showed three peaks after deconvolution which are 398.8 eV, 399.6 eV and 400.8 eV, which represent pyridinic N, N–H and amide C–N, respectively18.
Figure 3
Representative XPS spectra of N-CQD showing the lowest (N-0.25) and highest (N-10) nitrogen doped samples. The spectra show three typical peaks C1s (285.0 eV), N1s (399.0 eV) and O1s (531.0 eV). The deconvoluted N1s band showed three peaks representing pyridinic N, N–H and amide C–N.
The content of each nitrogen doping species (pyridinic, pyrrolic and graphitic) is identified and quantified from the XPS spectra of the NCQDs in order to understand their influence on the optical and chemical properties (Table S2). As commonly reported, the fluorescent property of CQDs can be enhanced by using nitrogen doping. However, only carbon-bound nitrogen can improve emission19. In addition, a higher N/C ratio was observed for N-CQD samples synthesized with a higher concentration of ammonia (Table S1). The O1s region contains three peaks at 530.9 eV, 532.2 eV and 533.3 eV for C–OH/C–O–C, C=O, H–O–H, respectively20. In addition, the oxygen content is also a key parameter in N-CQD emission, as it can maintain the balance between sp2 and sp321 carbon atoms. Therefore, Raman spectroscopy was used to investigate the disorder in the carbon bond arrangement of N-CQDs.
The Raman spectra (Fig. S5) of N-CQDs showed typical graphitic features consisting of the D mode (at 1368 cm−1) related to the symmetry transformation by defects, and the G band (at 1586 cm−1), which maps to the sp2 (graphite-like) bonds in the graphitic core. This is not surprising since the HRTEM images of N-CQD showed a typical lattice spacing of graphite (see Fig. 1d). When comparing the Raman spectra between N-CQDs, at first sight these spectra appear similar, a common ID/IG ratio of 0.95 revealed a balance between sp2 and sp3 bonds in the structure of N-CQDs. This is different from g-CQDs where an ID/IG ratio of 0.83 was observed and assigned to the carbon core (sp2 bonds)10. This is probably attributed to changes introduced by nitrogen doping that result in the transformation of C–C (sp2 bonds) into the sp3 bond between N, O, and C.
Optical properties of N-CQDs
The absorption spectra of the prepared N-CQDs measured by UV-Vis spectrophotometry are shown in Fig. 4. The N-CQD samples have a strong peak around 265 nm and a shoulder around 295 nm (Fig. 4a ). The absorption peak at 265 nm is characteristic of the π–π* transitions of the graphitic core (C=C or C–C) of the sp2 domains present in the sp3 environment, and the 295 nm is attributed to an–π* (C=O ) transitions and C–N/C=N bonds22,23. For comparison, the absorption spectrum of CQDs without nitrogen doping was also measured. It is observed that the absorption peaks related to N-CQDs are red-shifted compared to g-CQDs (synthesized from the same glucose source but without nitrogen doping), Figure 4b. These transitions are observed at 225 nm = π–π* (graphic core), and 280 nm = n–π* transitions (C=O)10. Therefore, the absorption peak observed at 295 nm in the case of N-CQDs is due to the formation of C–N/C=N bonds related to the doping effect caused by the presence of graphitic nitrogen24,25.
Figure 4
(a) UV-Vis absorption spectra of (a) N-CQD and (b) g-CQD without nitrogen doping. The presence of C–N/C=N bonds is observed at 295 nm.
The photoluminescence (PL) spectra of the prepared N-CQDs were measured using a range of different excitation wavelengths, as shown in Fig. 5. The PL emission of each sample clearly showed the PL dependent of excitation which is beneficial for a variety of applications. such as biosensors, bioimaging or LED devices26,27. The PL emission peaks were shifted when different excitation wavelengths were applied, and each sample exhibited an optimal excitation wavelength. Overall, the PL study revealed interesting optical properties of N-CQDs. First, the PL results are consistent with previous reports where the excitation-dependent emission phenomenon of CQD28 was observed. Second, the maximum excitation wavelengths varied from 360 to 320 nm with ammonia concentration.
Figure 5
Photoluminescence spectra of CQDs with and without nitrogen doping measured using excitation wavelengths in the range of 300–500 nm, (to) g-CQD (without nitrogen doping); (b) N-0.25; (b) N-2.5, (d) N-10.
However, the mechanisms behind the excitation-dependent properties of CQDs are unclear. One of the most comprehensive and widely accepted mechanisms to interpret the excitation-dependent PL of CQDs is the quantum confinement effect also known as the size effect14,21,28,29. In general, CQDs possess broad particle size distributions leading to a range of different energy gaps and is the reason for the variation in emission wavelengths30,31. But here, HRTEM image data analyzes confirmed that increasing the amount of nitrogen doping did not contribute to an increase in particle size for the as-prepared samples. Therefore, the observed redshift character can be attributed to the radiative recombination of eh pairs hosted in the sp232 clusters. Apart from the quantum confinement effect, the theory of surface states is quite widely adopted to interpret the excitation-dependent PL behavior of CQDs33,34,35. UV-Vis absorbance showed that the peak of N-CQDs at 265 nm is related to the π-π* transition, suggesting the existence of a large number of π electrons. Surface electronic states can be conjugated with these π electrons as a result of surface oxidation which results in modification of the electronic structure of N-CQD34,36.
To interpret the mechanism of this effect, the PL lifetime and PLQY of N-CQDs were measured. The results obtained (Table 1) showed an increase in both PL lifetime and PLQY after nitrogen doping and the highest lifetime and PLQY values were obtained for \(\left[N\right]\ge 7.5 M\). The obtained PLQY value of 9.6%\(\pm\) 0.9 for N-10 is a significant improvement compared to g-CQDs which showed PLQY <1%10. These results are comparable to the literature (shown in Table S3), where CQD and N-CQD were synthesized using different methodologies.
Table 1 The photoluminescence quantum yield (PLQY), half-life, 1/e lifetime, radiative (kr) and non-radiative (knr) rates of N-CQDs.
The radiative rate (kr) and non-radiative rate (knr) were calculated using Eq. (2) and (3) 37.
$${k}_{r}=\frac{\Phi }{{\tau }_{1/e}}$$
(2)
$$\Phi =\frac{{k}_{r}}{{k}_{r}+{k}_{nr}}$$
(3)
where \(\Phi\) is PLQY of N-CQD and \({\tau }_{1/e}\) corresponds to the lifetime when the fluorescence falls 1/e from its initial value.
Table 1 and Figure 6 show that when a higher concentration of ammonia was used, the non-radiative rates were significantly reduced. This is due to surface coating activities…