Gram-Scale Synthesis of Graphitic Carbon Nitride Quantum Dots with Ultraviolet Photoluminescence for Fe3+ Ion Detection.

A method for gram-scale synthesis of graphitic carbon nitride quantum dots (g-C3N4QDs) was developed. The weight of the g-C3N4QDs was up to 1.32 g in each run with a yield of 66 wt%, and the purity was 99.96 wt%. The results showed that g-C3N4QDs exhibit a stable and strong ultraviolet photoluminescence at a wavelength of 365 nm. More interestingly, the g-C3N4QDs can be used as a high-efficiency, sensitive, and selective fluorescent probe to detect Fe3+ with a detection limit of 0.259 μM.

1. Introduction

The graphitic carbon nitride quantum dots (g-C3N4QDs) have inspired extensive fundamental and application studies in recent years in the fields of photoelectronic devices [1], ion detection [2], photocatalysis [3], and biological imaging [4] due to their excellent stability, good water-solubility, biocompatibility, low toxicity, and excellent optical properties [5]. In recent years, researchers have carried out numerous works on the preparation of g-C3N4QDs. The synthesis methods are mainly divided into top-down synthesis and bottom-up synthesis [6]. In the top-down method, for example, Zhang et al. [7] prepared blue, fluorescent g-C3N4QDs by oxidizing bulk graphitic carbon nitride (g-C3N4) with a mixture of concentrated H2SO4 and HNO3. By combining a series of processes of acid neutralization, heat treatment, and ultrasonic treatment, Zhan et al. [8] synthesized high water solubility g-C3N4QDs with blue fluorescence by heating the mixture of g-C3N4 powder, ethanol, and concentrated KOH solution at 180 °C for 16 h. In the bottom-up method, for example, Liu et al. [9] successfully prepared g-C3N4QDs with a size of 3–5 nm by heating the mixture of urea and sodium citrate at 180 °C for 1.5 h. However, low yield and purity are still serious problems that will obstruct the widespread application of the g-C3N4QDs. Using excess organic molecules as precursors makes it especially difficult to separate g-C3N4QDs from the precursors, and the impurities inevitably remain in g-C3N4QDs. These difficult-to-remove impurities seriously restricts the accurate understanding of the intrinsic properties of g-C3N4QDs. Therefore, a simple and effective method for the synthesis of g-C3N4QDs with a high yield and purity is highly desirable. Although there are various methods to prepare g-C3N4QDs, the photoluminescence (PL) peak of g-C3N4QDs is generally located in the blue or green light region [6]. Ultraviolet fluorescence of g-C3N4QDs is rarely reported.

Here, we report a gram-scale method for the preparation of g-C3N4QDs from g-C3N4. The weight of the g-C3N4QDs is high; up to 1.32 g in each run. Typically, the yield and purity of the g-C3N4QDs is up to 66 wt% and 99.96 wt%, respectively. The results show that g-C3N4QDs exhibit a stable and strong ultraviolet PL at 365 nm with the excitation wavelengths in the range of 220 to 300 nm. More interestingly, the g-C3N4QDs can be used as a high-efficiency, sensitive, and selective fluorescent probe to detect Fe3+ with a detection limit of 0.259 μM.

2. Materials and Methods

2.1. The Preparation of g-C3N4QDs

Bulk g-C3N4 was synthesized by calcining 10 g of melamine (99.0%, Alfa Aesar, Shanghai, China) at 550 °C for 3 h in a muffle furnace with a heating rate of 0.5 °C·min−1 [10,11,12]. The g-C3N4QDs were prepared from g-C3N4, refluxed and oxidized, with concentrated nitric acid (HNO3, 65–68%) through a self-assembled experimental system [Figure 1a]. In brief, the mixture of 2 g of g-C3N4 powder and 200 mL of concentrated HNO3 were transferred into a 1000 mL round-bottom flask. Then, the mixture was heated at 135 °C in an oil bath followed by refluxing and stirring for 24 h. After cooling to room temperature, the mixture was diluted with deionized water and then evaporated to remove excess HNO3 in a rotary evaporator. Then, the obtained mixture was heated at 230 °C for approximately 3 h under a flowing Ar atmosphere (80 mL·min−1) to further evaporate HNO3, and a small pile of white solid was obtained. Thereafter, the solid was redispersed into deionized water, and then the solution was centrifuged (12,000 rpm) for 15 min to remove unoxidized g-C3N4 or big particles. The obtained supernatant was diluted and then vacuum filtered by 220 and 25 nm microporous membranes successively. Subsequently, the obtained solution of g-C3N4QDs was evaporated at 70 °C by rotary evaporators. Finally, the concentrated solution was freeze-dried by a vacuum freeze-drier to obtain 1.32 g of powdered g-C3N4QDs.

Figure 1

(a) Schematic diagram of the self-assembled experimental system. (b) TEM image and the particle size map of g-C3N4QDs.

2.2. Fluorescence Detection of Fe3+

A volume of 2.0 mL of various concentrations of Fe3+ solution (0–100 μM) was added to 2.0 mL of g-C3N4QDs solutions (0.025, 0.050, 0.075, and 0.100 mg/mL), respectively. The solution was mixed evenly and stood for 1 min to record the fluorescence emission spectra. The relative decrease of PL intensity at excitation wavelength of 247 nm was used for the quantitative analysis, wherein F0 and F are the PL intensities of the g-C3N4QDs in the absence and presence of Fe3+. In order to further verify the applicability of g-C3N4QDs as fluorescent probes to detect Fe3+ in practical applications, the PL response of g-C3N4QDs towards common cations were tested separately, including K+, Na+, Ca2+, Cd2+, Cu2+, Fe2+, Fe3+, Zn2+, and Ce4+. Selectivity tests were performed using the same procedure as the sensitivity assessment. The concentrations of Fe3+ and other metal ions taken were 100 μM.

2.3. Characterization

The morphology of each sample was investigated via a transmission electron microscopy (TEM, JEM-2100 F, JEOL, Kariya, Japan) with an acceleration voltage of 200 kV. The content of the chemical composition was measured by X-ray photoelectron spectroscopy (XPS, ESCALAB250, Thermo VG, Waltham, MA, USA) using Al Kα radiation. Fourier transform infrared (FT-IR) spectra were collected on a Brucker TENSOR 27 spectrometer. The absorption spectra were characterized by a UV-vis spectrometer (UV, UV-2700, Shimadzu, Kyoto, Japan) with a wavelength ranging from 200 nm to 800 nm. The impurity content in the sample was measured and analyzed by inductively coupled plasma (ICP, Agilent 720ES, Agilent, Waltham, MA, USA). The PL spectra of the samples were measured by a fluorescence spectrophotometer (PL, RF-5301PC, Shimadzu, Kyoto, Japan).

3. Results and Discussion

As shown in Figure 1b, the nearly spherical g-C3N4QDs were separated from each other and the diameter ranged from 4 to 12 nm. According to the statistics, the average particle size was approximately 7.2 nm. The obtained g-C3N4QDs were amorphous in structure without detectable lattices and the size was uniform. TEM and HR-TEM images of g-C3N4 are shown in Figure S1. The crystal structure of g-C3N4QDs was destroyed due to the introduction of defects. The high disorder of amorphous structure not only makes it have good elastic strain properties in a wide pH range, but also effectively reduces the exciton quenching, thus improving its fluorescence stability [13].

In order to evaluate the purity of g-C3N4QDs, the impurities’ content in the sample were measured and analyzed by ICP. It can be seen from Table 1 that the g-C3N4QDs have a low impurity content, and the content of Co, Cu, Mn, and Ni elements is less than 5 ppm. The content of Ca, Fe, and Na elements is 195 ppm, 45 ppm, and 135 ppm, respectively, which are inevitably introduced from deionized water in the preparation process. In general, the purity of g-C3N4QDs prepared by this work is up to 99.96 wt%.

Table 1

Inductively coupled plasma (ICP) analysis data of g-C3N4QDs.

Samples (ppm)	Ca	Co	Cu	Fe	Mn	Na	Ni
g-C3N4QDs	195	<5	<5	45	<5	135	<5

An XPS analysis is an effective method to understand the chemical composition and elemental chemical states of samples. The XPS survey spectra of the g-C3N4 and g-C3N4QDs samples are given in Figure 2a. It can be seen that there are three obvious peaks at the binding energies of 284 eV, 398 eV, and 532 eV of the sample, which correspond to the peaks of the binding energies of C, N, and O elements, respectively [13,14,15]. There are only a few O elements in g-C3N4, and the atomic percentage of C to N is close to 3:4. The relative content of O element in g-C3N4QDs is high; up to 26.72 at%, and it has higher oxygen-containing functional group and water solubility than g-C3N4. This is conducive to the research of ions detection based on g-C3N4QDs as a fluorescent probe [16]. Compared with g-C3N4, the content of sp2 graphitic carbon and oxygen in the C 1s spectrum of g-C3N4QDs increased [Figure 2b], indicating that the formation of sp2 graphitic carbon was promoted during oxidation and reflux. At the same time, due to the strong oxidizing property of nitric acid, more oxygen-containing functional groups were introduced. As show Figure 2c, the high-resolution N 1s XPS shows three peaks with centers about 398.2 eV, 399.3 eV, and 400.2 eV, corresponding to C-N=C, N-(C)3 and C-N-H bonds [17], respectively, indicating that the triazine unit of g-C3N4 remains in g-C3N4QDs. However, the ratio of C-N=C content in g-C3N4QDs decreased, indicating that some triazine units were damaged. The weak peak at about 404.22 eV is attributed to the charge effect or π electron delocalization in the heterocyclic ring [17]. As shown in Figure 2d, high-resolution O 1s XPS further confirmed that oxygen-containing functional groups of g-C3N4 mainly exist in the form of epoxy group (C-O-C) and carbonyl group (C=O), without carboxyl group (-COOH). The epoxy group (C-O-C) is gradually converted to carbonyl group (C=O) and carboxyl group (-COOH) in the process of oxidation and reflux. The content of each peak points of g-C3N4 and g-C3N4QDs can be seen in Tables S1–S3.

Figure 2

XPS survey spectra (a) and the high-resolution XPS spectra of (b) C 1s, (c) N 1s, and (d) O 1s of g-C3N4 and g-C3N4QDs.

As shown in Figure 3a, the FT-IR spectrum of g-C3N4QDs is mostly like that of g-C3N4. The characteristic absorption peak at 813 cm−1 represented the breathing vibration of tri-s-triazine, which is one typical out-of-plane ring bending vibration mode of g-C3N4, indicating that the obtained g-C3N4QDs have the same basic structure as g-C3N4 [2]. The peak position of 1046 cm−1 is mainly caused by the stretching vibration of the C-O bond [18]. The wide peaks in the 1300–1700 cm−1 region can be ascribed to the typical stretching vibration modes of C-N heterocycles [19]. The peaks with a wave number of 1388 cm−1 and 1731 cm−1 correspond to the typical stretching vibration modes of C-N bond and C=N bond, respectively [20]. The peak position of 1780 cm−1 is mainly characterized by C=O [21]. The broad absorption band between 3000 cm−1 and 3600 cm−1 is assigned to O-H stretching vibrations, while the sharp peak is assigned to the tensile vibration of terminal amino (N-H) [22]. These fully indicate that there are several functional groups in the sample, which is consistent with the results of the XPS analysis. Figure 3b shows the UV-vis absorption spectra of g-C3N4 and g-C3N4QDs. The main light absorption range of g-C3N4 and g-C3N4QDs are in the ultraviolet region. Compared with the bulk g-C3N4, the edge of g-C3N4QDs absorption band is blue shifted, which is attributed to the quantum size effect of g-C3N4QDs [7]. The absorption spectra of g-C3N4QDs show an obvious absorption peak at 264 nm, which is due to the π-π* electronic transition from HOMO to LUMO of the graphite carbonitride containing tri-s-triazine rings [23]. The UV-vis absorption spectrum can also confirm the existence of tri-s-triazine rings in g-C3N4QDs. Furthermore, the weak shoulder peak at 365 nm was assigned to the n-π* electronic transition of the C=N and C=O bonds in g-C3N4QDs [24].

Figure 3

FT-IR spectra of (a) and UV-vis absorption spectra of (b) of g-C3N4 and g-C3N4QDs.

Figure 4a shows the PL spectra of g-C3N4 at different excitation wavelengths. It can be seen that the emission peak of g-C3N4 is 435 nm when the excitation wavelength ranges from 220 to 400 nm. Figure 4b shows the PL spectra of g-C3N4QDs excited at a wavelength of 220–300 nm, and a strong PL emission peak at 365 nm can be seen. It can be found that the PL emission peak of the g-C3N4QDs is almost unchanged with the change of the excitation wavelength, which indicates that the g-C3N4QDs have excitation wavelength-independent PL behavior. In addition, the PL intensity of g-C3N4QDs is almost unchanged under the continuous irradiation by visible light and 365 nm for 12 h [Figure S2], indicating the g-C3N4QDs have excellent PL performance. The emission peak at 365 nm in the PL spectra is due to the π-π* transition in the unit system of the heterocyclic aromatic hydrocarbon [25], indicating that g-C3N4QDs still retains the structure of the unit ring of the heterocyclic aromatic hydrocarbon. When 3D g-C3N4 was broken into 0D g-C3N4QDs, the emission peak position showed blue shifts. Different from most of the previous g-C3N4 nanosheets and g-C3N4QDs with blue or green PL [26], the g-C3N4QDs in this work have ultraviolet PL.

Figure 4

PL spectra of g-C3N4 (a) and g-C3N4QDs (b).

As shown in Figure 5a–d, the fluorescence quenching intensity of g-C3N4QDs increases gradually with the increase of Fe3+ concentration from 0 to 100 μM. The results showed that g-C3N4QDs had different responses to Fe3+ at different concentrations. The fluorescence of the pristine g-C3N4QDs (≤0.075 mg/mL) could be completely quenched by the addition of 100 µM Fe3+. The Stern-Volmer plots of g-C3N4QDs at various concentrations for Fe3+ has good linearity for concentrations ranging from 0 to10 μM, 0–10 μM, 1–10 μM and 0–10 μM with the following equation: Y1 = (0.0544 ± 0.0026 )X1 + (0.0415 ± 0.0156)(R2 = 0.979), Y1 = (0.0745 ± 0.0020)X1 + (0.0337 ± 0.0120)(R2 = 0.993), Y1= (0.0540 ± 0.0048)X1 + (0.3010 ± 0.0299)(R2 = 0.940) and Y1 = (0.0584 ± 0.0023)X1 + (0.0315 ± 0.0134)(R2 = 0.987), respectively. These results indicated that the fluorescence response of g-C3N4QDs was best correlated with Fe3+ (0–10 μM) when the concentration of g-C3N4QDs was 0.050 mg/mL. Similarly, the Stern-Volmer plots for the fluorescence quenching of the g-C3N4QDs with Fe3+ has good linearity for concentrations ranging from 10 to 100 μM with the following equation: Y2 = (0.0036 ± 0.0003)X2 + (0.5880 ± 0.0189)(R2 = 0.945), Y2 = (0.0024 ± 0.0002)X2 + (0.7560 ± 0.0102)(R2 = 0.963), Y2 = (0.0019 ± 0.0001)X2 + (0.7840 ± 0.0069)(R2 = 0.972) and Y2 = (0.0027 ± 0.0002)X2 + (0.6000 ± 0.0127)(R2 = 0.973), respectively. These results indicate that the correlation between Fe3+ (10–100 μM) and g-C3N4QDs fluorescence response becomes better with the increase of g-C3N4QDs concentration. The limit of detection (LOD) is 0.259 μM. The detection limit is based on the Equation (1): where σ is the standard deviation of 15 replicate determinations of the blank g-C3N4QDs (Table S4); k is the slope of the calibration plot [27]. Compared with fluorescent probes based on carbon dots and GQDs (Table 2), the fluorescent probes based on g-C3N4QDs prepared in this work have obvious advantages in detection limit, reaction time and other analytical characteristics.

Figure 5

(a–d) Fluorescence responses of different concentrations of g-C3N4QDs in the presence of different concentrations of Fe3+ (excitation wavelength, 247 nm). (e–h) The relationship between the fluorescence quenched ratio and the concentration of Fe3+ from 0 to 100 μM.

Table 2

Based fluorescent probes for Fe3+ detection.

Materials	Linear Range(μmol·L−1)	Detection Limit(μmol·L−1)	Reaction Time (min)	Ref
Carbon dots	0–20	0.32	10	[28]
g-CNQDs	2–200	1	2	[25]
N-CQDs	3.32–32.26	0.7462	2	[29]
S-GQDs	0–0.7	0.0042	10	[30]
N and S doped Carbon dots	6.0–200	0.8	2	[31]
C-dots	12.5–100	9.97	-	[32]
g-C3N4QDs	0–100	0.259	1	This work

In order to further verify the applicability of g-C3N4QDs as a fluorescent probe to detect Fe3+ in practical applications, the PL response of g-C3N4QDs (0.05 mg/mL) towards common cations (100 μM) was tested separately, including K+, Na+, Ca2+, Cd2+, Cu2+, Fe2+, Fe3+, Zn2+, and Ce4+. As shown in Figure 6a, the fluorescence quenching of g-C3N4QDs by Fe3+ is strong, however the fluorescence quenching of g-C3N4QDs by other cations is weak. The fluorescence of g-C3N4QDs is quenched significantly only by Fe3+. Figure 6b shows the UV-vis absorption spectrum of common metal ions and the PL excitation and emission spectra of the g-C3N4QDs. According to the results, the wide range of absorption wavelengths of Fe3+ and PL emission peak of g-C3N4QDs (365 nm) are overlapped, which is very beneficial for fluorescent quenching. The outstanding selectivity and fluorescence quenching of the synthesized g-C3N4QDs may be ascribed to the absorption of light from g-C3N4QDs by Fe3+. The fluorescence quenching mechanism of the g-C3N4QDs in the presence of Fe3+ was proposed [Figure 6c]. When the excitation wavelength is 247 nm, the emission wavelength of g-C3N4QDs is 365 nm. In the presence of Fe3+, the light emitted by g-C3N4QDs is absorbed by Fe3+, thus causing fluorescence quenching. The excellent performance of the g-C3N4QDs reflects their potential application as a fluorescent sensor for the quantitative analysis of Fe3+ ions in an aqueous solution.

Figure 6

(a) PL response of the g-C3N4QDs (0.050 mg/mL) in the solution of common metal ions (100 μM). (b) UV-vis absorption spectrum of common metal ions and the PL excitation and emission spectra of the g-C3N4QDs. (c) Proposed fluorescence quenching mechanism of the g-C3N4QDs in the presence of Fe3+.

4. Conclusions

In summary, we reported a gram-scale method for the preparation of g-C3N4QDs. The weight of the g-C3N4QDs is up to 1.32 g in each run by using 2 g of g-C3N4 as precursor material, with a yield of 66 wt%, and the purity is 99.96 wt%. The results show that g-C3N4QDs exhibit a stable and strong ultraviolet PL at 365 nm with excitation wavelengths from 220 to 300 nm. More interestingly, using g-C3N4QDs as a fluorescent probe, Fe3+ can be detected within 1 min at the excitation/emission wavelength of 247/365 nm. The PL intensity decreased gradually in the concentration range of 0–100 μM of Fe3+. These results indicated that the fluorescence response of g-C3N4QDs was best correlated with Fe3+ (0–10 μM) when the concentration of g-C3N4QDs was 0.050 mg/mL. The linear equation is Y1 = (0.0745 ± 0.0020)X1 + (0.0337 ± 0.0120) (R2 = 0.993). The correlation between Fe3+ (10–100 μM) and g-C3N4QDs fluorescence response becomes better with the increase of g-C3N4QDs concentration. The best linear equation is Y2 = (0.0027 ± 0.0002)X2 + (0.6000 ± 0.0127) (R2 = 0.973), and the limit of detection is 0.259 μM. The high-efficiency, sensitive, and selective detection of Fe3+ was realized. All the results suggest that the obtained g-C3N4QDs with high UV fluorescence have potential applications in the fields of photoelectronic devices and ion detection.