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

Abstract

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.

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 $[(F_0-F)/F_0]$ 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. 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

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%.

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.

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 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.

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):

$$\mathrm{LOD}=\frac{3\mathsf{\sigma }}{k},$$

(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. 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

), the fluorescent probes based on g-C₃N₄QDs prepared in this work have obvious advantages in detection limit, reaction time and other analytical characteristics.

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.

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.