Introduction

A large number of ions released from wastewater have become a critical worldwide issue due to the severe hazards to the environment and organisms [1, 2]. Industrial wastes are important sources of Hg2+ pollution and exposure to mercury may cause damages to brain, kidney and neurological systems [3,4,5,6]. On the other hand, sulfide, as one of the highly toxic anion, irritates mucous membranes and in higher concentration may cause unconsciousness and respiratory paralysis [7, 8]. Conventional analytical techniques for ions detection include atomic absorption spectroscopy, inductively-coupled plasma mass spectrometry, and electrochemistry. These methods either are sensitive and selective but expensive, or are easy to operate but weak in sensitivity and selectivity. Therefore, there is an urgent need to develop novel approaches that are convenient and fast as well as sensitive and selective. So far, fluorescence spectroscopy has been a powerful optical technology that can provide a good flexible, sensitive, and simple detection method [9,10,11,12].

Carbon nitride as an organic semiconductor consists of carbon and nitrogen which is a promising candidate in carbon materials for applications in many fields [13, 14]. There are several allotropes of C3N4 such as α-C3N4, β-C3N4, pseudo-cubic C3N4, cubic C3N4 and g-C3N4. The history of carbon nitrides can be traced back to 1834 [15]. However, researchers became interested in it in 1990s owing to that β-C3N4 was predicted to have extremely high hardness values [16]. In fact, at ambient conditions, graphitic carbon nitride (g-C3N4) is the most stable allotrope among various carbon nitrides nanomaterials. g-C3N4 was first applied in water splitting as a metal free conjugated semiconductor photocatalysis [17, 18]. Since then, researchers have been paying more and more attention to g-C3N4. Owing to the strong C-N covalent bonding in-plane direction and weak van der Waals interactions between layered (tri-s-) triazine units, g-C3N4 can be easily synthesized into bulk, nanosheets, nanotubes, quantum dots or other nanostructures [19, 20]. Furthermore, a new type of fluorescent materials, graphitic carbon nitride quantum dots (g-CNQDs) have gained tremendous attentions. The g-CNQDs emerge as a novel fluorescent probe for biological and environment detection because they show bright fluorescence, good water solubility and biocompatibility, low cost, and low cytotoxicity [21,22,23]. For instance, Achadu et al. prepared g-CNQDs and their 2,2,6,6-tetramethyl (piperidin-1-yl)oxyl derivatives as a “turn off/on” fluorescence probe for ascorbic acid detection [24]. Yin et al. reported a novel one-pot evaporation-condensation strategy to synthesize g-CNQDs which were used as an efficient probe for Fe3+ trace analysis and live-cell imaging [25]. Wang et al. constructed an electrochemiluminescence and fluorescence sensor to detect riboflavin based on the resonance energy transfer between donor g-CNQDs and receptor riboflavin [26].

In this work, we report on the low temperature solid-state synthesis of g-CNQDs. Then g-CNQDs was applied as a fluorescence “off-on” probe for detection of Hg2+ and S2− in an aqueous solution. Significant fluorescence quenching of g-CNQDs occurs upon addition of different concentrations of Hg2+ while S2− can recover the fluorescence of g-CNQDs. As such, g-CNQDs can serve as an effective fluorescent sensing probe for detection of Hg2+ and S2− with high sensitivity and selectivity. The reaction mechanism of the “off-on” process has been preliminarily discussed. Finally, g-CNQDs has been successfully applied to the determination of Hg2+ and S2− in various spiked water samples.

Experimental

Materials

Sodium citrate dehydrate, urea, mercuric chloride, sodium sulfide nonahydrate, absolute ethanol were purchased from Aladdin Chemical (Shanghai, China, http://www.aladdin-e.com/). All chemicals of analytical grade reagents were used without further purification.

Apparatus and characterization

Transmission electron microscopic (TEM) image was taken using a JEOL 2010-H TEM (Tokyo, Japan). X-ray photoelectron spectra (XPS) were carried out with an AXIS ULTRADLD X-ray photoelectron spectrometer (Kratos, Tokyo, Japan). X-ray diffraction (XRD) was performed on a D8 Advance power X-ray diffractometer (Bremen, Germany). Fourier transform infrared spectrum (FT-IR) of g-CNQDs was measured on a Varian FT-IR-640 spectrometer (Agilent Technologies, Palo Alto, CA, US). UV–vis absorption and fluorescence spectra of g-CNQDs were recorded on a Metash 6100 UV/VIS spectrophotometer (Shanghai, China) and a Hitachi F-4500 spectrofluorometer (Tokyo, Japan), respectively. The fluorescence lifetimes were performed on a FLS920 steady-state transient fluorescence spectrometer (Edinburgh, UK).

Preparation of g-CNQDs

The synthesis of g-CNQDs followed the previous reported procedure with slight modifications [27]. Details of the synthesis are deposited in the Electronic Supporting Material.

“Turn off” detection of mercury(II)

Various concentrations of Hg2+ and phosphate buffer (10 mM, pH 8.0) were added to 3.0 mL g-CNQD solutions (25 μL, 6.5 mg/L). Afterwards, the mixture solutions were incubated for 8 min at ambient conditions. Fluorescence emission spectra were recorded at an excitation wavelength of 390 nm and the fluorescence intensities were recorded at excitation/emission wavelengths of 390/450 nm. The slit widths of excitation and emission were both set as 10 nm.

“Turn on” detection of sulfide

The g-CNQDs solution (25 μL, 6.5 mg/L) was mixed with Hg2+ (13.3 μM), different concentrations of S2− and phosphate buffer (10 mM, pH 8.0) were added to 3.0 mL and incubated for 6 min. The fluorescence emission spectra were recorded at an excitation wavelength of 390 nm. The fluorescence intensities were recorded at excitation/emission wavelengths of 390/450 nm. The slit widths of excitation and emission were both 10 nm.

Water samples pre-treatment and analysis

The lake water was from Yingze Park of Taiyuan, Shanxi Province, China. Tap water was collected from our laboratory. Wastewater was collected from an industry of Taiyuan city. Lake water and tap water samples were both first filtered and then centrifuged to remove large solids and debris. Wastewater was boiled for 10 min first, filtered through a 0.22-μm membrane filter and diluted to 10 times with phosphate buffer (10 mM, pH 8.0). The analyses of Hg2+ and S2− in water samples were performed as described in sections ““Turn off” detection of mercury(II)” and ““Turn on” detection of sulfide”.

Results and discussion

Total S(II) concentration in Na2S solution consists of S2−, HS− and H2S. There are two kinds of equilibrium in Na2S solution:

$$ {\displaystyle \begin{array}{l}{\mathrm{S}}^{2^{\hbox{--} }}+{\mathrm{H}}^{+}\rightleftharpoons {\mathrm{H}\mathrm{S}}^{\hbox{--}}\\ {}{\mathrm{H}\mathrm{S}}^{\hbox{--} }+{\mathrm{H}}^{+}\rightleftharpoons {\mathrm{H}}_2\mathrm{S}\end{array}} $$

Under acidic conditions, S2− is protonated and converted to HS− and H2S. In alkaline condition, the equilibrium shifts to the left due to the existence of more OH−. Both HS− and S2− exist in the solution when pH is 8.0 [28,29,30].

Choice of materials

Fluorescent semiconductor quantum dots (QDs) have shown attractive potential owing to their lots of promising applications, especially in ion detection, cell imaging and biosensing. Nevertheless, the QDs such as CdS and CdSe QDs usually suffer from their toxicity, health and environment hazards because of the heavy metals. Therefore, it is urgent to develop a kind of green, low toxicity, and environmentally QDs. g-CNQDs have attracted much attention because of their simpler synthesis method, low toxicity, good water-solubility, biocompatibility and chemical stability compared with other light-emitting quantum dots.

Characterization of g-CNQDs

The characterization methods for g-CNQDs were measured using TEM, XPS, XRD, FT-IR, UV-visible and fluorescence spectroscopy. Respective data and figures are displayed in the Electronic Supporting Material.

Optimization of method

The following parameters were optimized: (a) Sample pH, (b) Reaction time, (c) Concentration of g-CNQDs (d) Fluorescence quenching concentration of Hg2+; (e) Temperature of Hg2+ detection. Respective data and figures are deposited in the Electronic Supporting Material. The following experimental conditions were found to give the best results: (a) Sample pH 8.0, (b) The optimal fluorescence quenching and recovery reaction times are 8 min and 6 min, respectively, (c) The optimal concentration of g-CNQDs is 6.5 mg/L; (d) The optimal fluorescence quenching concentration of Hg2+ is 13.3 μM; (e) The working temperature of Hg2+ detection is 298 K.

“Turn-off” Hg2+ detection

The fluorescence behaviours of the prepared g-CNQDs toward different concentrations of Hg2+ were investigated. Figure 1a indicates that the fluorescence intensity of g-CNQDs is obviously reduced upon increasing the Hg2+ concentration, demonstrating the fluorescence “turn-off” detection of Hg2+. The inset of Fig. 1a shows that the fluorescence intensity against the concentration of Hg2+ (0.20–21 μM) is linear and the limit of detection (LOD) is 3.3 nM. The linear equation is 1/(F0–F1) = 0.01018/C + 0.00117 with the correlation coefficient (r) of 0.9999, where F0 and F1 are the fluorescence intensities of g-CNQDs in the absence and presence of Hg2+, and C is the concentration of Hg2+. The detection limit is based on the equation LOD = 3σ/k, where σ is the standard deviation of 11 replicate determinations of the blank g-CNQDs and k is the slope of the calibration plot.

Fig. 1
figure 1

a Effect of Hg2+ concentration on the fluorescence spectrum of g-CNQDs. The concentrations of Hg2+ are 0.0, 1.2, 2.4, 3.6, 4.8, 6.0, 7.2, 8.4, 9.6, 10.8, 12.0, 13.2, 14.4, 15.6, 16.8, and 18.0 μM from 1 to 16. The inset displays the Lineweaver–Burk plot for the g-CNQDs and Hg2+ concentration. b Effect of S2− concentration on the fluorescence recovery of g-CNQDs. The concentrations of S2− are 0.0, 8.0, 10.0, 12.0, 14.0, 16.0, 18.0, 20.0, 22.0, 24.0, 26.0, 28.0, 30.0, 32.0, 34.0, 36.0, and 38.0 μM from 1 to 17. The inset displays the linear relationship between the F2/F1 and S2− concentration. The excitation and emission wavelength are at 390 and 450 nm, respectively

Full size image

A comparison of the detection limits and linear ranges of different methods for Hg2+detection is shown in Table 1. It reveals that although several reported methods have lower LODs, our proposed method is still comparable to most other methods. But our materials for synthesizing g-CNQDs are relatively green and environmentally.

Table 1 An overview on recently reported nanomaterial-based optical methods for the determination of Hg2+
Full size table

“Turn-on” sulfide detection

Since Hg2+ has a great affinity towards S2− [32, 33], their strong coordination can recover the quenched fluorescence of g-CNQDs. Owing to the phenomenon of fluorescence recovery, g-CNQDs-Hg2+ system can be employed for sensitive and selective detection of S2−. The fluorescence recovery response of g-CNQDs-Hg2+ system was analysed at different concentrations of S2− under the optimal experimental conditions. The fluorescence intensity of g-CNQDs is recovered gradually with increasing the S2− concentration. A linear relationship is established between the g-CNQDs fluorescence intensity and the concentration of S2− ranging 8.0–45 μM (Inset: Fig. 1b) and the LOD is found to be 21.7 nM toward S2−, where LOD is determined from 3σ/k with σ as the standard deviation of 11 replicate determinations of the g-CNQDs-Hg2+ and k as the slope of the calibration plot. The linear equation for S2− is F2/F1 = 0.02782C + 0.7516, where F1 and F2 are the fluorescence intensities in the absence and presence of S2−, respectively. The linear range and LOD for S2− of this work are compared with other methods and displayed in Table 2. It can be seen that the LOD of our work is better than most other methods, and its sensitivity is good enough to detect sulfide in real water samples.

Table 2 An overview on recently reported nanomaterial-based optical methods for the determination of sulfide
Full size table

Selectivity study

To further verify the applicability of the fluorescence probe for detecting Hg2+ and S2− in practical applications, the effect of common ions was investigated. Figure 2a represents the interference of different cations to Hg2+ detection. In the control experiment, the concentrations of Hg2+ and each cation were 3.0 and 100.0 μM, respectively. Those cations have negligible interference on the quenching fluorescence of g-CNQDs by Hg2+. These results suggest that the g-CNQDs fluorescence probe for Hg2+ has high selectivity.

Fig. 2
figure 2

a The response of g-CNQDs and Hg2+ in the absence and presence of other cations. b Fluorescence response of g-CNQDs-Hg2+ in the presence of different anions. The concentrations of ions are 100 μM

Full size image

Analyses were carried out to investigate the interference effects of common anions on the recovery of g-CNQDs by S2− in Fig. 2b. The concentration of S2− and each anion were 10.0 and 100.0 μM, respectively. No significant change is observed in comparison to the blank. As such, the g-CNQDs-Hg2+ system is suitable for analysis of S2− in the presence of other anions.

Mechanism analysis

Fluorescence quenching usually originates from static and/or dynamic quenching. Static quenching involves the formation of the complex by combining the ground state photoluminescent molecule with the quencher, while dynamic quenching refers to the collision of excited photoluminescent molecule with the quencher [48, 49]. In general, the static and dynamic quenching process can be analysed by Lineweaver-Burk equation Eq. (1) and Stern-Volmer equation Eq. (2), respectively:

$$ \frac{1}{{\mathrm{F}}_0-\mathrm{F}}=\frac{1}{{\mathrm{F}}_0}+\frac{1}{{\mathrm{F}}_0{\mathrm{K}}_{\mathrm{LB}}{\mathrm{C}}_{\mathrm{q}}} $$
(1)
$$ \frac{{\mathrm{F}}_0}{\mathrm{F}}=1+{\mathrm{K}}_{\mathrm{SV}}{\mathrm{C}}_{\mathrm{q}} $$
(2)

where F0 and F are the fluorescence intensities of the fluorescence molecule with and without the quencher, respectively. Cq is the quencher concentration. KLB is the Lineweaver-Burk constant and KSV is the Stern-Volmer quenching constant. The Lineweaver-Burk plot can be used to determine if there is a linear relationship between the concentration of the metal ion of interest (Hg2+) and the change of fluorescence intensity. KLB is the static quenching constant, reflecting the efficiency of quenching or the accessibility of the fluorophores to the quencher. The sensitivity of the determination is directly related to the KLB value [50]. The fluorescence quenching of g-CNQDs by Hg2+ ion fits well with the Lineweaver-Burk equation as displayed in the inset of Fig. 1a. Fig. S10 shows that the absorption peak of g-CNQDs at 390 nm gradually decreases with the increase in Hg2+ concentration, suggesting that static quenching may play a major role in the interaction of Hg2+ with g-CNQDs. In addition, the fluorescence lifetimes of g-CNQDs in the presence of various concentrations of Hg2+ were investigated as depicted in Fig. 3. There is no change of the g-CNQDs lifetime, inferring that the quenching process may be governed by the static mechanism.

Fig. 3
figure 3

Time-resolved fluorescence spectra of g-CNQDs, g-CNQDs-Hg2+, and g-CNQDs-Hg2+-S2−

Full size image

There are lots of amino and hydroxyl groups on the surface of g-CNQDs. The selectivity and specificity of g-CNQDs to Hg2+ can probably attributed to the interaction between Hg2+ and the imine “N” (-C=N-C) or hydroxyl groups of g-CNQDs, making g-CNQDs to come close with each other and leading to a decrease of g-CNQDs fluorescence [51,52,53]. Thus, g-CNQDs can act as a “turn-off” fluorescent probe for Hg2+. On the other hand, the fluorescence recovery of g-CNQDs-Hg2+ system by adding S2− is ascribed to the competitive binding of S2− with Hg2+. S2− can extract Hg2+ from g-CNQDs surface to form a stable complex, resulting in the fluorescence “turn-on” of g-CNQDs. In this way, the g-CNQDs-Hg2+ system can function as a “turn-on” probe for S2−. The mechanism of the ‘off-on’ process is illustrated in Scheme 1.

Scheme 1
scheme 1

Schematic illustration of detection of Hg2+ and S2−

Full size image

Analysis of Hg2+ and S2− in water samples

The g-CNQDs probe was applied for analysis of Hg2+ and S2− in water samples. Different concentrations of Hg2+ (1.00, 9.00, and 18.00 μM) were added to tap water, lake water and wastewater samples. The recovery of water samples ranges from 99.0 to 110.0% as depicted in Table 3. The relative standard deviation (RSD) is less than 5%, suggesting that the analytical performance for the detection of Hg2+ in water samples is satisfactory. Moreover, spiked water samples of various concentrations of S2− (10.00, 20.00, and 30.00 μM) were also measured and the results are shown in Table 4. The recovery ranges from 98.0 to 110.0%, and the RSD is less than 5%. These results indicate that g-CNQDs may be a promising probe in real samples.

Table 3 Analytical results of Hg2+ in water samples
Full size table
Table 4 Analytical results of S2− in water samples
Full size table

Biomatter, such as microorganism and substance produced by them, in wastewater may absorb UV light, which may display background fluorescence under UV excitation. Therefore, wastewater samples need to be pretreated before detecting ions, for instance, boiling and microfiltration. The pretreatment of water samples may bring inconvenience to detection. In brief, this method has some limitations. The fluorescence probe may not be suitable for biological samples owing to the need for working in the UV excitation.

Conclusion

In this work, a probe for the analyses of Hg2+ and S2− is described. It is based on the “turn off-on” fluorescence phenomenon of g-CNQDs. Hg2+ induces significant fluorescence quenching of g-CNQDs. On the other hand, Hg2+ combines with sulfide and thereby recovers the fluorescence of g-CNQDs. The fluorescence ‘off-on’ probe enables rapid detection of Hg2+ and S2− with high sensitivity and selectivity. This probe has been applied to various spiked water samples to detect Hg2+ and S2− with satisfactory results. It is believed that the present strategy may offer a new approach for developing rapid, low-cost, highly sensitive and selective probe for real sample analyses.