Facilely synthesised sulphur-doped carbon dots for highly selective determination of picric acid and for biological imaging

ABSTRACT Here, yellow fluorescent sulphur-doped carbon dots (S-CDs) with high quantum yield (QY) of ~27.9% were synthesised from m-cresol purple by a fast and low-cost one-step hydrothermal method without other reagents. The surface morphology, functional groups and optical properties of S-CDs were characterised by high-resolution transmission electron microscope (HRTEM), X-ray photoelectron spectroscopy (XPS), Fourier transform infrared (FTIR) spectrum and ultraviolet visible (UV-vis) absorption. The fluorescence intensity of S-CDs can be specifically and stably quenched by picric acid (PA) through inner filtration effect (IFE). The linearity between fluorescent intensity and PA concentration was used for developing a fluorescent probe with detecting range of 0–259.0 μM and a detection limit as low as 14.6 nM. We developed the method by applying S-CDs to determine PA in lake water with recoveries in the range of 93.0–103.7%, which verified the feasibility of this probe in practical analysis. In addition, due to the low toxicity and good biocompatibility, S-CDs can be applied to biological imaging.


Introduction
2,4,6-Trinitrophenol, known as picric acid (PA), is a dangerous nitro aromatic compound with a strong electron withdrawing group.PA is widely used in the production of dyes, drugs and pesticides, and it plays a considerable role in the chemical industry [1].PA is an intermediate of dynamite, but its explosive power is bigger than 2,4,6-trinitrotoluene (TNT) [2].In addition, due to its phenolic hydroxyl and nitro functional groups, PA is toxic and poorly biodegradable.If it is accidentally discharged into the environmental water, it will cause serious environmental pollution problems [3], and the high concentration of PA can cause skin irritation.It reported that it is considered to be a possible carcinogen, which is harmful to the human centre, cardiovascular, liver, etc. [4].Therefore, quantitative determination of PA is essential.To date, the common methods for detecting PA include ultraviolet-visible spectroscopy (UV) [5], mass spectrum (MS) [6], surface-enhanced Raman spectroscopy (SERS) [7], gas chromatography (GC) [8] and high performance liquid chromatography (HPLC) [9], but these methods require expensive instruments, complex sample preparation and experimental procedures.Instead, fluorescence analysis is a fast, simple, accurate and cheap method for the trace analysis.Recently, diverse fluorescence nanomaterials, including light-emitting two-dimensional metal organic frameworks [10], organic fluorescent probes [11], nanoparticle fluorescent sensors [12], nanorods [13], lanthanide complexes [14] and nitrogen-doped carbon quantum dots [15], have been used to determine PA.
Carbon quantum dots, or carbon dots (CDs), are emerging fluorescent nanomaterials with a size of smaller than 10 nm and a high degree of monodispersion.Compared with other nanomaterials, CDs have unique properties, such as a wide range of sources, simple preparation, good solubility, tunable photoluminescence and excellent biological safety [16,17].Based on these advantages, CDs have a wide range of applications, including small-molecule detection [18], biological imaging [19], drug delivery [20], electrochemical sensor [21], anti-cancer [22], optoelectronic devices [23,24], etc.The preparation methods of CDs are generally classified into two groups: top-down and bottom-up approaches.The top-down synthesis method for the preparation of CDs is using the large carbon source as a raw material, such as electrochemical oxidation andlaser ablation [25].The other way is the bottom-up synthesis to let small organic molecules fusion and reconstruction, such as thermal decomposition and microwave-assisted method [16,26].Among them, hydrothermal method is the most popular.The synthetic materials include small organic molecules and organic natural substances [27].Heteroatom doping has been testified to be an effective method to modify the surface of CDs to enhance their photoluminescence and expand their applications [28].In recent years, various heteroatoms (such as nitrogen, phosphorus, sulphur and boron) have been reported to dope CDs Scheme 1. Clearly illustrate the synthesis and application of the S-CDs.[29][30][31][32], but most of them require modification by foreign chemicals.Self-doping of nitrogen and sulphur appears to be the most attractive and promising [33].Furthermore, the use of organic molecules as carbon sources can successfully meet this requirement.
In this paper, a single cheap raw material m-cresol purple was used to synthesise S-CDs with a quantum yield of 27.9% by the hydrothermal synthesis.The best excitation and emission wavelength of S-CDs are 400 nm and 553 nm, respectively.The CIE chromaticity diagram (0.38, 0.58) further proves that it emits yellow light.S-CDs have good linearity (R 2 ¼ 0.997), high sensitivity (detection limit as low as 14.6 nM), wide linear range (0-259.0μM), and have been successfully used for the detection of PA in actual samples.(The recovery rate is between 93.0-103.7%,RSD < 3%).Furthermore, S-CDs are a promising candidate material for the bioimaging.

Materials
M-cresol purple (Shanghai Reagent No. 3 Factory), absolute ethanol (Tianjin Zhiyuan Chemical Reagent Co., Ltd., China), H 3 P0 4 , H 3 BO 3 and HAc (Chongqing Chuandong Chemical (group) Co., Ltd., China), and other reagents (such as NaCl and PA) were purchased from Aladdin (Shanghai, China).All chemical reagents used in the experiment were of analytical grade.Britton-Robinson (BR) buffer solution was used for pH adjustment.The water used in the whole experiment was ultrapure water (18.2MΩcm).Caenorhabditis elegans (C.elegans) and Arabidopsis thaliana (A.thaliana) were provided by the Institute of Biomedicine, Shanxi University.

Preparation of S-CDs
0.3 g m-cresol purple was dissolved in 15 mL absolute ethanol and 15 mL water.Then, it was transferred to a 50 mL polytetrafluoroethylene reactor and heated at 200°C for 4 h.After being naturally cooled to room temperature, the obtained solution was centrifuged at 10,000 rpm for 30 mins, and a portion of the supernatant was freeze-dried to obtain powder for characterisation.The remaining supernatant was placed in a refrigerator at 4°C for later use.

Quantum yield measurement
Rhodamine 6 G (quantum yield is 0.94, dissolved in ethanol) is a standard reference substance for measuring the quantum yield of S-CDs, and equation (1) is the calculation formula for quantum yield.

Determination of PA with S-CDs
At room temperature, 10 μL of S-CDs solution was mixed with 2 mL of water.Subsequently, PA (20 μL) standard solution with different concentrations was pipetted into the system, and the mixture was gently shaken for 10 s.The mixture was then moved into a quartz cuvette, and its fluorescence spectrum was measured (λ ex = 400 nm, λ em = 553 nm).The selectivity of PA detection was explored by the addition of other interferences following the same procedure.All experiments were repeated at least three times.

Analysis of PA in real water samples
To demonstrate the application of the method we proposed, a fluorescent system comprising the S-CDs was utilised to analyse PA in lake water.Lake water was collected from Lingde lake in Shanxi University.In brief, lake water was filtered through membrane (0.22 μm) to remove impurities.A series of known concentrations (1 μM, 3 μM and 5 μM) of PA were added to the filtered lake water sample, and then analysed with the description in 2.4 Determination of PA with S-CDs.Each experiment was performed five times.

Cytotoxicity experiment
The procedure of the biological toxicity test is as follows.A stock solution of S-CDs with different concentrations (0, 12.5, 25, 50, 100, 200 and 300 mg mL −1 ) were prepared.C. elegans and A. thaliana were incubated with S-CDs solutions in the different confocal dishes for 24 h.Based on the mortality data, the half-lethal concentration was calculated, and the appropriate S-CDs concentration was selected for biological imaging.

Biological imaging
The C. elegans were seeded in confocal dishes and cultured with S-CDs solution (50 mg mL −1 ) for 30 mins at 23°C, and then dishes were washed with saline to remove the remaining S-CDs solution.Immediately, confocal fluorescence imaging was performed under the excitation of 405 nm laser.Also, we performed the imaging of C. elegans, which co-incubation with PA solution of different concentrations.The C. elegans incubated with S-CDs solution (50 mg mL −1 ) for 30 minutes were placed on three glass slides, respectively.Afterwards, 2 μL PA solution of different concentrations (0.025 M, 0.05 M, 0.1 M) was respectively added to glass slide, kept for 10 minutes, and then shared confocal fluorescence imaging (C PA = 0.025 M, 0.05 M, 0.1 M).The imaging procedure for A. thaliana is similar to that for C. elegans, but just incubated with 0.1 M PA solution.

Characterisation of S-CDs
TEM images suggest that S-CDs are uniformly monodispersed spherical nanoparticles (Figure 1a).The average diameter of S-CDs was 3.18 nm (Figure 1b).The HRTEM displays a particular crystal structure with a lattice spacing of 0.2 nm, which corresponds to the (100) diffraction planes of graphitic structure [34] (Figure 1a).The surface functional groups of S-CDs were investigated by the FT-IR spectroscopy.As shown in Figure 2a, the characteristic peak near 3435 cm −1 belongs to O-H stretching vibration.The two peaks near 2975 cm −1 and 2915 cm −1 were attributed to C-H stretching vibration.The peaks at 1447 cm −1 and 1020 cm −1 revealed the presence of C-OH and C-O, respectively.632 cm −1 was the stretching vibration of C-S.The peak at 1179 cm −1 represents the presence of -SO 3  or C = S on the surface of S-CDs [33,35].The structure of S-CDs was analysed by XRD.The XRD pattern of the S-CDs has an obvious bulging peak, and the peak position is about 2 θ = 20.2°(Figure 2b).Therefore, it was ascribed to the (002) lattice plane of carbon based materials with some of the disordered carbon atoms.The satisfactory degree of graphitisation was assigned to the introduction of sulphur-containing and oxygen-containing groups on the surface of S-CDs [36].In the Raman spectrum (Figure 2c), distinct peaks at 1376 and 1588 cm −1 were attributed to the disordered D band and the crystalline G band, respectively [37].
The surface element composition of S-CDs was characterised by XPS.As manifested in Figure 2d, the three peaks centred at 166.1, 287.1 and 534.2 eV correspond to S 2p (3.00%), C 1s (80.44%) and O 1s (16.51%), respectively.The C 1s spectrum (Fig. S1a) described three peaks at 284.6, 285.2 and 286.9 eV, which belong to C-C, C-S and C-O, respectively.Two characteristic peaks near 168.3 and 169.4 eV depicted in S 2p spectrum (Fig. S1b) had proved the existence of S = O and C-S-O.In the O 1s spectrum (Fig. S1c), the absorption peaks at 531.4, 532.2 and 532.9 eV are from C = O, C-O-C and C-OH, respectively.The results inferred by XPS are consistent with the results of FT-IR characterisation [35].

Optical properties of S-CDs
We use UV-vis absorption spectroscopy and fluorescence spectroscopy to further understand the optical properties of m-cresol purple and S-CDs.Compared with the unheated m-cresol purple solution (Fig. S2), the S-CDs have no obvious UV-vis absorption at 350-600 nm (Figure 3a).The characteristic absorption peaks of S-CDs at 271 and 317 nm belong to π-π* and n-π* [38] (Figure 3a), respectively, which are different from the absorption peak positions of m-cresol purple solution at 200-350 nm (Fig. S2).The S-CDs hold the excitation-independent optical behaviour (Figure 3b).The best excitation wavelength is 400 nm, and the best emission wavelength is 553 nm (Figure 3c).In order to further study the fluorescence emission of S-CDs, the chromaticity diagram was studied.The CIE diagram (CIE coordinates: 0.38, 0.58) further confirmed that S-CDs emit yellow fluorescence, which further proved the luminescence properties of S-CDs (Figure 3d).However, under the excitation of 400 nm wavelength, the m-cresol purple solution showed extremely weak fluorescence intensity at 523 nm (Fig. S2).The results proved that the highly fluorescent S-CDs were successfully synthesised.The quantum yield of S-CDs is about 27.9%.

Optimisation of S-CDs synthesis conditions
To obtain the S-CDs with the better fluorescence intensity, we had optimised the synthesis conditions, which involved the reaction time and the temperature.The fluorescence intensity was tested below the emission wavelength of 553 nm (the testing excitation wavelength is 400 nm).As shown in Fig. S3, low temperature leads to weak fluorescence intensity due to incomplete carbonisation.With an increasine in temperature, the fluorescence intensity gradually increases [39].In addition, the bright fluorescent S-CDs can be obtained with longer reaction time.It is possible that long reaction time promotes the carbonisation of reactants, leading to an increase in the fluorescence intensity.
Considering the optimal fluorescence property, for the sake of energy saving and cost reduction, we finally chose 200°C for 4 h as the synthesis conditions to obtain the bright fluorescent S-CDs [40,41,42].

Fluorescence stability of S-CDs
In order to explore the best conditions for S-CDs as fluorescent probes in the determination of PA, we optimised some analysis parameters (pH, ionic strength, reaction time and storage time).In the pH-dependent experiment, we measured the fluorescence intensity of S-CDs in the pH range of 2-12.The results revealed that under neutral conditions, the fluorescence intensity of S-CDs reaches the maximum (Figure 4a).The reason for this phenomenon may be that excessive hydrogen ions and hydroxyl groups will affect the functional groups on the surface of S-CDs, and the fluorescence intensity of S-CDs is inhibited [35].The influence of NaCl concentration on the fluorescence intensity of S-CDs is almost negligible (Figure 4b), thus ensuring the application of S-CDs in bioimaging and environmental analysis.Finally, we explored the changes in the fluorescence intensity of the S-CDs after adding PA, and found that the fluorescence intensity of the S-CDs decreased rapidly within 1 min, and then kept steady (Fig. S4a).The best response time of S-CDs and PA is 1 min.We investigated the effect of storage time on the fluorescence intensity of the S-CDs.Five parallel groups of S-CDs solutions were placed in a refrigerator at 4°C, and the fluorescence intensity was measured every week.As displayed in Fig. S4b, the fluorescence intensity of the S-CDs is almost unchanged, indicating that the prepared S-CDs have good stability.

Selectivity of S-CDs for determination of PA
In order to evaluate the selectivity of S-CDs to PA, 20 μL of interfering substances (the concentration of all substances is 0.01 M), 10 μL of S-CDs solution were added to 2 mL ultrapure water, and then fully after stirring and equilibrating for 1 min, the fluorescence spectrum (λ ex = 400 nm, λ em = 553 nm) was measured.The results demonstrated that the S-CDs could specifically discriminate PA from other nitroaromatic explosives, different small molecules (Figure 5a), metal ions, and anion (Fig. S5).All in all, the S-CDs have high selectivity to PA.
In the S-CDs solution with the presence of PA (20 μL PA, 10 μL S-CDs, 2 mL water), the 20 μL of interfering substances was added into the above solution, and then we measured the fluorescence spectrum (λ ex = 400 nm, λ em = 553 nm).The fluorescence intensity quenched by PA is only slightly changed or neligible, which indicate that the interfering substances do not affect the determination of PA with S-CDs (Fig. S6).On the other hand, the interfering substances not interrupt the interaction between PA and S-CDs.

Fluorescence determination of PA
To determine the optimal conditions for PA testing, the temperature and pH were optimised.The variation of fluorescence quenching efficiency (F 0 /F) was applied to monitor the influence of temperature and pH on the quenching effect of PA.The BR buffer solution was used to prepare solutions of different pH to explore the role of pH.From Fig. S7a, no obvious change of F 0 /F was observed within pH 6-10.Hence, water (neutral solution) was chosen as the reaction system so that the pH will not interfere with the reaction result.Fig. S7b shows the effect of temperature, and the fluorescence intensity decreases with the increase in temperature.Although the S-CDs have stronger fluorescence intensity at low temperature, the room temperature is easier to control and remains stable during the experiment.Therefore, the room temperature was selected for PA detection.
The fluorescence intensity was significantly reduced by PA (Fig. S8).As the concentration of PA increased, the fluorescence intensity of S-CDs at 553 nm decreased steadily, as displayed in Figure 5b.In the range of 0-19.6 μM, the logarithm of the S-CDs fluorescence intensity ratio ( ln F 0 =F) has a good linear relationship with the concentration of PA, and the corresponding linear regression equation is ln F 0 =F = 0.0205 C PA + 0.00823 with the correlation coefficient R 2 = 0.997 (Figure 5c).In Figure 5d, when the concentration of PA ranges from 19.6 to 259.0 μM, the linear regression equation is ln F 0 =F = 0.01384C PA þ 0.22227, R 2 ¼ 0.997 (F 0 and F represent the fluorescence intensity of S-CDs with/without PA, respectively; λ ex = 400 nm, λ em = 553 nm).The detection limit is 14.6 nM (based on the equation: LOD ¼ 3δ=s, where 3δ represents the standard deviation of the 11 blank measurements, s represents the linear simulation resultant curve slope).Table 1 compares the emission wavelength, the detection limit, the linear range and the quantum yield of the different fluorescent probes for determining PA.Obviously, the method we propose has relatively low detection limits and a wide linear range in comparison with other probes.

Analysis of quenching mechanism
We have carried out a series of experiments to explore the possible fluorescence quenching mechanism.We tested the relative fluorescence lifetime of S-CDs.The average lifespan of pure S-CDs is 3.475 ns.After PA is added (C PA = 0.01 μM, V PA = 20 μL), the average lifespan of S-CDs is 3.417 ns (Figure 6a).There is no obvious change in the average lifespan before and after PA is added, so the static quenching effect is not a luminescence  quenching mechanism.The absorption band of the PA and the excitation band of the S-CDs have the spectral overlap (purple area) (Figure 6b).Thus, the emitted energy of S-CDs was absorbed by the PA during the quenching process.From Fig. S8, it is evident that the emission wavelength remains at 553 nm after the addition of PA.These results imply that the quenching mechanism could be IFE [44][45][46].IFE is considered to be a simple and flexible mode of interaction in fluorescence analysis because it does not require a chemical connection between the analyte and the CDs [47].

The real sample detection
In order to evaluate the feasibility and repeatability of this method, a standard addition method was carried out to test the content of PA in lake water (no PA was detected in the lake water, and the standard concentrations of PA added to lake water were 1 μM, 3 μM and 5 μM).The measurement results are shown in Table 2.The measured value of PA content in the actual sample is highly consistent with the added amount.The corresponding recovery rate is 93.0-103.7%,and the relative standard deviation is below 2.76% (n = 5).Therefore, the S-CDs we designed can be accurately applied to the determination of PA in actual water samples.

Cytotoxicity studies of S-CDs and Biological imaging
We tested the viability of C. elegans and A. thaliana incubated in the different concentrations of S-CDs solutions.As depicted in Fig. S9, the survival rate was still above 82% at the highest concentration of S-CDs (300 mg mL −1 ).From the results, we infer that the S-CDs solution has low toxicity.In addition, at a concentration of 100 mg mL −1 , the viability of C. elegans was higher than 92%, so we chose this concentration for subsequent imaging.Accordingly, 50 mg mL −1 was selected for A. thaliana imaging.C. elegans and A. thaliana often exist in contaminated soil.They can not only be used for biological imaging but also as a typical representative of animals and plants.Therefore, it is potential for us to detect the content of PA in its body to evaluate the degree of pollution.Using a confocal laser scanning microscope, we performed the live biological imaging of C. elegans and A. thaliana.When excited with a 405 nm laser, the C. elegans coincubated with the S-CDs solution have a good morphology and emit a bright yellow fluorescence (Figure 7a).This phenomenon indicates that S-CDs have low toxicity, small size and good biocompatibility; moreover, S-CDs can enter the body through endocytosis.
After incubating with S-CDs and the different concentrations of PA solutions, it was found that the yellow fluorescence of C. elegans gradually weakened as the concentration of PA solution increased (Figure 7d).Our study indicates that the S-CDs can not only perform the live biological imaging of C. elegans but they are also expected to be able to measure the PA content in C. elegans through changes in fluorescence intensity.We also implemented confocal fluorescence microscopy to obtain the imaging of A. thaliana.The roots of A. thaliana co-incubated with S-CDs solution showed good morphological structure and emitted a bright yellow fluorescence (Figure 8a).After the A. thaliana was incubated with the S-CDs and PA solution, we found that the intensity of yellow fluorescence was highly reduced (Figure 8d).From this, we can explain that S-CDs can potentially be used for the detection of PA in A. thaliana.

Conclusion
In this paper, S-CDs, with a high quantum yield (~27.9%), are obtained by the facile hydrothermal treatment of m-cresol purple.The characterisation of S-CDs proved the formation of sulphur-doped carbon quantum dots and revealed the functional groups on the surface.S-CDs showed the high intensity of photoluminescence and the emission wavelength is independent of excitation wavelength.Therefore, they are used for the determination of PA with high selectivity, high sensitivity and wide linear range.Meanwhile, the developed fluorescent probe S-CDs have been successfully applied to determine the PA content in the lake water with satisfactory results (recoveries: 93.0-103.7%,RSD < 3%).Furthermore, S-CDs also exhibited great potentials for application in bioimaging and biological sensing.

Figure 3 .
Figure 3. (a) The UV-vis absorption spectrum and fluorescence spectra of the S-CDs, Inset: photograph of the S-CDs under daylight (left) and light (right).(b) Excitation wavelength dependence, (c) Excitation-emission matrix and (d) CIE coordinates of the S-CDs.

Figure 4 .
Figure 4.The effect of (a) pH and (b) ionic strength on the fluorescence intensity of S-CDs.

Figure 5 .
Figure 5. (a) The (F 0 -F)/F of the S-CDs in the presence of different small molecules and other nitroaromatic explosives (F 0 is the fluorescence intensity of the S-CDs solution, F is the fluorescence intensity of the S-CDs solution after added interfering substances) (b) Fluorescence emission spectra of the S-CDs against different PA (0-259.0μM) (c) The relationship between ln(F 0 /F) and PA concentration in the range of 0-19.6 μM (d) The relationship between ln(F 0 /F) and PA concentration in the range of 19.6-259.0μM (F 0 and F represent fluorescence intensities of the S-CDs at 553 nm in the absence and presence of PA).

Figure 6 .
Figure 6.(a) Fluorescence lifetime curve of S-CDs in the absence/presence of PA; (b) Overlap (purple area) between absorption spectrum of PA and excitation spectrum of S-CDs.

Table 1 .
Comparison of different fluorescence methods for the determination of PA.

Table 2 .
Determination of PA in actual water samples (n = 5).