Highly selective fluorescent carbon dots probe for mercury(II) based on thymine–mercury(II)–thymine structure

A novel thymine-functional fluorescent sensor was developed for Hg2+ detection with high sensitivity and selectivity. The synthesis of the fluorescent sensor took two steps (1) the synthesis of amine-functionalized carbon dots (CDs-PEI); and (2) the attaching of thymine moieties on to the surface of the CDs-PEI through EDC/NHS coupling chemistry to obtain the thymine-functional fluorescence carbon dots (CDs-Thy). The CDs-Thy were successfully applied to detect Hg2+ by quenching fluorescence with a rapid response through photoinduced electron transfer with the formation of T–Hg2+–T structures. The linear concentration range of Hg2+ is 0–1.0 μmol L−1 and a limit of detection (LOD) as low as 3.5 × 10−8 mol L−1 was obtained. Moreover, the CDs-Thy can resist interference from other metal ions and anions. The CDs-Thy were also used for the Hg2+ detection in water samples and the recoveries were from 90% to 104%. Due to the simplicity and effectiveness, it shows great promise as a potential sensing platform for Hg2+.


Introduction
Mercury(II) is a toxic pollutant that is widely distributed in soil, water, and air. 1 Metal smelting, coal production, waste disposal, rening and manufacturing, and the production of chlor-alkali processes will discharge large quantities of mercury into the environment. [2][3][4] Inorganic mercury can be transformed into the highly toxic form such as organic methylmercury under the catalysis of microorganism. These organic mercury that can amass in human body throughout the food web 2,5,6 and cause cinesipathy, disgnosia, cerebral lesion, impaired vision, and even death. [7][8][9] For these reasons, the development of straightforward, highly sensitive and selective methods for monitoring and detecting Hg 2+ is of great importance.
In recent years, uorescent probes for Hg 2+ detection have attracted extensive attention due to their convenient and timeefficient procedures, [10][11][12][13][14][15][16] which make them good alternatives to traditional analytical techniques, such as electrochemical method, 17 atomic absorption spectroscopy (AAS), 18 and inductively coupled plasma mass spectrometry (ICP-MS), 19,20 etc. Another alternative could be the photoluminescent carbon dots (CDs) because of their obvious advantages such as easy in preparation, 21 environmental friendliness, 22 good biocompatibility, 23 water solubility 24 and low toxicity 25 by comparing with metal quantum dots and traditional organic dyes. 26 They have also been applied to the analysis and detection of ions and biomolecules. 21,[27][28][29][30][31][32] However, in their recent status, the photoluminescent CDs showed low sensitivity and poor selectivity. [33][34][35][36] Therefore, it remains a challenge to develop uorescent CDs for highly selective and sensitive recognition of Hg 2+ in biological and aqueous systems.
In this work, we developed a new kind of uorescent CDs for Hg 2+ detection. We redesigned the functional surfaces of existing photoluminescent CDs 37 that had found application in Cu 2+ detection 38 with commendable selectivity. We found that there are amino groups on the surface of the CDs, which can be modied by thymine (T). As the thymine can form T-Hg 2+ -T structure (Fig. 1A), 39,40 we conjectured that the newly developed thymine-functional CDs (CDs-Thy) will have high selectivity for Hg 2+ . The results showed that the CDs-Thy can be applied to determine the Hg 2+ in water samples. Satisfactory results were obtained.

Instruments
The UV-visible absorption and the FT-IR spectra were scanned using UV 3150 UV-vis spectrometer (Shimadzu) and IR-Nicolet Avatar 330 spectrometer (Bruker), respectively. The uorescence measurements and quantum yields were carried out by using a RF5301 spectrometer (Shimadzu). A PHS-25 pH meter (INESA Scientic Instrument Co., Ltd) was employed to determine pH values. The 1 H NMR spectra were obtained by using Avance III-400 MHz spectrometer (Bruker). The transmission electron microscope (TEM) images were measured by using JEM-2010HR (Jeol Ltd). X-ray photoelectron spectroscopy (XPS) and uorescent lifetime were recorded using an ESCALab250Xi spectroscopy (Thermo Fisher) and FLSP920 photoluminescence spectrometer (Edinburgh Instruments Ltd), respectively.

Synthesis of amine-functionalized CDs (CDs-PEI)
The CDs-PEI were prepared according to published papers elsewhere. 37,38

Synthesis of thymine-functionalized CDs (CDs-Thy)
The CDs-Thy were synthesized by using excess thymine-1-acetic acid to react with the CDs-PEI. Before the modication of thymine on the CDs-PEI, the thymine-1-acetic acid (128 mg, 0.698 mmol) was activated with EDC (161 mg, 0.840 mmol) and NHS (385.4 mg, 3.35 mmol) in 140 mL MES (0.1 M, pH 5.5) solution for 20 min at 25 C with constant stirring. Then the CDs-PEI (44.0 mg) were added into the solution and kept reacting for another 48 h (Fig. 1B). Aer reaction, the concentrated solution was column chromatographed on silica-gel (0.01 M HCl) to remove the by-products.

Fluorescence measurement
Stock solution of the CDs-Thy (200 mg mL À1 ) was prepared using deionized water. Stock solutions of cations (0.001 M) were prepared under suitable pH conditions. The pH optimization experiments of CDs-Thy in the presence and absence of Hg 2+ between 5.8 and 8.2 were obtained through the uorescence emission intensities, respectively. Real time uorescent signals of the CDs-Thy were obtained with different concentrations of Hg 2+ to optimize the optimal reaction time. All uorescence spectra were recorded at optimized detection conditions in the PBS buffer solution (0.1 M, pH 7.4) at 25 C.

Measurement of water samples
Tap water and Pearl River water were used for Hg 2+ detection. All these samples were centrifuged and ltered with micropore membranes. The obtained ltrate was added with standard Hg 2+ samples with different concentrations and used for further analysis.

Characterization of CD-PEI and CDs-Thy
To investigate the surface modication of the CDs-Thy, FT-IR spectroscopy was used to characterize the CDs-PEI, thymine-1acetic acid and the CDs-Thy, respectively. As shown in the spectrum a of Fig. 2A, the peaks at 3414 cm À1 (y N-H ) and 1630 cm À1 (d N-H ) conrmed that the surface of CDs-PEI contain amine groups. The presence of these amine groups provides CDs with the potential to be modied with the thymine-1-acetic acid consisting of -COOH group. Compared to the CDs-PEI, the CDs-Thy have the characteristic peaks in 2974 cm À1 (y -CH 3 and y -CH 2 ) and 1479 cm À1 (d -CH 2 ) as the spectrum b shown in the Fig. 2A, ascribing to vibration of -CH 3 and -CH 2 , which can also be found at 2961 cm À1 and 1483 cm À1 in the IR spectrum of thymine-1acetic acid as the spectrum c shown in Fig. 2A. In addition, the CDs-Thy show more peaks around 1641 cm À1 (y C]O ) and 1268 cm À1 (y C-N and d N-H ) that are all attributed to the thymine-1acetate groups on the surface of CDs-Thy, which also can be found in the IR spectrum of thymine-1-acetic acid at 1633 cm À1 and 1258 cm À1 as the spectrum c of Fig. 2A. These results revealed that the thymine-1-acetic group was modied on to the surface of the CDs. Moreover, with the formation of carbonyl amide, the absence of the amide-II band at 1568 cm À1 (d -CONH-) can further prove the successful conjugation of the CDs-PEI and the thymine-1-acetic acid. Fig. 2B shows the 1 H NMR spectra of the CDs-Thy, the CDs-PEI and the esteried product of thymine-1-acetic acid, respectively. In the 1 H NMR spectrum of the CDs-Thy, the multiple peaks at 3.1 ppm corresponded to the resonance signals of alkyl group connected with N atoms and C]O. One can also nd that the chemistry shis at 1.8 ppm, 4.4 ppm and  Fig. 2B). The reason to measure the 1 H NMR spectrum of methyl thymine-1-acetate instead of that of thymine-1-acetic acid is that the later is not soluble in water. These results conrmed that the CDs-Thy were successfully prepared from the CDs-PEI and thymine derivative.  (Fig. 3d). The N 1s spectrum (Fig. 3e) of the CDs-PEI can be resolved two components which are attributed to C-NH 2 at 400.2 eV and C]N at 401.8 eV, respectively. However, the tting analysis of the N 1s XPS spectrum of the CDs-Thy (Fig. 3f) reveals three distinct peaks at 399.6 eV, 400.0 eV, and 401.9 eV corresponding to O]C-NH, C-N-C and C]N, respectively. It means that the primary amine of CDs-PEI almost totally translate to secondary amine aer the reaction with thymine-1-acetic acid. Futhermore, the O 1s XPS spectra of the CDs-PEI (Fig. 3g) and the CDs-PEI (Fig. 3h) reveal two types of  oxygen bonds, which are associated with C]O at 531.7 eV and C-O at 533.0 eV, respectively. 44,45 And the relative contribution of C]O in the CDs-Thy has an signicant increase compared to the CDs-PEI, from 68.4% to 88.8%. In conclusion, the results conrm that the surfaces of the CDs-PEI were successfully functionalized with thymine groups. Fig. 4A and B show the TEM results of the CDs-PEI and the CDs-Thy, respectively. It can be found that the CDs-PEI mainly consist of uniform nanoparticles in size ranging from 2 to 5 nm. The TEM showed that the average particle size of the CDs-PEI was about 3.0 nm (inset of Fig. 4A). In contrast, the particle size range and the average size of the CDs-Thy were 2-7 nm and 3.5 nm aer the CDs-PEI modied with thymine group, respectively (inset of Fig. 4B). These results demonstrated that the conjugated thymine contributed about 0.5 nm to the average particle diameter. Fig. 4C shows the UV/vis spectra of the CDs-Thy, the CDs-PEI and thymine-1-acetic acid, respectively. It can be seen that a relatively weak and broad absorption band appeared in a region from 320 to 400 nm, which was ascribed to the n-p* transition of the C]O and C]C conjugated structure of both CDs-PEI and CDs-Thy. The strong absorption peak around 272 nm on the spectrum of CDs-Thy proved the successful modication of thymine group on the CDs because the similar peak was also observed in the spectrum of thymine-1-acetic acid. Furthermore, the content of thymine group was measured to be 6.35 mg per 100 mg by using the standard addition experiments (Fig. S1 †). Fig. 4D shows the uorescence excitation and emission spectra of the CDs-PEI and the CDs-Thy, respectively. When excited at 354 nm, both CDs-PEI and CDs-Thy exhibit maximum uorescence emission around 440 nm. However, the uorescence emission spectra of CDs-PEI and CDs-Thy have no signicant difference because of their same photoluminescence mechanism. Fig. 4E shows the emission spectra of the CDs-Thy at different excitation wavelengths which varied from 300 nm to 400 nm. The emission of the CDs-Thy showed the excitationindependent property and may originated from a single species. 46 Fig. 4F shows the uorescence decay spectra of CDs-PEI and CDs-Thy. Both of the CDs-PEI and the CDs-Thy consists of a short s 1 and a long s 2 lifetime components (Table S1 †). The uorescence decay curves were tted with a double-exponential function. 47 The lifetime of CDs-Thy was 10.17 ns which was closed to the CDs-PEI (10.97 ns). Furthermore, the uorescence quantum yields of CDs-PEI and CDs-   Thy, referenced to a quinine sulfate solution in 0.1 M H 2 SO 4 (F F ¼ 0.55), were determined to be 23.6% and 31.0%, respectively (Table 1).

Hg 2+ sensing properties of the CDs-Thy
Experimental conditions including pH and reaction time for UV/vis and uorescence experiments of the CDs-Thy were optimized (see Fig. S2 †). Fig. 5A shows the UV-vis spectra of the CDs-Thy in the presence of different concentrations of Hg 2+ . One can see that absorption peaks appeared around 307 nm, which could be attributed to the evolution of T-Hg 2+ -T structures. The intensities of the absorption band increased gradually with the increasing of Hg 2+ concentrations. Fig. 5B shows a good linear relationship between the absorbance at 307 nm and the concentrations of Hg 2+ with a correlation coefficient of R 2 ¼ 0.999. Fig. 5C shows the uorescence emission response of the CDs-Thy reacting with increased concentrations of Hg 2+ from 0.1 to 50 mM. One can nd that the intensity at 440 nm was continuously decreased with the addition of Hg 2+ . Fig. 5D shows a very steep decrease of uorescence intensities at 441 nm when adding Hg 2+ into the solution of the CDs-Thy. Such an outstanding uorescence quenching was assigned to the photoinduced electron transfer from the excitation state of CDs-Thy to the vacant d orbital of Hg 2+ when the T-Hg 2+ -T structure was formed. 34,36 Furthermore, the uorescence intensity of the CDs-Thy toward Hg 2+ decreased linearly over the Hg 2+ concentration range of 0-1.0 mM (Fig. S3 †). The limit of detection (LOD) of the CDs-Thy for Hg 2+ was thus determined to be about 3.5 Â 10 À8 mol L À1 according to the formula: LOD ¼ 3s/k, where the s means the standard deviation of the response, k means the slope of the calibration curve. Table 2 shows the comparison of linear range and limit of detection with other recently reported Hg 2+ probes. The CDs-Thy provide a remarkable lower detection limits and a satised linear range. These results indicated that the CDs-Thy are highly sensitive for Hg 2+ . Fig. 6 shows the TEMs of the solutions of the CDs-Thy with and without Hg 2+ , respectively. One can nd that large reticular aggregates were formed when Hg 2+ was introduced to the solution of CDs-Thy (Fig. 6B). The quenching of the uorescence emission by Hg 2+ could be the photoinduced electron transfer (PET) process when the T-Hg 2+ -T structures were formed.

Selectivity and competition of CDs-Thy for Hg 2+ detection
Fig . 7A shows the selectivity experiments of the CDs-Thy with 16 different metal cations. The uorescence responses of the CDs-Thy were recorded in the presence of different metal cations, including Cu 2+ , Hg 2+ , Zn 2+ , Mg 2+ , Fe 2+ , Au 3+ , Co 2+ , Ca 2+ , Pb 2+ , Ba 2+ , Cr 3+ , Na + , Ag + , Ni 2+ , K + , and Fe 3+ . As expected, an outstanding uorescence intensity descent was found only with   the addition of Hg 2+ , while no noticeable changes in the uorescence spectra were observed upon addition of other metal cations. These results demonstrated that the CDs-Thy possess a high selectivity for Hg 2+ detection. Competition measurements were also measured by adding the mixed solution of Hg 2+ and other metal cations to the solution of CDs-Thy as shown in the Fig. 7B. Severe uorescence quenching were observed when the CDs-Thy were exposed to Hg 2+ , which shows a superior reaction ability between the CDs-Thy and Hg 2+

The analysis of water samples
The CDs-Thy were applied in several tap water and river water samples to investigate the practicability of the CDs-Thy for Hg 2+ sensing. Since no Hg 2+ was detected in these samples, certain amount of standard Hg 2+ was added to the raw water samples. The results (Table 3) shows that the recoveries ranged from 90% to 104%, and the relative standard deviations are less than 7%. These results veried that the analysis of Hg 2+ using the CDs-Thy in water samples was satisfactory.

Conclusions
We developed a novel thymine-functionalized carbon dots, the CDs-Thy, which can be a useful tool for selective and sensitive determination of Hg 2+ in a totally aqueous solution. The CDs-Thy showed a specic on-off uorescence response to Hg 2+ based on T-Hg 2+ -T metal ion-mediated base pairs with a low detection limit. In addition, the CDs-Thy also had a rapid and selective response to Hg 2+ at natural pH range. Moreover, the CDs-Thy also showed its ability for the measurement of Hg 2+ in both tap and river water with satisfactory results. In view of these desirable features, the CDs-Thy could be a potential tool for detection and monitoring of Hg 2+ .

Conflicts of interest
There are no conicts to declare.