Simultaneous determination of Hg(II) and Cu(II) in water samples using fluorescence quenching sensor of N-doped and N,K co-doped graphene quantum dots

2.1 Chemicals and materials

All the chemicals used were of analytical grade. Citric acid and sodium hydroxide used were from Carlo Erba (Italy). All of selected mineral acids, potassium nitrate, sodium nitrate and sodium acetate were obtained from QRec (New Zealand). Mercury nitrate and copper nitrate trihydrate were purchased from Sigma-Aldrich (Germany). Acetic acid was from Merck (Germany). Potassium dihydrogen phosphate and dipotassium hydrogen phosphate were from LOBA (India). Paraffin oil was brought from Ajex Finechem (Australia). Deionized water (Simplicity Water Purification System, Model Simplicity 185, Millipore, U.S.A.) was used throughout the experiments.

2.2 Instruments and apparatus

Spectrofluorophotometer (Shimadzu RF-5301PC, Japan) with excitation and emission slit widths of 3 nm was mainly used. pH meter (Model Proline B21, Becthai Equipment & Chemical, Thailand), analytical balance (Model BSA224S-CW, Scientific Promotion, Thailand), quartz cell with 1-cm path length (Fisher Scientific, U.S.A.) were also used. Round bottom flask (Pyrex®, England) and hot plate with a magnetic stirrer in association with paraffin oil bath were set for the citric acid pyrolysis.

2.3 Synthesis and characterization of GQDs and N-GQDs & N, K-GQDs

Graphene quantum dots were prepared by citric acid pyrolysis (Dong et al., 2012). Briefly, 2.0 g citric acid (QRec™, New Zealand) was transferred into a 100 mL round bottom flask which was heated to 250 °C in an oil bath for about 5 min. Citric acid was liquated and its color was changed rapidly to yellow. This liquid was then dissolved by dropwise addition of 100 mL 0.25 M NaOH (QRec™, New Zealand) with continuous stirring for 30 min. The obtained GQDs solution was stored at 4 °C before use. For N-GQDs or N,K doped GQDs synthesis, the GQDs, as also prepared via citric acid pyrolysis, were simultaneously treated with 1–5% (w/v) nitric acid (QRec™, New Zealand) or 1–5% (w/v) potassium nitrate (QRec™, New Zealand) to find the suitable amount of N or N,K doping condition; starting with 2.0 g of citric acid and 1% (w/v) nitric acid or 1% (w/v) potassium nitrate was transferred into a 100 mL round bottom flask, which was heated to 250 °C in an oil bath for about 5 min and then each of the two mixture solutions was dissolved in 100 mL 0.25 M NaOH solution as mentioned above. The as-prepared N-GQDs and N,K-GQDs solution were also stored at 4 °C before use.

2.4 Quantum yield measurement

Determination of the fluorescence quantum yield (QY) of undoped GQDs, N-GQDs and N,K-GQDs was carried out by comparing with standard solution of quinine sulfate (QY = 0.54 at 360 nm) used as a reference compound. The quantum yield is calculated using the slope of the regression line generated by plotting the integrated fluorescence intensity at the emission wavelength of 460 nm against with its absorbance at 370 nm for multiple concentrations of each of undoped GQDs, N-GQDs and N,K-GQDs, and quinine sulfate solutions. The QY was calculated by the following equation (Amini et al., 2017): $$\text{Q}={\text{Q}}_{\text{R}}\times [\text{m}/{\text{m}}_{\text{R}}]\times [{\text{n}}^2/{{\text{n}}^2}_{\text{R}}]$$ where, Q is the quantum yield, m is the slope of the mentioned regression line, n is refractive index that assuming equal to 1.33 for both solution and the subscript R refers to quinine sulfate solution.

2.5 Selective fluorescence quenching detection of Hg²⁺ and Cu²⁺ using N-GQDs & N,K-GQDs

For these following experiments, about 100 mg/L of each GQDs, N-GQDs or N,K-GQDs solution and 0.1 M phosphate buffer solution pH 7 were mixed well in a 10 mL volumetric flask. Then various concentrations of each of Hg²⁺ and Cu²⁺ were added into an aliquot of each of undoped GDQs, N-GQDs or N,K-GQDs solution (10 mL final volume). These Hg²⁺ and Cu²⁺ quenching fluorescence spectra of each of the undoped GQDs, N-GQDs or N,K-GQDs solution were recorded immediately at λ_ex/λe_m 370/460 nm. Then, the spectral measurements were used to plot the quenching external calibration curve for both Hg²⁺ and Cu²⁺ with different concentration ranges.

2.6 Recovery study of real water samples

To evaluate the undoped GQDs, N-GQDs and N,K-GQDs based fluorescence quenching sensor for Hg²⁺ and Cu²⁺ detection simultaneously in an artificial solution, the applicability of the proposed method for real water samples including drinking water and tap water was carried out. Tap water samples were collected with polyethylene bottle from Khon Kaen area and were analyzed without any pretreatment. All water samples were filtered through Whatman filter paper No.42 and then adjusted to pH 7 with 0.1 M phosphate buffer. In this procedure, 1 mL of the water sample and 0.5 mL of undoped GQDs, N-GQDs or N,K-GQDs (100 mg/L) solution were transferred into 10.0 mL volumetric flask. Then, each of the sample mixture was spiked with three concentration levels of standard solution of Hg²⁺ (20.0, 50.0 and 100.0 µM) and standard solution of Cu²⁺ (100.0, 300.0 and 500.0 µM).

2.7 Statistical analysis

Regarding statistical data analysis in an analytical method validation, statistics for those experimental results obtained at least in triplicate include mean, standard deviation, confidence intervals, and linear regression. Data analysis and their curve fitting using all registered statistical packages as such common Excel 2016 is reported.

3.1 FT-IR characterization of the undoped GQDs, N-GQDs & N,K-GQDs

The FTIR spectra of undoped GQDs, N-GQDs and N,K-GQDs are shown in Fig. 1(a)–(c). From the vibration spectrum of undoped GQDs (Fig. 1(a)), the absorption peaks at 3436, 2950, 1576 and 1080 cm⁻¹ were observed, indicating the vibration modes of hydroxyl stretching, C—H stretching, C⚌C stretching of polycyclic aromatic hydrocarbons (Wang et al., 2014), C—O in COH/COC (epoxy) groups (Wang et al., 2009), respectively. These results confirm the undoped GQDs successfully synthesized by citric acid pyrolysis. The FTIR spectrum of N-GQDs is shown in Fig. 1(b), with the broad band at 3000 – 3500 cm⁻¹ corresponding to the stretching vibrations of O-H and N-H (Kaur et al., 2017; Mondal et al., 2018). The absorption peaks at ∼1700, 1585 and 1218 cm⁻¹ indicate C⚌C, C⚌O and C—O—C stretching vibrations, respectively. Surprisingly, the IR spectrum peak at ∼1400 cm⁻¹ was from the vibration mode of C-NH, indicating that the successful doped of nitrogen atom onto the surface of GQDs (Ju and Chen, 2014; C. Zhang et al., 2016). Furthermore, Fig. 1(c) shows an additional peak at 1777 cm⁻¹ for the N,K-GQDs, which is probably resulted from potassium related bonding (Qian et al., 2016; Ain et al., 2016). Therefore, these FTIR spectra can be commonly used to confirm their chemical structures of the as-prepared N-GQDs and N,K-doped GQDs.

Close

Fig. 1

FT-IR Spectra of (a) GQDs, (b) N-GQDs and (c) N,K-GQDs.

Export to PPT

3.2 Effect of mineral acids on GQDs yield

For preliminary study, fluorescence spectra of both undoped GQDs and GQDs doped with some mineral acids are shown in Fig. 2. Their absorption spectra were found around 370 nm. When the mineral acids were thermally degraded and decorated into the defaulted surface structure of the GQDs by pyrolysis situation, their emission spectra appeared nearly around 460 nm with either an enhanced or quenched fluorescence process. Thus confirming that the mineral acids doping in the GQDs was successfully obtained and such GQDs doped with nitric acid (HNO₃) was chosen as the starting material because of much higher fluorescence intensity. The enhancing fluorescence emission of GQDs doped with HNO₃ possibly resulted from the N-doping induced modification of the chemical and electronic characteristics of the GQDs (Liu et al., 2018; Cai et al., 2014; Wang and Fan, 2018).

Close

Fig. 2

Fluorescence spectra of GQDs (black) and 1%(w/v) each of some mineral acids doped with GQDs (nitric acid/GQDs – green; phosphoric acid/GQDs – blue; hydrochloric acid/GQDs – red; perchloric acid/GQDs – violet).

Export to PPT

3.3 Quenching effect of metal ions on the fluorescence intensity of N-GQDs

The effect of some metal ions including Ag⁺, Na⁺, Hg⁺, Cu²⁺, Pb²⁺, Mn²⁺, Mg²⁺, Fe^2+, Fe³⁺ and Al³⁺ ions on the fluorescence intensity of the N-GQDs was tested. The fluorescence intensity and the normalized fluorescence intensity of N-GQDs solution showed the quenching effect with adding Hg²⁺ (Fig. 3(a)) and Cu²⁺ (Fig. 3(b)). Hg²⁺ can interact with negative charge on the surface of N-GQDs because of their electrostatic interactions on the quenching process. Some metal ions such as Cu²⁺ can also interact with negative charge on the surface of N-GQDs but their ion bindings are somewhat weak and its binding affinity is not much as strong as that of the Hg²⁺ (Li et al., 2015). Therefore, the quenching process of Cu²⁺ would be weaker than that of Hg²⁺. The quenching processes of both Hg²⁺ and Cu²⁺ on the surface of N-GQDs are considerably due to facilitating non-radiative electron/hole recombination annihilation through an effective electron transfer (Huang et al., 2013; Liu et al., 2016; He et al., 2016). The electron transfers from the excited state of N-GQDs to an empty d-orbital of the metal ion lead to the fluorescence quenching effect of the N-GQDs.

Close

Fig. 3a

Fluorescence spectra of 1%(w/v) HNO₃ doped GQDs in the presence of various metal ions

Export to PPT

Close

Fig. 3b

Normalized fluorescence intensities of 1%(w/v) HNO₃ doped GQDs in the presence of various metal ions

Export to PPT

3.4 Selective detection of Hg²⁺ and Cu²⁺ using N-GQDs

The fluorescence intensities of N-GQDs increased with the increase in HNO₃ concentration (Fig. 4). The enhancement of florescence intensity would be due to the emissive traps of the nitrogen doped on the surface of GQDs. The nitrogen atom is comparable in an atomic size to the carbon atom and their effects turn on the electronic nature of GQDs due to difference in an electronegativity between C and N that appears large enough to result in significant charge transfer in the C-N composites (Chen et al., 2017). Therefore, the chemical doping of nitrogen atom into the framework of GQDs would directly be attributed. Figs. 5 and 6 show such calibration curves plotting between F_o − F and concentrations of Hg²⁺ and Cu²⁺ (µM) using GQDs doped with 1 to 5% (w/v) of HNO₃. It was shown that their slopes between 20–200 µM of Hg²⁺ and slopes between 100–500 µM of Cu²⁺ were found linearly increasing with the percentage of nitric acid used, of which their calibration curves of GQDs doped with 3% and 5% (w/v) nitric acid gave higher sensitivity to detect both Hg²⁺ and Cu²⁺, respectively.

Close

Fig. 4

Fluorescence spectra of GQDs and 1–5% (w/v) HNO₃ doped GQDs

Export to PPT

Close

Fig. 5

Calibration curves plotting between F_o − F and concentrations of Hg²⁺ (µM) using GQDs doped with 1–5% (w/v) HNO₃

Export to PPT

Close

Fig. 6

Calibration curves plotting between F_o − F and concentrations of Cu²⁺ (µM) using GQDs doped with 1–5% (w/v) HNO₃

Export to PPT

Fig. 7 shows the fluorescence spectra of GQDs doped with potassium nitrate. The results showed the fluorescence spectrum is rather difference in shape from that of the N-GQDs, and their intensities also increased with the increase in KNO₃ concentrations. So, GQDs doped with this nitrate salt may has distinct chemical structure, the so-called N,K-GQDs form. The enhancement of their florescence intensity would due to attribution in between π and π* orbital of the N,K-GQDs (C. Zhang et al., 2016; Ain et al., 2016). The change in energy structure would therefore lead to the change in their optical properties as well.

Close

Fig. 7

Fluorescence spectra of GQDs and 1–5% (w/v) KNO₃ doped GQDs

Export to PPT

Also, the N,K-GQDs can be used to detect both Hg²⁺ and Cu²⁺ as shown in Figs. 8 and 9. The calibration curves show that their slopes between 20–200 µM for Hg²⁺ and 100–500 µM for Cu²⁺ increase with increasing the percentage of potassium nitrate, of which the calibration curve of GQDs doped with 5% (w/v) KNO₃ give higher sensitivity to detect both of Hg²⁺ and Cu²⁺.

Close

Fig. 8

Calibration curves plotting between F_o − F and concentrations of Hg²⁺ (µM) using GQDs doped with 1–5% (w/v) KNO₃

Export to PPT

Close

Fig. 9

Calibration curves plotting between F_o − F and concentrations of Cu²⁺ (µM) using GQDs doped with 1–5% (w/v) KNO₃

Export to PPT

The strong fluorescence intensity of N-GQDs and N,K-GQDs was also investigated by using quinine sulfate solution as a reference case. The quantum yields of different doping materials are compiled in Table 1. The fluorescence quantum yields of the obtained N-GQDs and N,K-GQDs was found to be 83.42% and 94.07%, respectively, which they were higher than that of the undoped GQDs and quinine sulfate, so the doping N and K on the surface of GQDs gave significant enhancement of the fluorescence quantum yields of the GQDs (Yan et al., 2016).

Comparison of detection sensitivity for Hg²⁺ and Cu²⁺ in terms of their slopes of its calibration curves using the undoped GQDs, N-GQDs and N,K-GQDs as fluorescence quenching sensor is shown in Table 2. Thus, quantitative analysis of both Hg²⁺ and Cu²⁺ was validated and their analytical parameters were listed in Table 3. The fluorescence intensity of the optimal N-GQDs decreased with an increasing in concentrations from 20 to 100 µM of Hg²⁺ and 100 to 500 µM of Cu²⁺. Both limit of detection (LOD) and limit of quantitation (LOQ) of the proposed florescence quenching sensor were 0.42 μM & 1.41 μM for Hg²⁺, and 13.2 μM & 44.0 μM for Cu²⁺, respectively. The intra-day precision and inter-day precision expressed as the percentage of RSDs were 6.98 and 11.35 for Hg²⁺, and 11.78 and 9.43 for Cu²⁺, respectively, derived from the detection of the solution model of 40.0 μM Hg²⁺and 300.0 μM Cu²⁺. These results suggest that the proposed method exhibits a very good repeatability and is therefore acceptable for simultaneous determination of Hg²⁺ and Cu²⁺ with different sensitivity ranges in the same sample solution.

The proposed protocol was applied for Hg²⁺ and Cu²⁺ determination in drinking water and tap water samples. From the results (Tables 4–6), their recoveries of the spiked Hg²⁺ and Cu²⁺ by using undoped GQDs were obtained in the ranges of 96.75 ± 2.88 − 104.95 ± 6.01% and 96.38 ± 6.18 − 106.30 ± 7.06%, respectively. The recoveries when using N-GQDs were found in the ranges of 99.00 ± 1.29 − 101.25 ± 1.66% for Hg²⁺ and 99.06 ± 2.06 − 101.73 ± 0.65% for Cu²⁺, while the recoveries obtained from using N,K-GQDs were also in the ranges of 95.75 ± 7.11 − 102.75 ± 2.39% for Hg²⁺ and 97.32 ± 7.17 − 101.10 ± 5.55% for Cu²⁺. This method evaluation is very satisfactorily attributed in terms of an analytical accuracy of the trace metal detection. However, comparison of different fluorescence sensing considered as LOD for Hg²⁺ and Cu²⁺ detection is complied. Even though its sensitivity of N-GQDs probe for Hg²⁺ detection is a little bit higher when compared with other literatures, it might be due to differences in optimum conditions used under specific detection studied. In such case of Cu²⁺ detection using N-GQDs sensing probe it is still limited. Therefore, this study is aimed to report the use of N-GQDs as fluorescence sensing probe for simultaneous detection of both Hg²⁺ and Cu²⁺ existing in the same sample solution with different sensitivity range as mentioned above. It is concluded that the proposed method can be successfully applied to real water samples including drinking water and tap water Table 7.