Introduction

The rise of various industries has been accompanied by a sharp increase in environmental contamination by toxic substances like heavy metals. Heavy metals are those metals having relatively large density like mercury, lead, arsenic, cadmium etc. discharged into the earth primarily through industrial effluents, volcanic emissions, fuel combustion etc. Interaction of these metal ions with the thiol group of biomolecules results in their cyanogenic impact on animals [1]. This builds the demand for the development of various refinement techniques and sensors for such pollutants. To monitor the presence of these metals, instrumental methods like Atomic Absorption Spectroscopy (AAS), Ion chromatography, Industrially Coupled Plasma Mass Spectrometry (ICP-MS) etc. are generally used.

Electrochemical sensors are good alternative for the spectroscopic and chromatographic techniques as the latter methods require complex methodology and significant expenses [2]. Various kinds of heavy metal sensors are available recently, created based on the chemical interactions between the electrode and metal ions. In order to improve selectivity in sensing application, chemically modified electrodes are used. Nanoparticles, organic molecules, polymer materials etc. and their composites can be used as electrode modifiers [3, 45, 6]. Graphene based nanoparticles like graphene oxide and graphene quantum dots show awesome electrical, optical and mechanical properties [7,8,9,10,11]. Several graphene based electrode modification were reported till date towards metal ion detection. The oxygen functionalities present in graphene oxide can act as metal binding site as well as it can be modified covalently or non-covalently, with a suitable molecule which can interact with the specific analyte [12,13,14,15,16,17,18,19].

Ion imprinting is a technique by which specific cavities can be introduced in the material. This ensures specificity and selectivity for a particular target analyte [20,21,22,23,24]. Combination of ion imprinting with nanomaterials is a good choice for developing sensors with high selectivity and sensitivity since nanomaterials improve electrical conductivity and imprinting provides better selectivity to the material. T. Alizadeh et al. have reported an imprinted polymer-MWCNT composite towards the detection of mercury(II) ions using itaconic acid as the monomer and fabricated a carbon paste electrode with the ITA-IIP/MWCNT composite [25]. An IIP composite with rGO was used by Motlagh et al. [26] for mercury(II) detection. The molecule 2,2′-(9E,10E)-1,4-dihydroxyanthracene-9,10-diylidene)bis(hydrazine-1-carbothioamide) (DDBHCT) was synthesised by them for chelating with mercury(II) ion. Methyl methacrylic acid was taken as the functional monomer, EGDMA as cross linker and ammonium persulfate as the initiator for imprinted polymer formation. They achieved a detection limit of 0.02 μg/L using SWASV. In these reported cases either the ligand synthesis involved many steps or the cost of CNT would be a concern.

Compared to carbon nanotubes, graphene is a cheaper and easily synthesizable material. Herein we report a novel nanosensor based on GQD composite with an ion-imprinted polymer, synthesized by a simple method which is used for modification of glassy carbon electrode for the electrochemical sensing of mercury(II) ions.

Experimental

Materials and reagents

Graphite fine powder was obtained from Loba Chemie. 4-vinyl aniline, Ethylene glycol dimethyl acrylate (EGDMA), Polyvinyl alcohol (PVA), Azobisisobutyronitrile (AIBN), and metal salts were purchased from Aldrich and Nafion was purchased from Merck. All other reagents utilized were of analytical grade. Acetate buffer solution was prepared by mixing sodium acetate and glacial acetic acid in appropriate quantities. It was used for all electrochemical measurements.

Apparatus

The materials synthesised were characterised by Fourier Transform Infrared spectroscopy (JASCO FT-IR- 5300 spectrometer) using KBr pellets. The surface morphology of the material was determined using SEM (Jeol JSM-6390LV/JED–2300) and TEM (HR-TEM Jeol/JEM 2100) imaging. To analyse the elemental composition of the material, CHNS analyser was used (Elementar Vario ELIII CHNS analyser). Electronic spectra of the sample is recorded using Thermo scientific Evolution 220 UV–Vis spectrophotometer. Electrochemical measurements were carried out in CHI 660E electrochemical workstation (CH instruments) with a three-electrode system in which GQDTU-IIP modified glassy carbon was used as the working electrode, Ag/AgCl as reference electrode and a Pt wire as counter electrode in acetate buffer solution (pH 5). ICP-MS analysis was done using Thermo Scientific, iCAP RQ ICP-MS instrument (Helium KED mode). For the GQD synthesis, probe sonicator (SONICS vibra cell model-VCX 750) was used. Thermogravimetric analyses were done using a Metler 851 Model SDTA/TGA instrument at a heating rate 20 °C min−1. X-ray photoelectron spectroscopy was recorded using PHI 5000 Versaprobe Scanning Esca Microprobe.

Synthesis of Graphene Quantum Dot (GQD)

GQD was synthesised from Graphene Oxide (GO) using probe sonication method according to a reported procedure. For this GO was synthesised initially using Tour method [27]. 1 g of NaOH was dissolved in 30 mL of ethyl acetoacetate by stirring for 30 min. It was filtered and to the filtrate about 0.2 g of GO was added and kept for ultra-sonication for 2 h under 10 kHz frequency (50%). The dispersion obtained was centrifuged to remove unexfoliated GO. The centrifugate was then dialysed and the solution outside was collected as GQD solution which exhibited strong fluorescence. It was lyophilised to obtain GQD in powder form [28].

Functionalization of GQD to form the ligand monomer (GQDTU)

About 0.1 g of GQD powder was treated with 4 mL thionyl chloride to convert the acid functionalities (-COOH) of GQD into acid chlorides [29]. The acid chloride modified GQD (GQD-COCl) was dissolved in dried acetone. About 0.08 g of dry ammonium thiocyanate was dissolved in dried acetone and cooled to 0–5 ˚C. To this, a cold solution of GQD-COCl in dried acetone was added drop wise by keeping the temperature below 5 ˚C. After complete addition it was stirred for 1 h to permit complete precipitation of ammonium chloride. The ammonium chloride formed was sifted through rapidly. The filtrate was taken in a round bottom flask, cooled and kept for stirring. To this a cold solution of 4-vinyl aniline (0.24 g in dried acetone) was added drop wise. Mixing was continued for 60 min. The resultant mixture was poured over a large quantity of ice water. A yellow brown precipitate was formed which was separated and dried in air [30].

Polymerisation of functionalised GQD monomer (GQDTU)

The polymer was prepared via suspension polymerisation method [31]. 0.1 g of polyvinyl alcohol (PVA) was dissolved in about 30 mL hot water and cooled to room temperature. The polymerization mixture was prepared by taking Template (Hg2+): Monomer (GQDTU): Cross linker (EGDMA) in the proportion 1:2:4. About 0.5 mL of free radical initiator Azobisisobutyronitrile (AIBN) was added to it. This mixture taken in a round bottom flask was subjected to N2 purging. To this, PVA solution was added drop wise with stirring. It was kept in an oil bath with vigorous stirring at 80 ˚C in N2 environment for 2 h. The stirring was continued overnight. The beads obtained were collected, washed and dried to get the polymer GQDTUP [32, 33].

Removal of the imprint ion from the polymer

Hg(II) ions were leached out from the polymer beads by stirring with 100 mL of 50% HCl for 18 h [31]. The beads were dried in an oven to get the ion imprinted polymer material (GQDTU-IIP). To confirm the removal of the template Hg(II) ions, ICP–MS analysis was carried out by digesting the material in nitric acid followed by filtration and dilution.

Electrode fabrication

Glassy carbon electrode (GCE) was cleaned using alumina powder (0.05 micron).Then adsorbed alumina was removed by sonication using methanol, 1:1 HNO3, acetone and distilled water successively and the surface was dried in air. For electrode modification about 0.1 g of GQDTU-IIP was dispersed in 100 μL nafion solution. 3 µL of this solution was used for drop casting in the cleaned GCE surface and kept for drying in air for 2 h. The modified electrode was kept at 4 ˚C until use [34, 35].

Electrochemical detection of Hg(II) ions

The detection of Hg(II) ions was carried out using the three electrode system in which the modified GCE was used as the working electrode, Ag/AgCl as the reference electrode and Pt wire as the counter electrode. The electrode characterisation was done using Cyclic Voltammetry (CV) and Electrochemical Impedance Spectroscopy (EIS) in a solution of [Fe(CN)6]4−/3− (5 mM) and 0.1 M KCl. Differential Pulse Voltammetry (DPV) measurements were carried out in acetate buffer electrolyte of pH 5 from -0.4 V to 0.6 V. Pulse amplitude, width and period used were 0.05 V, 0.05 s and 0.5 s respectively [36].

Real sample analysis

For the real sample analysis, water was collected from Periyar river near industrial area, Eloor, Kochi, Kerala, India and tap water sample was collected from residential area. The samples were filtered using Whatman No.1 filter paper to remove bigger particles. The pH of the samples were then adjusted to 5. Different concentrations of Hg(II) ions were spiked into the samples and analysed using the developed sensor [37, 38].

Results and discussion

Synthesis and characterisation of GQDTU-IIP

The synthesis of functionalised GQD and subsequent polymerisation to form the electrode modifier GQDTU-IIP is shown in Scheme 1. The polymer before (GQD–TUP) and after the removal (GQDTU–IIP) of the ions were digested and ICP–MS analysis was carried out to confirm the removal of the imprinted Hg(II) ions. The result obtained is given in Table 1, which shows the removal of Hg(II) ion from the polymer material. In order to understand the absorption capacity of GQDTU–IIP, 0.08 g of the polymer material was dispersed in 20 mL of 25 ppm mercury solution and stirred for about 6 h. Then the material was filtered out, washed with distilled water multiple times and dried in an oven to get GQDTU–IIP2. The amount of Hg(II) ions up taken was determined by ICP-MS analysis after the digestion of the polymer. The result obtained as given in Table 1 which shows that the material has good absorption capacity, confirming the coordination of Hg(II) ions into the imprint sites.

Table 1 ICP–MS analysis results obtained for GQDTU–IIP, GQD–TUP and GQDTU-IIP2
Full size table

The modified GQD and the polymer material were characterised using FT-IR analysis. Figure 1 shows the FT-IR spectra of GQD (a), GQDTU monomer (b) and GQD-TUP polymer(c). The peaks in the range of 1723 cm−1, 1604 cm−1, 1345 cm−1, 1116 cm−1 etc. in the case of GQDs corresponds to C \(=\) O stretching, C \(=\) C stretching, O \(-\) H bending of carboxyl group and C \(-\) O stretching vibrations respectively [37, 38]. The amide C \(=\) O bond formation is confirmed by the peak observed at 1670 cm−1 in the functionalised GQD (b). In this spectrum, the peaks near 3500–3420 cm−1 is due to antisymmetric stretching vibrations of N \(-\) H bond and that between 3420–3340 cm−1 is due to the symmetric N \(-\) H stretching. C \(-\) N stretching is found near 1350 cm−1 [39, 40, 30, 41, 42]. It provides information regarding the functionalisation of GQDs. The IR spectra of polymer material confirms the EGDMA cross-linked polymer formation. The peaks at 2959 cm−1 and 2870–2840 cm−1 in the spectrum (c) corresponds to antisymmetric and symmetric stretching vibrations of CH2 bonds respectively. The peak at 1725 cm−1 is due to C \(=\) O bond [43].

Fig. 1
figure 1

FT-IR spectra of (a) GQD (b) functionalised GQD (GQDTU) (c) GQDTUP composite

Full size image
Scheme 1
scheme 1

a. Functionalisation of GQD b. synthesis of the imprinted polymer material GQDTU-IIP

Full size image

SEM and TEM analysis were used to carry out surface characterisation. Figure 2a and b show the TEM images of GQDs and GQDTU monomer [44]. The mean particle size of GQD is found to be 9.5 nm and that of functionalised GQD is about 18 nm in diameter. The SEM image of the polymer material is also shown in Fig. 2c. The suspension polymerisation resulted in bead like polymer material having almost uniform size. This is used to modify the electrode surface. The FE–SEM image of modified electrode using about 3 µL of the polymer dispersion in nafion is shown in Fig. 2d. The image shows the uniform distribution of polymer beads on the electrode surface [43, 45].

Fig. 2
figure 2

(a) TEM image of GQD synthesised via ultra-sonication process and (b) TEM image of GQDTU (c) SEM image of polymer GQD-TUP (d) SEM image of the electrode surface modified with GQD-TUIIP inset of a and b shows histograms of GQD and functionalised GQD respectively

Full size image

The composition of the functionalised GQD was verified using CHNS analysis. The result obtained is given in Table 2 indicating the presence of N and S atoms which could be attributed to the successful functionalisation of GQDs [46].

Table 2 CHNS analysis result for the functionalized GQD (GQD-TU)
Full size table

The chemical composition of GQD and GQDTU-IIP were studied using X-ray photoelectron spectroscopy. The wide spectrum of GQD shows two prominent peaks at 284.9 eV and 531.3 eV binding energy corresponding to C1s and O1s respectively. In the polymer material (GQDTU–IIP) in addition to C1s peak (at 284.6 eV) and O1s peak (532.6 eV) two more peaks are present at 398.2 eV and 166.2 eV corresponding to N1s and S2p signals respectively (Fig. 3a and b). De-convoluted C1s spectrum of GQD and GQDTU–IIP is shown in Fig. 3c and d. For GQD, the peaks corresponding to C \(-\) C/C \(=\) C appears at 284.8 eV, that of C \(-\) O appears at 285.7 eV and C \(=\) O signals appears at 288.2 eV. In the polymer material, C \(-\) C/C \(=\) C appears at 284.6 eV, C \(-\) O at 286.2 eV and C \(=\) O at 288.5 eV binding energy. O1s signal formed at 532 eV and 533.5 eV for GQD (Fig. 3e and f). While for GQDTU-IIP it is shifted to 531 eV and 535 eV. The peak corresponding to N1s and S2p are very weak (Fig. 3g and h). This may be due to the formation of the cross linked polymer material which interrupt the signal from sulphur and nitrogen. Compared to the weight percentage of other elements in the matrix, the weight percentage of sulphur and nitrogen is very low. This also can be a reason for the weak signal in the spectrum [45,46,47,48].

Fig. 3
figure 3

(a) XPS spectra of GQD (b) GQDTU-IIP material (c) C1s spectra of GQD and (d) GQDTU-IIP (e) O1s spectra of GQD and (f) GQDTU-IIP (g) N1s and (h) S2p spectra of GQDTU-IIP

Full size image

Thermogravimetric curve of the material GQDTU–IIP, GQD and functionalised GQD monomer (GQDTU) are shown in Fig. 4. The initial weight loss of GQD below 100 °C is due to the elimination of moisture or solvent. In the case of GQDTU the initial weight loss is much lower than that of GQD. The weight loss above 150 °C is due to the decomposition of oxygen functional groups. In the case of GQDTU, the weight loss above 150 °C is only about 10% which can be due to the modification of oxygen functional groups. Compared to GQD, GQDTU shows much higher thermal stability. The polymer composite, GQDTU–IIP also showed thermal stability more than that of GQD. Degradation of the polymer material occurs in the temperature range 290–380 °C with maximum weight loss at ~325 °C. Thus the polymer GQDTU–IIP is found to be stable up to ~300 °C [49,50,51,52].

Fig. 4
figure 4

TGA curves of GQD, functionalised GQD (GQD–TU) and GQDTU-IIP composite

Full size image

Electrochemical characterisation of GQDTU-IIP modified glassy carbon electrode

The electrochemical characterisation of the proposed sensor was carried out using CV and EIS measurements. In the impedance spectra the semi-circular part corresponds to the electron transfer process and the linear part corresponds to the diffusion process. The radius of semi-circular part of Nyquist plot gives the charge transfer resistance. For comparison another imprinted polymers BAIIP and GOIIP was also prepared using benzoic acid and graphene oxide respectively instead of GQD by following the same procedure. Non imprinted polymer (NIP) was prepared corresponding to GQDTU-IIP in the absence of the template ion during the synthesis. The modified electrodes using these polymers were also characterised using EIS. Figure 5 shows the Nyquist plot for GQDTU–IIP–GCE, BAIIP–GCE, GOIIP–GCE and NIP–GCE electrodes taken in a 1:1 solution of [Fe(CN)6]3−/4− (5 mM) and KCl (0.1 M) [51]. The charge transfer resistance of GQD–TUIIP was found to be lower when compared to that of GOIIP and BAIIP. For the non-imprinted polymer (NIP) also the charge transfer resistance is higher which can be due to the inhibition of electron transfer between [Fe(CN)6]3−/4− and GCE surface.

Fig. 5
figure 5

Nyquist plots of different modified electrodes GQDTU–IIP–GCE, BAIIP–GCE, NIP–GCE and GOIIP–GCE

Full size image

Figure 6 shows the CV plot of bare GCE, GQDTU–IIP–GCE and NIP-GCE towards Hg(II) ions [52,53,54,55]. The peak shifted towards lower potential (0.38 V to 0.32 V) with enhanced current response of about six times (3 μA to 17 μA) obtained with the GQDTU–IIP–GCE which confirms effective modification, whereas no considerable response was seen in the case of NIP–GCE. This can be due to the presence of imprinted cavities at the GQDTU–IIP modified GCE.

Fig. 6
figure 6

Cyclic voltammograms of different modified electrodes GQDTU–IIP–GCE, NIP–GCE and bare GCE towards Hg(II) ions in acetate buffer of pH 5

Full size image

Sensing performance of the modified electrode

The detection of Hg(II) ions using the GQDTU-IIP modified electrode was studied using CV as well as DPV techniques. CV and DPV were taken in acetate buffer solution of pH 5. The cyclic voltammograms of the modified electrode with increasing concentration of Hg (II) ions is shown in Fig. 7a. The cyclic voltammogram shows an oxidation peak near 0.32 V in the presence of Hg(II) ions. The results show linear increase in current with increase in concentration of Hg(II) ions. In the absence of Hg2+ ions, no redox peaks were observed in the voltammogram. The corresponding calibration plot is shown in Fig. 7b. The linear response was obtained from 6 × 10–8 M to 8.5 × 10–7 M and 1.4 × 10–6 M to 8 × 10–6 M with regression equations,

Fig. 7
figure 7

(a) Cyclic voltammograms of the sensor in response to 60 × 10–9 M to 15 × 10–6 M Hg(II) respectively (100 mV/s scan rate) (b) Calibration graph

Full size image
$$\mathrm y=0.0074\mathrm x+0.4878\;\left(\mathrm R^2=0.98\right)$$

and

$$\mathrm y=0.0023\mathrm x+5.6393\;\left(\mathrm R^2=0.97\right)$$

From this the LOD was calculated using the equation,

$$\mathrm{LOD}=3\sigma/\mathrm S$$

where σ is the standard deviation of blank and S is slope of the calibration plot. Thus the detection limit obtained was 30.2 nM. There is a very small reduction peak found at the potential 0.15 V which shows that the interaction of the Hg(II) ions with the imprint sites on the electrode is quasi reversible [56,57,58].

The DPV graph of the sensor performance towards Hg(II) ions is shown in Fig. 8a. The DPV also shows a linear increase in current with increase in concentration of Hg(II) ions in the range 5 × 10–8 M to 2.3 × 10–5 M (R2 = 0.984, y = 0.0001x + 1.084). The corresponding calibration plot is shown in Fig. 8b. The LOD obtained from DPV technique is 23.5 nM. The mechanism of the sensing was studied using scan rate and the result is shown in Fig. 9a and b. The plot of peak current versus square root of scan rate shows a linear relation graph. This shows the electron transfer is diffusion controlled process. Table 3 shows comparison of recent reports on Hg(II) detection by ion imprinted polymer composite based electrode. 

Fig. 8
figure 8

(a) Differential pulse voltammograms of the sensor in response to increasing concentration of 50 nM to 23 µM Hg(II) ions. (100 mV/s scan rate) (b) Calibration graph

Full size image
Fig. 9
figure 9

(a) CVs of GQDTU-IIP sensor at various scan rates (10 to 100 mV/s) studied in 0.1 M KCl containing 5 mM [Fe(CN)6]3−. (b) Calibration graph of square root of scan rate vs peak current at various scan rates

Full size image

It is found that the modified electrode shows no much change in response after 10 days (RSD value 7.1%). The successive measurement upto three times also shows comparable results with RSD value 4.78%. This indicates that the sensor shows good stability and repeatability. Reproducibility was checked by taking the response of three different electrodes and the RSD value 2.93% shows the results are reproducible.

Optimisation studies

Buffer was chosen as the supporting electrolyte for the electrochemical analysis. Different buffers like acetate buffer, phosphate buffer, citrate buffer and HCl buffer of pH 5 were made and corresponding electrochemical response towards 4 µM concentration of Hg(II) ions were measured using DPV technique. The result shows that acetate buffer shows better current response compared to other buffers and hence it was chosen for all other studies (Fig. 10a). Then the pH of the buffer was optimised by measuring corresponding electrochemical response at different pH and was found that buffer of pH 5 exhibited highest current response and hence it was chosen for further studies (Fig. 10b). The incubation time was studied from 0 to 10 min and found that about 8 min gave highest current response and above that time the current remained same and then decreased after some more time (Fig. 10c). The effect of graphene quantum dots in the analytical performance was studied by comparing the resulting current response with imprinted polymer made using the same procedure using benzoic acid (BAIIP) and graphene oxide (GOIIP). The result shows that a well defined peak was obtained by using GQDTU-IIP compared to the other polymer material modified electrode and bare electrode. The study was done by using DPV technique (Fig. 10d).

Fig. 10
figure 10

(a) The DPV graphs of GQDTU-IIP modified electrodes towards Hg(II) ions in different buffers of pH 5 as supporting electrolytes (b) Effect of pH on peak current (c) Effect of incubation time on peak current (d) DPV response of electrodes modified using GOIIP, BAIIP, GQDTU–IIP and Bare GCE towards Hg(II) ions

Full size image

Selectivity studies

The selectivity of the developed sensor was analysed in the presence of interfering ions like Pb2+, Cd2+, Cu2+, Ni2+, Mn2+, Co2+, Zn2+, As3+, Al3+ and Fe2+ using DPV. Initially the modified electrode was taken in the buffer solution and the response was measured using DPV. Then 5 µM concentration of Hg(II) was added into it and the corresponding voltammogram was recorded. Then the current response at 0.28 V towards each metal ions of concentration equivalent to that of Hg(II) was measured separately. The result is summarised in the Fig. 11. The study shows the peak current obtained by the common interfering ions relative to Hg(II) is much lower than that of Hg(II) ion in the potential range.

Fig. 11
figure 11

DPV response of GQDTU-IIP modified GCE towards different metal ions

Full size image

Real sample analysis

The modified electrode was used to study the presence of mercury ions in real water samples containing 0.2 µM Hg(II) ions using spike recovery test. The result obtained is summarised in Table 4. The study shows a satisfactory result to use the sensor towards real sample analysis.

Table 3 omparison of recent reports on Hg(II) detection by ion imprinted polymer composite based electrode
Full size table
Table 4 Real sample analysis in River water and Tap water by response current measurements (n = 3) with different spike concentration under optimum conditions
Full size table

Conclusion

Nanobeads of Graphene Quantum Dot based ion imprinted polymer (GQDTU–IIP) has been prepared through a simple synthesis procedure. GQDTU–IIP based electrochemical sensor has been successfully developed. The ion imprinted material was used for the sensing of mercury ions with a linear range of response from 5 × 10–8 M to 2.3 × 10–5 M using DPV. The sensor provided a limit of detection of 23.5 nM using DPV technique with reproducible results. The imprinting process has enhanced the selectivity of the material towards Hg(II) ions with negligible interference from other heavy metal ions. The electrochemical sensor was used for the analysis of water samples collected from river and tap. The results obtained show that the sensor can be reliably used for the sensing of Hg(II) ions in real samples.