Introduction

Heavy metal concentration in food samples is very important for human health (Afkhami et al. 2011; Cayir and Coskun 2010; Ieggli et al. 2010). This is the cause of various health problems, such as cardiovascular diseases, kidney failure, cancer, etc. (Kadıoğlu et al. 2010). It is important to accurately determine levels by using various analytical techniques. Techniques such as spectroscopic and electrochemical methods are used for metal analysis (Szymczycha-Madeja and Welna 2013; Üstündağ et al. 2007; Tormen et al. 2011). Electrochemical techniques on modified electrodes are especially very sensitive for metal solutions with very low concentrations (Wei et al. 2012a; Wang et al. 2014; Ensafi 2010).

The scientists have been discussing the impact of environmental exposure to toxic metals on human health especially caused by lead and cadmium for more than 30 years. General evaluation of the toxic influences of Pb and Cd on people living in polluted areas is obviously based on analysis of the amount these toxic elements in urine and blood (Minoia et al. 1990; Hamilton et al. 1994). Not only toxic metals such as Pb and Cd but also some essential metals like Mn, Ca, Fe, Zn, Ni, Co, Cu, Cr, Na and K generally accumulate on noils, hair, and teeth.

Modified electrode has become very attractive for nanosensors in the last two decades (Sanghavi and Srivastava 2011). Differential pulse voltammetry (DPV) (Ensafi 2010), square wave voltammetry (SWV) (Li et al. 2011), amperometry, and electrochemical impedance spectroscopy (EIS) are very highly recommended, sensitive, and low-cost techniques on the modified surfaces for quantitative analysis of various molecules (Ensafi et al. 2011) and metals. A generally modified surface can drastically improve the electrocatalytic properties of electroactive groups, increase the electron transfer rate, and increase the sensitivity of the electrode response (Omidinia et al. 2014). Various methods, such as electrochemical (Baraket et al. 2013), self-assembled monolayers (SAMs) (Güzel et al. 2010), chemical (Üstündağ and Solak 2009), and physical (Gupta et al. 2013) techniques have been reported for the electrochemical quantitative analysis in various environmental and food samples. Most of the transition metals, such as Cu, Cd, Co, Fe, Ni, Zn, etc., and a lot of other metals, such as Sn, As, Sb, Pb, Bi, etc., are analyzed via electrochemical voltammetric methods. In the literature, chemical analysis of heavy metals in food samples by using voltammetry has greatly increased in recent years. Wei et al. studied toxic metal composition in drinking water on AlOOH(boehmite)@SiO2/Fe3O4 porous magnetic microspheres modified glassy carbon (Wei et al. 2011) via square-wave anodic stripping voltammetry (SWASV). They reported that the LOD was calculated as 0.0103 and 0.0156 ng L−1 for Pb2+and Cd2+, respectively. Abbasi et al. investigated the performance of the Pb2+ and Cd2+ analysis on 2-mercaptobenzothiazole modified hanging mercury drop electrode via voltammetry (Abbasi et al. 2011). In this paper, linear calibration curves were established in the range of 0.5–70 and 0.2–30 ng mL−1, for Pb2+ and Cd2+, respectively, with a detection limit of 0.017 ng mL−1 for Pb2+ and 0.01 ng mL−1 for Cd2+. They used the rice, soya, and sugar as real samples in the article. Illuminati et al. investigated a new method for the determination of Cd2+, Pb2+, and Cu2+ in wine (Illuminati et al. 2013). They prepared an epoxy-impregnated graphite electrode. The LOD values were determined as 1.2 and 7.0 ng L−1 for Pb2+ and Cd2+, respectively. In another study, some trace metals were analyzed in a honey sample on a boron-doped diamond electrode via DPASV (Honório et al. 2014). LOD was calculated as 0.40 and 1.28 ppb for Pb2+ and Cd2+, respectively, in the article.

Although a definition for graphenes does not yet exist, graphene has given direction to the scientific world since its first synthesis. The name graphene is already confusing to scientists as graphite oxide and GO. Many researchers describe graphene as consisting of a single atomic sheet of conjugated sp2 carbon atoms and a 2D single layer of graphite (Goh and Pumera 2010; Loh et al. 2010). The GO has a C:O ratio of 2–3, and the same characteristic properties are observed in graphene. However, GO has functional (such as C=O, –OH, and –COOH) and aromatic groups with different properties compared to graphene. Reduced graphene oxide (rGO) is more aromatic than GO and has the same properties as graphene, which include electrical, thermal, and mechanical properties. Case graphite has well-stacked layers parallel to each other in a three-dimensional crystalline structure. The chemical bonds within the layers are covalent, with the same sp2 hybridization as other members of the graphene family. Graphite oxide has many layers of GO (Fitzer et al. 1995). Graphene and its derivatives are advantageous to science in a variety of technological areas, such as electronics (Novoselov et al. 2004; Gilje et al. 2007), sensors (Schedin et al. 2007), electromechanics (Bunch et al. 2007), solar cells (Wang et al. 2008), memory devices (Standley et al. 2008), hydrogen storage (Sofo et al. 2007), ultracapacitors (Stoller et al. 2008), field-effect transistors (FET), and biomedical applications (Geim and Novoselov 2007). In recent years, graphene-derivative nanomaterials and nanotubes have garnered much popularity in the field of electroanalytical study. This dramatic interest can be based on the use of electrocatalytically affected GOs (Eng and Pumera 2014). Several electrochemical nanosensors, based on graphene and its derivatives, have been developed for various samples, such as biological (Hasanzadeh et al. 2013) and environmental (Hasanzadeh et al. 2012). The physicochemical properties of GO provide researchers an opportunity for metal detection. The functional groups of GO or binding some molecules on GO provides for detection of heavy metals very easily (Lü et al. 2013). Li et al. (2014) studied Cd2+ and Pb2+ analysis on electroreduced GO (erGO) modified bismuth-film electrode by SWV. They calculated the LOD as 0.1 μg L−1 for Cd2+ and 0.05 μg L−1 for Pb2+. Shaoo et al. (2013) determined heavy metals on a reduced GO/Bi nanocomposites modified surface by DPV. In the study, LOD of Cd2+ and Pb2+ were determined as 2.8 and 0.55 μg L−1, respectively. Ouyang et al. (2011) analyzed Zn2+, Cd2+, and Pb2+ on the Hg-Bi included carbon nanotube modified GC electrode by SWV. They reported that the LODs for Cd2+ and Pb2+ were lower than 2 ppm and 0.12 ppb, respectively. In an article, Zhu et al. (2014) studied analysis of Pb2+ and Cd2+ on a gold nanoparticle–graphene–cysteine composite modified bismuth film electrode via voltammetry. LOD was calculated as 0.10 ppb for Cd2+ and 0.05 ppb for Pb2+.

The present work, therefore, aims to apply GO modified GC electrode as a working electrode for the simultaneous measurement of Pb2+ and Cd2+ concentrations by the DPASV method. The nanostructure was characterized using XPS, TEM, AFM, and electrochemical techniques. The sensitivity of response for the ions and the detection limits was also investigated to show the applicability of the method.

Experimental

All highest purity chemicals were obtained from Merck, Sigma-Aldrich, Fluka, or Riedel de Hean chemical companies. The water was purified in ultrapure water (UPW) by a resistance of 18.3 MΩ (Human Power 1+ purification system, S. Korea). All experiments were performed under a purified Argon gas (99.999 %) atmosphere. All experiments were performed at room temperature (25 ± 1 °C) in the triple electrode system. The reference electrode was an Ag/AgCl/KCl (sat) and Ag/Ag + (0.01 M) in aqueous and non-aqueous media, respectively. Pt wire was used as auxiliary. The GC working electrode was gathered from BAS (Bioanalytical Systems, MF-2012, USA) for electrochemical modification and characterization, such as CV and EIS measurements. The other surface characterization of the nanofilm was performed on a GC-20 (Tokai, Japan) electrode. Carbon electrodes were cleaned with 100 and 50 nm Al2O3 suspension (Baikowski Int. Corp., USA) on polishing clothes (Buehler, Lake Bluff, IL, USA) for approximately 5 min and then washed with UPW and acetonitrile (CH3CN). Polished GC electrodes were sonicated (Ultrasonic Cleaner, SK1200H, China) in UPW and isopropyl alcohol/ CH3CN (1:1, v/v) for approximately 5 min.

Preparation of GO and GC–O–GO

GO were prepared using the method described earlier (Hou et al. 2010). Briefly, the GO was synthesized from natural graphite by a modified Hummers method. To start, graphite powders were oxidized with sulfuric acid. Five grams of graphite oxide was placed in 25 mL of sulfuric acid; then 5 g K2S2O8 and 5 g P2O5 were added. The mixture was heated at 80 °C for 6 h. Then it was cooled to about 30 °C. The mixture was washed with water and sonicated to remove the residual acid. The product was kept under ambient conditions for 12 h. The pretreated graphite powder was added to cold (0 °C) H2SO4 (250 mL). Then KMnO4 (30 g) was inserted into the mixture by stirring at less 20 °C. When the addition was completed, the mixture was kept and stirred at 35 °C for 4 h, and then 500 mL UPW was added. During the dilatation step, the temperature of the oxidation reaction was carried out under 50 °C with an ice bath; 500 mL UPW was added and stirred for 2 h. Following the stirring, 1.5 L UPW and 40 mL %30 H2O2 were added. The color of the mixture changed to brilliant yellow and began to bubble. The GO suspension was centrifuged and washed with 0.1 M HCl three times to remove metallic pollution, followed by 1.0 L UPW water.

Two hundred milligram GO was diluted in 25 mL 0.01 M HCl. The bare GC was modified with the GO in the suspension solution under the positive potential for three CV cycles. The modified surface was denoted as GC–O–GO. The electrochemical measurements were acquired with a Gamry Reference 300 electroanalyzer (USA). The proposed structure of the GC–O–GO surface constructed by the electrochemical binding mechanism is shown in Scheme 1.

Scheme 1
scheme 1

Schematic diagram of GC–O–GO

Full size image

Characterization of GC–O–GO

CV and EIS characterization was carried out in 1 mM K3Fe(CN)6 + 1 mM K4Fe(CN)6 solution in 0.1 M KCl supporting electrolyte using the electroanalyzer system. The pKa of the GC–O–GO was determined with the redox couple by CV methods. The second redox probe characterization of a modified surface was performed with 1 mM ferrocene in 0.1 M tetrabutylammonium tetrafluoroborate (NBu4BF4) in CH3CN. EIS data of the Fe(CN)6 3−/4− redox couple was measured at 300 kHz to 0.1 Hz at 10 mV wave amplitude with 0.165 V of DC potential.

XPS measurements were performed by using a PHI 5000 Versa Probe (ϕ ULVAC-PHI, Inc., Japan/USA). The modified surface was characterized via an AFM microscopy (NT-MDT atomic force microscopy), with a non-contact mode of resonant freq. 150 kHz force constant 5 N/m. The GO was imaged using a JEOL 2100 HRTEM instrument (JEOL Ltd., Tokyo, Japan).

Determination of Simultaneous Cd2+ and Pb 2+ in Aqueous Media with DPASV

Cd2+ and Pb2+ ions were adsorbed on the GC–O–GO surface in 0.1 M acetate buffer solution with various pH, accumulation time, and temperature. The optimum condition of the sensor application was calculated via anodic stripping DPV. Optimization of pH, deposition time, and temperature were investigated for the range of 3.5–5.5 pH, 10–50 min, and 20–40 °C, respectively. Before DPASV measurements, Pb2+ and Cd2+ ions on the modified surface were reduced to metallic forms at −1.0 V for 10 s in 0.1 M acetate buffer solution. A traditional milk sample was 1:1 (v/v) diluted with UPW. The pH value of the milk sample was adjusted at 4.5 via 0.1 M acetate buffer solution. The tap water sample was taken from our research laboratory in Dumlupinar University without pretreatment before determination; the pH value was adjusted to 4.5 with 0.1 M acetate buffer solution. One gram of the rice and soya seed samples was weighed and powdered with an agate mortar. In a beaker, 15 mL of concentrated HNO3 was added to the soya and rice powdered samples and was kept overnight. The suspensions were evaporated near to dryness on a magnetic hotplate at about 120 °C for 4 h. The residues were dissolved in 0.5 M HNO3 and centrifuged to remove particles from the suspension (Rajabi et al. 2014). The pH of the media adjusted at 4.5 via 0.1 M acetate buffer solution. The measurements of the all samples were repeated five times (n = 5).

Results and Discussion

Electrode Modification and Characterization

The GO (TEM imaging of the GO is given in Fig. 1a) was electrochemically attached on GC electrode via CV voltammetry. Oxidation voltammogram of GO on GC electrode in 0.01 M HCl is given in Fig. 1b as four cycles. The modification process was completed after the first cycle. The modification of the GO on GC electrode was checked through a second experiment. Under the same conditions, the bare GC electrode was scanned in a positive direction in 0.01 M HCl without GO. The voltammogram is given in Fig. 1c. In an article, during the positive scanning, various functional groups such as –OH, C = O, and –COOH are composed on the GC electrode surface (McCreery and Bard 1991). Behavior of the ferrocene on GC, GC–ox, and GC–O–GO electrodes is shown in Fig. 2a. In the figure, it is explicitly shown that the electron transfer rate of ferrocene on GC–ox is faster than on GC–O–GO. The electrochemical behavior of Fe(CN)6 3−/4− redox couple on GC, GC–ox, and GC–O–GO electrodes in 0.1 M KCl in aqueous media is given in Fig. 2b. GC–O–GO surface shows a higher blocking ability than GC–ox for the Fe(CN)6 3−/4− electron transfer reaction. These two redox probe characterizations show that GO was covalently attached on a GC surface via electrochemical oxidation.

Fig. 1
figure 1

TEM imaging of the GO (a), cyclic voltammograms of presence of GO (b), and without GO (c) on GC in 0.01 M HCl media. Scan rate is 200 mV s−1 vs. Ag/AgCl(sat) reference electrode

Full size image
Fig. 2
figure 2

Cyclic voltammograms of 1 mM ferrocene (a) and Fe(CN)6 3−/4− (b) on GC, GC–ox, and GC–O–GO; cyclic voltammograms of 1 mM Fe(CN)6 3−/4− on GC and GC–O–GO at pH = 7.0 and 2.1 (c), scan rate is 200 mVs−1 vs. Ag/Ag+(0.01 M) and Ag/AgCl(sat) reference electrode, non-aqueous and aqueous media, respectively; cathodic peak currents versus pH plot of the surfaces in Fe(CN)6 3−/4− redox probe and derivative curve (d)

Full size image

The pH dependence of Fe(CN)6 3−/4− occurred when the pH of the bulk solution changed. Carboxylic acid groups on the modified surfaces have a major effect with the various pH of the buffer solution. The cyclic voltammogram of the redox probe on GC–O–GO is given in Fig. 2c. The electron transfer rate of the redox couple on GC–O–GO was changed with the pH of the media. When the pH of the solution was high enough from the pKa value of the surface, the carboxylic acid groups were charged as negative. The negatively charged GC–O–GO surface significantly decreased the negatively charged Fe(CN)6 3−/4− by repelling them with the electrostatic effect. The cathodic peak current data of the redox couple with various pH on the GC–O–GO surface is given in Fig. 2d. The surface pKa of the GC–O–GO was determined as 3.18 ± 0.30 by Gaussian fitting. In an article, pKa of colloidal GO was determined as 4.3 (Konkena and Vasudevan 2012). The carboxyl and hydroxyl groups involved with a graphene surface behave like carbon nanotubes, which probably have pKa values ranging from 3 to 5 (Tarley et al. 2006).

The experimental data of EIS was fitted with a diffusion-affected constant phase element (CPE; Y o and α value) or Warburg (Z w)-affected Randles equivalent circuit as shown in Fig. 3a–c. The experimental data was obtained at a 300-kHz–0.1-Hz frequency range using a 1 mM Fe(CN)6 3−/4− redox couple in 0.1 M KCl in aqueous solution by using EIS under 0.165 mV of DC potential. The fitting values are shown in Table 1. The resistance of the solution with supported electrolyte (R s) was determined to be between 85 and 140 Ω. The charge transfer resistances (R ct) of the redox couple on the GC electrode and GC–ox were nearly 3–4 kΩ. The R ct of redox couple probes on GC–O–GO was fitted as 120 kΩ. The GC–O–GC surface has exhibited different behavior according to GC electrode and GC–ox.

Fig. 3
figure 3

The Nyquist plots of GC (a), GC–ox (b), and GC–O–GO (c) and their fittings curves

Full size image
Table 1 Fitting values of Warburg affected CPE circuit for 1 mM Fe(CN)6 3−/4− redox couple on bare GC and GC–ox and GC–O–GO in 0.1 M KCl in aqueous media
Full size table

The formed GC–O–GO was confirmed by XPS measurements, which are effective tools to characterize the presence of C and O elements. The C1s and O1s binding energy spectra of the GC–O–GO surface are given in Fig. 4a, b. The narrow region high-resolution C1s core spectra of GC–O–GO fit as four peaks, which were O–C=O at 288.8 eV, C=O at 287.8 eV, C–O at 286.7 eV, and C–C at 284.9 eV in Fig. 4a (Tu et al. 2014). The O1s spectrum of the modified surface shows peaks at 533.8, 532.5, 531.8, and 530.7 eV, respectively, which could be assigned to C–C=O/O–C=O; C–OH/C–O, –OH (hydroxides), C=O, and other carbon complexes (Fig. 4b).

Fig. 4
figure 4

High-resolution XPS core spectra of GC–O–GC: C1s (a) and O1s (b); AFM images of bare GC (c) and GC–O–GO (d)

Full size image

AFM imaging of the bare GC electrode is given in Fig. 4c. The AFM images of the GO terminated thin film (GC–O–GO) are shown in Fig. 4d. The GC electrode surface has a few aluminum polishing matter-affected pinholes or scratches at the 1–10-nm scale. The imaging of the modified surface has nanometer and micrometer scaled changeable roughness because of horizontal or vertical located GO with a micrometer scale. Some regions of the modified surface are smoother than bare GC electrode surfaces.

Calibration Curves of Cd2+ and Pb2+ on the GC/GO

Some analysis parameters that affect the simultaneous analysis of the Pb2+ and Cd2+, such as the pH factor, deposition time, and deposition temperature, were optimized via DPV. The maximum current of the Pb2+ and Cd2+ were measured at pH 4.5 (Fig. 5a). The optimum deposition time was determined as achieved at 30 min (Fig. 5b). The incubation temperature of the ions was calculated as 28 °C (Fig. 5c).

Fig. 5
figure 5

Optimization, pH at 28 °C for 30 min (a), deposition time at 28 °C for pH = 4.5 (b) and deposition temperature at pH = 4.5 for 30 min (c) for Pb2+ and Cd2+

Full size image

Figure 6a shows DPASV responses of concentrations of Cd2+ and Pb2+ on the GO covalently modified carbon electrode surface and their Gaussian fits. The determination limits of Cd2+ and Pb2+ have been tested by DPASV measurements on the GO covalently modified carbon electrode in 0.1 M acetate buffer (pH 4.5) range of 1 × 10−6 M–1 × 10−12 M Cd2+ and Pb2+. The electrical signals were observed at −631.2 mV for Cd2+ and −329.2 mV for Pb2+, approximately. The signals of the stripping peaks of Cd2+ were more intensive than that of Pb2+; the stripping peaks of Cd2+ were twofold, according to the stripping peaks of Pb2+. This means the electron transfer of Cd2+ ions on the GO surface is faster than that of Pb2+ ions. Hence, the GC–O–GO surface features the highest sensitivity and the best stability for trace concentrations of Pb2+ and Cd2+. The carboxylate and hydroxyl groups of the GO surface can electrostatically attract Pb2+ and Cd2+ ions. This means that Pb2+ and Cd2+ ions also show chemical attraction with oxygen atoms on the GO surface (Wei et al. 2012b).

Fig. 6
figure 6

DPASV calibration voltammograms (a) of 1 × 10−8, 1 × 10−9, 1 × 10−10, 1 × 10−11, and 1 × 10−12 M Cd2+ and Pb2+ on the GC–O–GO and their Gaussian fittings (b); calibration curves of Cd2+ and Pb2+ on the GC–O–GO (c) and DPASV voltammogram for interference effects of the method (d) in aqueous media

Full size image

Figure 6a, b presents calibration voltammograms of different concentrations of Pb2+ and Cd2+ on the GO covalently modified carbon electrode according to Gaussian data. At the optimal conditions, simultaneous analysis of Cd2+ and Pb2+ was performed. Also, the analysis of Cd2+ and Pb2+ions was performed by changing the concentration of both species. As shown in Fig. 6c, the linear graphs of Pb2+ and Cd2+ are y = −0.2262x + 2.8834 (R 2 = 0.9973) and y = −0.5771x + 7.3078 (R 2 = 0.9959), respectively. The mutual responses of species are practically unaltered with the increase of another species concentration. The detection limits of Cd2+ and Pb2+ were 2.5 × 10−13 and 2.8 × 10−12 M, respectively. The Cd2+ calibration curve slope is especially higher than that of Pb2+. A GC–O–GO electrode surface is more selective for Cd2+ than Pb2+. LOD results of the methods are satisfactory as compared to the values given in many articles in the literature.

Three different concentrations of 1.0, 4.0, and 7.0 nM Pb2+ and Cd2+ in the linear range were analyzed in six independent series on the same day for intra-day precision and six consecutive days for inter-day precision from six measurements of every series. The precision and accuracy results are given in Table 2. The RSD or precision % values varied from 0.25 to 1.98 for intra-day and from 0.25 to 2.94 for inter-day precision. Accuracy of this method was calculated as the percent relative error. Both of the results obtained for intra-day and inter-day accuracy were ≤2.00 %.

Table 2 Precision and accuracy results of the methods for intra-day and inter-day
Full size table

Analysis of Food Samples

The applicability of the GC–O–GO electrode for the analysis of food and water samples with a different variety of real samples with different matrix was assessed by its application to the simultaneous determination of the ions in various real samples, including tap water, rice, soya, and milk. The analysis results of the samples are collected in Tables 3. The results obtained by the DPASV method show a good agreement with those obtained by ICP-OES. The results confirm the applicability of this method for precise and accurate determination of the Cd2+ and Pb2+ ions in a wide variety of some real samples with different complex matrices. The recovery values of Cd2+ and Pb2+ in real samples (n = 5) are shown in Table 2. The recoveries of the Cd2+ and Pb2+ were up to 98 %, so analytical recovery was obtained for all species and in the real samples.

Table 3 Simultaneous determination of Cd2+ and Pb2+ in real samples (n = 5)
Full size table

Interference Responses of Pb2+ and Cd2+ Presence in Cu2+ and Hg2+

Figure 6d presents DPASV responses of Pb2+ and Cd2+ in the presence of Cu2+ and Hg2+ on a GC–O–GO electrode surface. It is well known that matrix affects the ion-selective electrode response. Given strong signals in electrochemical methods, Cu2+ and Hg2+ metals can affect the response of the electrode (Okcu et al. 2005). We select Cu2+ and Hg2+, consciously, because signals of Cu2+ and Hg2+ are very good. The testing of the matrix effect provides our method a confidence for the matrix effect of Cu2+ and Hg2+. The peak current of Pb2+ and Cd2+ did not affect the equimolar range. The peak current responses of the ions were reduced by about 2 % by the presence of 100-fold excess of Cu2+ and Hg2+ ions.

Conclusion

The GO was synthesized via the electrochemical exfoliation method as described in the literature. The GO covalently modified on the carbon electrode surface. This prepared electrode was characterized by using TEM, AFM, and XPS. The modified electrode was electrochemically characterized with CV and EIS techniques, assisted with ferrocene and Fe(CN)6 3−/4− redox couples. The proposed method, using GO covalently modified to a glassy carbon surface, is proven to be efficient, sensitive, and rapid and can be used for Pb2+ and Cd2+ in food samples such as tap water, soya, milk, and rice. Simultaneous determination of Pb2+ and Cd2+ ions by DPASV is one of the most important advantages of the proposed method, which has been applied on food samples. Effects of this procedure were investigated for electrochemical parameters such as calibration curves and matrix effect. The detection limits of Pb2+ and Cd2+ were very low. The GO covalently modified carbon electrode surface shows very ultrasensitive response for these cations. The synergistic effect of the GO materials was obtained for Pb2+ and Cd2+ detection with improved sensitivity and reproducibility.