Study the voltammetric behavior of 10-Hydroxycamptothecin and its sensitive determination at electrochemically reduced graphene oxide modified glassy carbon electrode

2.1 Apparatus and reagents

Model CHI650 electrochemical system (Chenhua Instrument Company, Shanghai, China) was employed for electrochemical techniques. A standard three-electrode electrochemical cell was used for all electrochemical experiments with a bare GCE or modified electrode (d = 3 mm) as a working electrode, a platinum (Pt) wire as an auxiliary electrode and an Ag/AgCl as a reference electrode.

All reagents were of analytical reagent grade. 10-Hydroxycamptothecin was purchased from Aladdin Industrial Corporation (Qigang Rd, Fengxian, Shanghai). Graphite powder was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Double distilled water was used for all preparations. A stock solution of HCPT (1.0 × 10⁻³ mol L⁻¹) was prepared with ethanol and stored at 4 °C darkly. Dilutions were done just before use. Phosphate buffer solution (PBS, pH 2.0) was used as the supporting electrolyte for cyclic voltammetry (CV) and liner sweep voltammetry (LSV). Each assay was performed at room temperature.

2.2 Synthesis of graphene oxide

Firstly, the graphene oxide was prepared from natural graphite powder according to the literature (Prabakaran and Pandian, 2015) with a little modification. Then, exfoliation of graphite oxide to GO was achieved by ultrasonication of the dispersion for 40 min (1000 W, 20 μm amplitude) according to a previous report with a slight modification (Wang et al., 2009). A homogeneous aqueous dispersion of GO (1.2 mg mL⁻¹) was obtained.

2.3 Electrode pretreatment and ERGO-modified procedure

Prior to modification, the GCE was polished with finer emery paper and 0.1 μm alumina slurry to get a mirror surface, and successively rinsed thoroughly with acetone, ethanol, and distilled water in an ultrasonic bath respectively, each for 1 min. The GO/GCE was established by depositing the above – composition (6 μL, 1.2 mg mL⁻¹ GO solution) on a fresh GCE surface using a micro-injector, and then the solvent was evaporating for 15 min naturally. The ERGO/GCE was obtained by treating GO/GCE in a 0.1 mol L⁻¹ deoxygenized phosphate buffer solution (PBS, pH 7.0). The cyclic voltammetry was performed within potential window between 0 V and −1.5 V for 10 cycles with a scan rate of 0.1 V s⁻¹. When not used, the ERGO/GCE was stored darkly under 4 °C.

2.4 Analytical procedure

The quantitative analysis of HCPT was carried out in PBS buffer solution (pH 2.0) at room temperature unless otherwise specified. Then, ERGO/GCE was placed into the test solution and the cyclic voltammetry (CV) or linear sweep voltammetry (LSV) was performed. In determination, an accumulation step was performed under open circuit along with solution stirring for 180 s. After each determination, the ERGO/GCE went for successive one cycle between −0.1 and 1.3 V in 0.1 mol L⁻¹ PBS (pH 8.0) to give a regenerated electrode surface.

3.1 Electrochemical properties of ERGO/GCE

In general, K₃[Fe(CN)₆] is used as an electrochemical probe to investigate the properties of the pretreated electrode surface. Here, we used it as the proof to characterize the ERGO/GCE by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS), respectively. Fig. 1 shows the cyclic voltammograms of K₃[Fe(CN)₆] (1.0 × 10⁻³ mol L⁻¹ solution containing 0.1 mol L⁻¹ KCl) at bare GCE (curve a), GO/GCE (curve b) and ERGO/GCE (curve c) respectively. Well-defined CV, characteristic of a diffusion-limited and reversible electron transfer redox process, was observed at the bare GCE (ΔE_p = 70 mV). When the electrode was coated with GO, both the redox peak currents and the quasi-reversibility (ΔE_p = 115 mV) of Fe(CN)₆³⁻ were decreased, which could be attributed to the negatively charged oxygen-containing groups from GO, blocking the diffusion of Fe(CN)₆³⁻ from solution to the electrode surface. Nevertheless, after GO was electrochemically reduced, the redox peak current increased significantly compared with bare GCE and ΔE_p = 68 mV, indicating that ERGO can effectively increase the electron transfer rate of Fe(CN)₆³⁻ due to its electric conductivity and the increase of electrode surface area. Fig. 2 presents the Nyquist diagrams of the GCE (curve a), GO/GCE (curve b) and ERGO/GCE (curve c) in 5.0 × 10⁻³ mol L⁻¹ Fe(CN)₆^3−/4− (1:1) containing 0.1 mol L⁻¹ KCl. It can be seen that a small well-defined semi-circle at higher frequencies was obtained at the bare GCE. This indicated that a small interface electron resistance (R_ct) was existed. When GO was deposited on the GCE surface, the R_ct increased obviously. This phenomenon could be attributed to the GO film itself, and it indicated that a resistance into the electrode/solution system was existed. However, the R_ct decreased remarkably after GO was electrochemically reduced to ERGO, which suggested the excellent electro-conductibility of ERGO. In further research, an equivalent circuit was designed (insert of Fig. 2), and the R_ct obtained were about 25.38 Ω, 219.8 Ω, and 12.87 Ω for the GCE, GO/GCE and ERGO/GCE, respectively. This result also demonstrated that the GO was successfully reduced on the GCE surface just as designed.

Close

Figure 1 Cyclic voltammograms of [Fe(CN)₆]³⁻ (1.0 × 10⁻³ mol L⁻¹) containing 0.1 mol L⁻¹ KCl at bare GCE (a), GO/GCE (b) and ERGO/GCE (c) with a scan rate v = 0.1 V s⁻¹.

Export to PPT

Close

Figure 2 Nyquist plots of [Fe(CN)₆]^3−/4– (5.0 × 10⁻³ mol L⁻¹) containing 0.1 mol L⁻¹ KCl at bare GCE (a), GO/GCE (b) and ERGO/GCE (c). EIS condition: frequency range: 100 kHz–0.01 Hz; perturbation amplitude: 5 mV; the inset are the equivalent circuit and the zoomed image of the curves a and c.

Export to PPT

3.2 Electrochemical behavior of HCPT at ERGO/GCE

Fig. 3 illustrated the cyclic voltammograms (CVs) of HCPT (3.0 × 10⁻⁶ mol L⁻¹) obtained at the bare GCE (curve a), GO/GCE (curve b), and ERGO/GCE (curve c) in 0.1 mol L⁻¹ PBS (pH 2.0). In order to explain all phenomena thoroughly, each scan was performed 2 cycles and the enlarged curves a and b were shown as inset in Fig. 3. HCPT showed electrochemical activation at all these three electrodes. At the bare GCE and GO/GCE, weak electrochemical reactivity could be discerned (Fig. 3, curve a, b). At the ERGO/GCE, a sensitive anodic peak at 0.986 V (P1) and a cathodic peak at 0.311 V (P2) were presented in the 1st cycle. In the 2nd cycle, a new anodic peak at 0.345 V (P3) was appeared and the P1 was disappeared. The P2 was no change. From these experimental data, we preliminary estimated that the P2 and P3 were a pair of redox peaks and its active group came from the product of P1 oxidation. The P1 was an irreversible anodic peak. The electrochemical properties of HCPT obtained at the ERGO/GCE were very different from the previous reports (Zhang, 2005; Sun et al., 2006). In these two studies, there was only one anodic peak appearance, which should be the P1 in this study.

Close

Figure 3 Cyclic voltammograms of HCPT (3.0 × 10⁻⁶ mol L⁻¹) at bare GCE (a), GO/GCE (b) and ERGO/GCE (c); Inset: the zoomed image of the curves a and b. Supporting electrolyte: 0.1 mol L⁻¹ PBS (pH 2.0); Scan rate: 0.1 V s⁻¹.

Export to PPT

To further investigate the electrochemical properties of HCPT in detail, the scan potential window was controlled within different ranges to observe the electrode response of HCPT. Firstly, when the potential window was controlled between −0.1 V and 0.65 V for 4 cyclic scans, there was no any redox peak appearance (Fig. 4A). Secondly, when the potential window was set between 0.65 V and 1.3 V for 4 cyclic scans, the P1 was observed in the first cycle and then no any peak was obtained (Fig. 4B). Thirdly, when the potential window was controlled between 1.3 V and −0.1 V, the initial potential was set at 0.65 V going negatively for 4 cycles (Fig. 4C). Just as expected, the P2 was not observed in the first cycle and the three peaks were obtained just as the curve c in Fig. 3. Finally, successive 4 cyclic scans were performed between −0.1 V and 1.3 V (Fig. 4D). The P1 disappeared after the first cycle and the peak currents and potentials of P2 and P3 almost kept unchanged in the next 3 cycles. These data demonstrated the above conjecture. That is, the P2 and P3 were a pair of redox peaks and its active group came from the product of P1 oxidation.

Close

Figure 4 Voltammograms of HCPT within different potential window. (A) Between −0.1 V and 0.65 V; (B) between 0.65 V and 1.3 V; (C) and (D) between −0.1 V and 1.3 V. Other conditions were the same as in Fig. 3.

Export to PPT

3.3 Effect of pH and scan rate

To further elucidate the electrode reaction of HCPT at the ERGO/GCE, the influence of solution pH and scan rate (v) were investigated by CV. The data of P1 were obtained from the first cycle and P2 and P3 were from the second cycle. Fig. 5A displays CVs of HCPT (3.0 × 10⁻⁶ mol L⁻¹) in 0.1 mol L⁻¹ PBS with different pHs, ranging from 2.0 to 6.0. The peak potentials of P1, P2 and P3 shifted negatively with increasing pH. Plots of peak potentials versus solution pH were found to be linear over the pH range of 2.0–6.0. And the linear regression equations were: E_P1 (V) = −0.0365pH + 1.065 (R = 0.9949), E_P2 (V) = −0.0553pH + 0.424 (R = 0.9982), E_P3 (V) = −0.0547pH + 0.463 (R = 0.992) (Fig. 5B). From the slope of −0.0365 V pH (P1), we knew that the process of P1 involved proton and electron in a ratio of 1:2. And the slope of −0.0553 V pH⁻¹ (P2) and −0.0547 V pH⁻¹ (P3) was close to the Nernst slope of 0.059 V pH⁻¹ at 25 °C. This result indicated that there were equal number of proton and electron taking part in the redox of P2 and P3.

Close

Figure 5 (A) The cyclic voltammograms of HCPT (3.0 × 10⁻⁶ mol L⁻¹) with different pH PBS (curves a to i: 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0). (B) pH-dependences of peak potential (a was for P1, b was for P2, c was for P3). pH of PBS. Other conditions were same as in Fig. 3.

Export to PPT

It was also observed that the peak current of HCPT at the ERGO/GCE decreased along with the increase of pH from 2.0 to 6.0. In particular in pH 6.0, the electrochemical response of HCPT was very small. Hence, pH 2.0 of PBS was chosen as the optimum solution in further analyses.

The effect of scan rate on the redox of HCPT at the ERGO/GCE was investigated and voltammograms were shown in Fig. 6A. Likewise, every CV scan was cycled twice. With the scan rate increasing, the currents of three peaks increased and the peak potential of P1 and P3 shifted positively and P2 shifted negatively. For the P1, a good linear relation between peak currents and scan rates could be described by the following equation: i_p1 = −8.629v − 129.13 (i_p1 in μA, v in V s⁻¹, R = 0.9978), suggesting that the anodic peak P1 was controlled by adsorption. Meanwhile, a good linear relationship was exhibited between peak potential (E_p1) and ln v (Fig. 6B): E_p1 = 0.0302ln v + 1.0579 (E_p1 in V, v in V s⁻¹, R = 0.9999). According to Laviron theory (Laviron, 1979) for an irreversible process, following equation exists:

Close

Figure 6 (A) The voltammograms of HCPT with different scan rates (from inner to outer): 0.02, 0.04, 0.06, 0.08, 0.1, 0.15, 0.2 V s⁻¹. (B) The relationship between E_p1 and lnv.

Export to PPT

For P2 and P3, the peak shapes were roughly symmetrical at relatively slow scan rates. When the scan rates were increased from 20 to 200 mV s⁻¹, the peak potentials of P2 and P3 (E_p2, and E_p3) were shifted slightly and the peak-to-peak separation (ΔE_p) was augmented from 16 to 57 mV s⁻¹. The small ΔE_p separation values indicated a fast electron transfer rate. At the same time, the peak currents of i_p2 and i_p3 were proportional to the scan rates with linear regression equations of i_p2 = 7.228v + 65.90 (i_p2 in μA, v in V s⁻¹, R = 0.9906) and i_p3 = −5.553v − 54.216 (i_p3 in μA, v in V s⁻¹, R = 0.9921), indicating the electrode process driven by adsorption. A quasi-reversible adsorbed system should satisfy the following equation:

3.4 Chronocoulometry studies

For an adsorption controlled electrode process, it is necessary to calculate the saturated adsorptive capacity (Γ_max) of electroactive substance at the electrode surface. Then, multi-potential step chronoamperometry was employed and potentials steps were performed from 0.65 V to 1.3 V (first step), 1.3 V to −0.1 V (second step) and −0.1 V to 0.65 V (third step). Two experiments were carried out in 0.1 mol L⁻¹ PBS (pH 2.0) with absence and presence of 2.0 × 10⁻⁵ mol L⁻¹ HCPT, respectively. Fig. 7A shows the Q–t curves with HCPT presence (a, b and c) and absence (a′, b′ and c′). The markers a, b and c correspond with P1, P2 and P3 in Fig. 3. The corresponding Q–t^1/2 curves were obtained and displayed in Fig. 7B–D. At the first step (0.65–1.3 V), the corresponding Q–t^1/2 plots of curves a′ and a were calculated with the linear equations of Q(10⁻⁴ C) = 0.825t^1/2 + 0.6267 (R = 1.000) and Q(10⁻⁴ C) = 0.86t^1/2 + 1.047 (R = 1.000), respectively (Fig. 7B). A bigger intercept difference and same slope were obtained between a and a′, which further meant that the oxidation of HCPT at P1 was controlled by adsorption. According to the formula given by Anson (1964),

Close

Figure 7 (A) Chronocoulometric curves obtained at the ERGO/GCE in the presence (a, b, c) and absence (a′, b′, c′) of HCPT (2.0 × 10⁻⁵ mol L⁻¹). (B–D) The dependency of charge Q (10⁻⁴) versus t^1/2, corresponding data were derived from (A).

Export to PPT

For the next two steps from 1.3 to −0.1 V and from −0.1 to 0.65 V, the corresponding Q–t^1/2 curves were plotted in Fig. 7C and D. Obviously, the two slope values of the Q–t^1/2 (b, b′) plots were equal, which was an additional evidence for an total adsorption-driven electrode process. So same as Fig. 7D (c, c′), which was an total adsorption-driven electrode process. According to the formula mentioned above, saturated adsorption capacity values of 3.474 × 10⁻⁹ mol cm⁻² and 6.656 × 10⁻⁹ mol cm⁻² were calculated for the oxidative and reductive HCPT, respectively.

3.5 Calibration curve and determination limit of HCPT at ERGO/GCE

Here, calibration curve was achieved by linear sweep voltammetry and the potential was going from 0.65 V to 1.3 V with a scan rate of 0.05 V s⁻¹. In this way, only the P1 was appeared and it was used as the detect signal. The LSV curves were shown in Fig 8A. It was found that the anodic peak currents were two linear responses with HCPT concentrations. One is from 1.5 × 10⁻⁸ mol L⁻¹ to 5.0 × 10⁻⁷ mol L⁻¹ with the regression equation of i_p (10⁻⁶ A) = 1.187 + 9.669C(10⁻⁶ mol L⁻¹) (R = 0.9976) and the other is from 5.0 × 10⁻⁷ mol L⁻¹ to 3.0 × 10⁻⁶ mol L⁻¹ with the regression equation of i_p (10⁻⁶ A) = 4.92 + 2.293C(10⁻⁶ mol L⁻¹) (R = 0.9992), respectively. It was obvious that the slope of the latter was smaller, which could be ascribed to a decrease of the accumulation efficiency. Based on the first linear range, a determination limit of 1.5 × 10⁻⁸ mol L⁻¹ was obtained.

3.6 Reproducibility, stability, interference

The reproducibility and stability of the proposed sensor were evaluated using LSV in a 1.0 × 10⁻⁶ mol L⁻¹ HCPT solution. Five ERGO/GCEs were fabricated in parallel and then their responses were detected with a detected relative standard deviation (RSD) of 5.3%. Meanwhile, 10 interval measurements were carried out using one ERGO/GCE under optimal conditions. The obtained RSD was about 4.2%, which demonstrated good reproducibility of the present sensor. Five measurements were carried out on a new electrode surface using one electrode and the obtained RSD was about 3.9%. This suggested that the proposed sensor had a good repeatability. To estimate the stability of the sensor, one ERGO/GCE was stored for one week, and it maintained about 93.2% of its initial response for a same HCPT solution. It demonstrated that this sensor had good long-term storage stability.

The interference of some normal anions and cations and some organic compounds was investigated in the presence of 1.0 × 10⁻⁶ mol L⁻¹ HCPT. The results suggested that 100-fold concentration of Cu²⁺, Zn²⁺, Fe²⁺, Mg²⁺, K⁺ had no influence on the signals of HCPT with deviations below 5%. 50-fold glucose, starch, uric acid, ascorbic acid also showed no influence. In short, the interference effects of the studied compounds were negligible, which clearly proved the reasonable selectivity for the proposed method.

3.7 Standard addition recovery experiment of HCPT in the urine sample

In order to evaluate the applicability of the present method for the determination of HCPT in real sample, fresh human urine samples were obtained from volunteer who had not taken HCPT and collected in sterile bottles to avoid contaminating stored. The samples were centrifuged for 10 min (3000 rpm) to remove some proteins and then part of supernatant was diluted 100 times with 0.1 M PBS (pH 2.0) as a determining solution. The LSV was used for the standard addition recovery experiment (Fig. 9). Sweeping in this analytical solution, no anodic peak of HCPT was emerging. That is, no HCPT contains in health human urine sample and no interference exists also. By adding a defined concentration of HCPT to the detect sample solution, recoveries were determined in the range of 94.2–104% (Table 1), which clearly indicated the applicability and reliability of the proposed method.

Close

Figure 8 The LSV curves of HCPT with different concentrations (from a to g): 1.5 × 10⁻⁸, 8 × 10⁻⁸, 2.0 × 10⁻⁷, 4.0 × 10⁻⁷, 1.0 × 10⁻⁶, 2.0 × 10⁻⁶, 3.0 × 10⁻⁶ mol L⁻¹. (B) Calibration plot of peak current versus HCPT concentrations. Scan rate: 0.05 V s⁻¹.

Export to PPT

Close

Figure 9 LSV curves obtained at the ERGO/GCE in analytical urine sample with adding different HCPT concentrations (from inner to outer): 0, 3.0 × 10⁻⁷, 1.0 × 10⁻⁶, 2.0 × 10⁻⁶ mol L⁻¹. Scan rate: 0.05 V s⁻¹.

Export to PPT