Graphene nanoplatelet-based sensor for the detection of dopamine and N-acetyl-p-aminophenol in urine

2.1 Apparatus and reagents

Electrochemical experiments were conducted using PGSTAT204 Autolab Potentiostat/Galvanostat/EIS FRA32M Module (Metrohm-Autolab, The Netherlands) with Nova 2.1 Software for data acquisition and experimental control. Electrochemical impedance spectroscopy in 5.0 mM potassium hexacyanoferrate ([Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻) was carried out at open circuit within the frequency range of 100 kHz–0.1 Hz at an applied potential of 0.25 V. Under the optimized conditions, unless otherwise stated, Differential pulse voltammetry (DPV) were recorded with pulse amplitude, pulse time, voltage step time, voltage step and scan rate of 50 mV, 50 ms, 500 ms, 5 mV, and 50 mV·s^-1 respectively. The disposable screen-printed carbon electrodes (Ref DS 410) utilized in the sensor design have a carbon working electrode, carbon counter electrode and silver reference electrode and were purchased from DropSens, Spain. Scanning electron microscopy (SEM) was performed by JEOL JSM-610PLUS/∼LA SEM (JEOL Ltd, Japan). Raman spectrum was obtained with a LabRAM 300 system (HORIBA Scientific, UK) using He-Ne (632.8 nm) laser. Graphene nanoplatelets were purchased from the US Research Nanomaterials, Inc, Houston, TX. Nafion, potassium hexacyanoferrate ([Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻), hydrogen peroxide (H₂O₂) (30% (w/w) in H₂O), Phosphate Buffered Saline (PBS) tablets, dopamine hydrochloride and N-acetyl-p-aminophenol were purchased from Sigma-Aldrich, USA. All other chemicals were of analytical grade and used without further purification. Pooled normal human urine samples were purchased from Innovative Research, USA.

2.2 Procedures

3.1 Characterization of SPE and GNPs-Naf/SPE

The surface morphological features of the SPE and GNPs-Naf/SPE were examined by SEM and Raman spectroscopy. Fig. 1A, B, C show the view of the SPE and GNPs-Naf/SPE, respectively. The morphology of the SPE is typical for graphite materials with grains that are stacked in flakes (Fig. 1A). The heterogenous distribution of graphene nanoplatelets sheets of submicron dimension is clearly visible throughout the GNPs/Naf/SPE with wrinkles on the surface (Fig. 1B) and well-formed architecture of large open area porous surface structure (Fig. 1C). Further characterization of the GNPs/Naf/SPE (Fig. 1D) show D, G and 2D peaks, which are typical features of thick graphene stacks (Ferrari, 2007; Cancado et al., 2008).

Close

Fig. 1

Scanning electron micrographs of (A) bare SPE; (B) top and (C) cross-section of GNPs-Naf/SPE, and (D) Raman spectrum of GNPs-Naf/SPE measured using 632.8 nm laser.

Export to PPT

The interface of the SPE and GNPs-Naf/SPE were further characterized by Faradaic electrochemical impedance spectroscopy (EIS) in the presence of 5.0 mM Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻. Fig. 2 illustrates experimental Nyquist spectra for both the bare SPE and GNPs-Naf/SPE. The impedance spectrum associated with the bare SPE (curve a, Fig. 2) consists of a semicircle in the high frequency region and a linear part in the low frequency region, corresponding transfer and diffusion processes, respectively. The diameter of the semicircle represents the charge-transfer resistance (R_ct) at the electrode (Suni, 2008; Kanyong et al., 2016b). In comparison to what was observed at the GNPs-Naf/SPE (curve b, Fig. 2), the diameter of the semicircle of the bare SPE was larger. This shows that the charge-transfer rate increased upon employing the GNPs, which facilitated the Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ redox process.

Close

Fig. 2

Nyquist plots observed for electrochemical impedance spectroscopy at bare SPE (curve a) and GNPs-Naf/SPE (curve b) in PBS (pH 7.4) containing 5.0 mM [Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ and 0.1 M KCl.

Export to PPT

The electrocatalytic behavior of the sensors was characterized by cyclic voltammetry. Fig. 3A is comparison of cyclic voltammograms recorded at the bare SPE and GNPs-Naf/SPE in PBS (pH 7.4) containing 5.0 mM Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ and 0.1 M KCl at 100 mV s⁻¹ scan rate. When compared to what occurred on the bare SPE (curve a, insert of Fig. 3A), the GNPs-Naf/SPE (curve b, Fig. 3A) exhibited a characteristic increase of both the anodic and cathodic peak currents for the Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ redox couple, thus, confirming the successful modification of the SPE with the GNPs composite. Higher peak currents and smaller peak-to-peak potential separation (ΔE_p) were observed at the GNPs-Naf/SPE (I_pa = 1.35 mA, I_pc = 1.38 mA; ΔE_p = 200 mV) when compared with the bare SPE (I_pa = 0.03 mA, I_pc = 0.07 mA; ΔE_p = 276 mV). This more than 40-fold increase in the anodic peak current and 20-fold increase in the cathodic peak for Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ redox couple can be attributed to the higher electrocatalytic properties of the GNPs composite which led to an enhancement of the total electroactive area of the GNPs-Naf/SPE (Krampa et al., 2017). The presence of the GNPs also produced a negative and a positive shift in the anodic and cathodic potentials, respectively; thus, giving rise to a smaller peak-to-peak separation (ΔE_p = 200 mV). This cyclic voltammetry data agrees with the results obtained from faradaic impedance analysis; thereby confirming the successful modification of the SPE with graphene nanoplatelets.

Close

Fig. 3

(A) Cyclic voltammograms recorded using bare SPE (curve a) and GNPs-Naf/SPE sensor (curve b) at 100 mV·s⁻¹ scan rate; insert is CV of bare SPE; (B) CVs recorded using GNPs-Naf/SPE at 10, 20, 35, 50, 75, 100, 150, 175, 200, 250, and 300 mV·s⁻¹ scan rates; insert is peak current vs. square root of scan rate. All CVs were recorded in 5.0 mM [Fe(CN)₆]³⁻/[Fe(CN)⁶]⁴⁻ in PBS (pH 7.4) containing 0.1 M KCl.

Export to PPT

The effect of scan rate (v) on the voltammetric behavior of GNPs-Naf/SPE towards Fe(CN)₆]³⁻/[Fe(CN)₆]⁴⁻ redox couple was then examined. At the scan rates investigated (Fig. 3B), a plot of the square root of the scan rate (v) vs. the anodic (I_pa) and cathodic (I_pc) peak currents exhibited a linear relationship (insert of Fig. 3B), which suggests a behavior consistent with surface confined voltammetry and corresponding ‘thin-layer’ type voltammetry (Kanyong et al., 2016b). The slight deviations from linearity may reflect on the transport processes to and from the electrode resulting from the porosity of the GNPs composite films (Streeter et al., 2008; Kanyong et al., 2016b). A linear relationship was also observed when the absolute values of log I_pa and log I_pc were plotted against log v (not shown) with slope values of 0.50 and 0.69, respectively. These slope values are comparable with the theoretically expected value of 0.5 for purely diffusion-controlled currents (Gosser, 1993), which suggests that the electrochemical process is diffusion-controlled and that the surface of the modified SPE is not fouled.

3.2 Electrochemical behavior of DA and APAP at SPE and GNPs-Naf/SPE

Cyclic voltammograms (CVs) were recorded for 0.5 mM each of DA and APAP at the bare SPE and GNPs-Naf/SPE in PBS (pH 7.4) at a scan rate of 100 mV s⁻¹. A comparison of the CVs for DA and APAP at the SPE and GNPs-Naf/SPE is illustrated in Fig. 4A. A pair of well-defined, quasi-reversible anodic and cathodic peaks for DA was observed at 160.4 mV and 102.1 mV, respectively, on the GNPs-Naf/SPE. The anodic peak can be attributed to the oxidation of DA to dopaminequinone while the cathodic peak can be attributed to the reduction of the dopaminequinone back to DA (Loget et al., 2013; Kanyong et al., 2016b). At the bare SPE, DA has an anodic peak and a broadened cathodic peak; which is an indication of a sluggish electron transfer process. This behavior of DA at the bare SPE can be attributed to the conductivity of the SPE material and/or electrode fouling caused by the deposition of DA and its associated redox products on the electrode (Kanyong et al., 2016b). The redox peaks shifted to more positive potentials with a marked decrease in the peak currents (I_pa = 2.5 mA; I_pc = 0.4 mA) at the bare SPE. A significantly enhanced peak currents (I_pa = 5.7 mA; I_pc = 1.5 mA) with peak-to-peak potential separation (ΔE_p) of 58.3 mV for DA was observed at the GNPs-Naf/SPE; which is considerably close to the 59.0 mV value expected for Nernstian one-electron reactions (Kanyong et al., 2016b). Similar electrochemical behavior of dopamine has been reported elsewhere (Morrin et al., 2003; Kanyong et al., 2016b, 2016c; 2016d). The smaller ΔE_p value suggests that the reversibility of DA at GNPs-Naf/SPE is remarkably improved and the enhanced peak currents also suggests an enhancement of the electron-transfer rate by the GNPs at the surface of the electrode.

Close

Fig. 4

CV of (A) 0.5 mM dopamine (DA); (B) 0.5 mM N-acetyl-p-aminophenol (APAP); and; (C) simultaneous determination of binary mixtures of DA and APAP (0.5 mM each) in PBS (pH 7.4) at bare SPE (curves a) and GNPs-Naf/SPE (curves b).

Export to PPT

Fig. 4B shows CV recorded in 0.5 mM APAP in PBS (pH 7.4) at 100 mV s⁻¹. The voltammogram shows an anodic peak (at 328.0 mV) in the positive-going scan and a cathodic counterpart peak (at 298.5 mV) in the negative-going scan which corresponds to the transformation of N-acetyl-p-aminophenol (APAP) to N-acetyl-p-benzoquinoneimine and vice versa in a quasi-reversible two-electron (ΔE_p = 29.5 mV) process (Eq. (1)) (Nematollahi and Niroomand, 2009):

However, at the SPE, there was a positive shift in the anodic peak (E_pa = 442.2 mV) position and a negative shift in the cathodic peak (E_pc = 70.0 mV) position; which increased the peak-to-peak potential separation from 29.5 mV to 372.2 mV. This is an indication of a sluggish electron transfer process. The electrochemical behavior of both DA and APAP on both sensors had non-unity peak current ratios (I_pc/I_pa); this can be a criterion for the instability of reaction products (Nematollahi and Niroomand, 2009) at the sensor surfaces under the experimental conditions.

The CVs for a binary mixture of 0.5 mM DA and APAP in PBS (pH 7.4) at the SPE and GNPs-Naf/SPE were also recorded (Fig. 4C). The voltammogram for the binary mixture shows two broadened anodic peaks (a1 = 232.4 mV; a2 = 456.2 mV) with no corresponding distinguishable reduction peaks at the bare SPE. However, at the GNPs-Naf/SPE, the voltammetric profiles have well-defined peaks at 166.7 mV and 374.2 mV corresponding to the oxidation of DA and APAP, respectively. The oxidation peaks of DA (a1 = 166.7 mV) and APAP (a2 = 374.2 mV) observed on GNPs-Naf/SPE formed a quasi-reversible couple with peak potentials (c2 = 135.7 mV; C1 = 135.7 mV), as observed during the oxidation of DA (Fig. 4A) and APAP (Fig. 4B) alone. The peak currents are also significantly higher at GNPs-Naf/SPE than at SPE. Clearly, the modification of the SPE with GNPs leads to significant increase in peak current. The presence of Nafion served as a binding agent to prevent the detachment of the GNPs from the SPE surface (Wen et al., 2012; Yao et al., 2013a; Yao et al., 2013b; Yao et al., 2014). Without the presence of the Nafion, the GNPs are easily washed off the surface of the SPE when immersed in the electrochemical cell with measuring solution. Consequently, the addition of Nafion enhanced the GNPs film formation on the SPE surface.

3.3 DPV analysis of DA and APAP at GNPs-Naf/SPE

Further analysis of DA and APAP was carried out at the GNPs-Naf/SPE using differential pulse voltammetry (DPV) in PBS (pH 7.4) over concentration range of 0.25–40.0 µM and 0.25–30.0 µM for DA and APAP, respectively. A fresh electrode was used for each measurement, and this was done in triplicate for each concentration. The resulting plots (Fig. 5A, C) show that the oxidation peak currents (I_pa) for DA and APAP increased linearly with increasing concentrations, suggesting a stable and efficient electrocatalytic activity at the GNPs-Naf/SPE. For DA detection, the corresponding linear regression equation is defined by the expression $Ipa\left(\mu A\right)=0.321\left[DA\right]\left(\mu M\right)+0.014,R^2=0.9994$ in the range of 0.25–40.0 µM (Fig. 5B) while for APAP detection, the corresponding linear regression equation is defined by the expression $Ipa\left(\mu A\right)=0.398\left[APAP\right]\left(\mu M\right)+0.033,R^2=0.9994$ in the range of and 0.02–30.0 µM (Fig. 5D). Using the regression equations, the calculated limits of detection for DA and APAP (based on 3x the baseline noise) were found to be 0.13 µM and 0.25 µM, respectively; these were deemed to be satisfactory for the analysis of elevated levels of both compounds in human bodily fluids.

Close

Fig. 5

Analysis of DA and APAP at GNPs-Naf/SPE. Differential Pulse Voltammograms (DPV) with corresponding linear calibration plots for (A and B) DA (C and D) APAP, and (E and F) in PBS (pH 7.4).

Export to PPT

3.4 Analysis of DA and APAP in binary mixtures at GNPs-Naf/SPE

DPV analysis of binary mixtures of DA and APAP was also performed with different concentrations (DA in the range of 1–30 µM; APAP in the range of 0.1–10.0 µM) and the voltammogram for this analysis is shown in Fig. 5E. With increasing concentrations of each compound, there was a linear increase in the anodic peak currents (Fig. 5F) for each measurement. The corresponding linear regression behavior for DA and APAP is defined by the equations $Ipa\left(\mu A\right)=0.221\left[DA\right]\left(\mu M\right)+0.011,R^2=0.9992$ and $Ipa\left(\mu A\right)=0.950\left[APAP\right]\left(\mu M\right)+0.101,R^2=0.9999,$ respectively. The limits of detection for DA and APAP were estimated to be 0.15 µM and 0.32 µM, respectively. These analytical performance characteristics are superior to published results obtained from other graphene-based sensors (Table 1), which tended to have reduced linear ranges (Adhikari et al., 2016; Kang et al., 2010; Xiong and Jin, 2011; Zhang et al., 2014;) and/or poor limits of detection (Kim et al., 2010; Kim et al., 2017; Mallesha et al., 2011; Xiong and Jin, 2011; Wang et al., 2009; Yang et al., 2014) when compared to this sensor.

The reproducibility of the GNPs-Naf/SPE is highlighted in Fig. 5B, D, F where the responses to individual increasing concentrations of DAP (Fig. 5B) and APAP (Fig. 5D) and in their binary mixtures (Fig. 5F) are detailed for triplicates measurements. Standard deviations for peak current intensities for the two compounds were found to be in the range of 1.2–1.9%. The stability of the GNPs-Naf/SPE was also determined in a similar fashion after storage in room temperature for 30 days; the peak current intensities for DA and APAP decayed by less than 3.1%, suggesting that the electrodes were highly stable over the 30-day period.

3.5 Analytical application of GNPs-Naf/SPE

To demonstrate the feasibility of the GNPs-Naf/SPE for routine analysis, the sensor was used to determine DA and APAP concentrations in human urine samples. All the urine samples were diluted (1:10) with PBS (pH 7.4), and then appropriate amounts were transferred to the electrochemical cell and each species determined using the GNPs-Naf/SPE and DPV. The simultaneous determination of DA and APAP was also evaluated; firstly, their binary mixture in PBS (pH 7.4) was spiked into the urine samples and DPV was used to simultaneously recover them. A fresh GNPs-Naf/SPE was used for each measurement. The recovery data (Table 2) shows that the methods is highly accurate and reproducible, indicating potential application of the GNPs-Naf/SPE for the determination of DA and APAP in real clinical samples.