2.1 Characterizations of NAHGO
X-ray photoelectron spectrum of the hybrid NAHGO is done to assess more about the surface composition and chemical states of the elements present in NAHGO. The wide scan spectrum (survey) of GO and NAHGO are displayed in Fig. 1a. GO shows only the peaks of C1s and O1s, where as NAHGO exhibits the presence of carbon, nitrogen and oxygen, with major prominent peaks at 285.12, 401.23 and 533.41 eV corresponds to C1s, N1s and O1s respectively. The C1s high-resolution spectrum is de-convoluted into four components, and is shown in Fig. 1b. The peaks at 284.5, 285.4, 286.1 and 287.6 eV attributed to C-C/C=C of graphene oxide framework, C-N, C-O and HN-C=O bonds. The deconvoluted peaks of O1s at 531.5 and 532.9 eV, as indicated in Fig. 1c, and are due to the characteristic C=O, C-O of hydroxyl and epoxy bonds of graphene oxide28,29. The individual peaks of N1s at 399.6 and 400.8 eV, which are exhibited in Fig. 1d, attribute to the presence of pyridinic N and pyrrolic N, respectively. The XRD patterns of GO and NAHGO is presented in Fig. 1e. The intense 2θ peak at 10.7o (001) diffraction plane of as prepared GO shows an inter-planar distance 0.83 nm, calculated by the Bragg’s equation. In NAHGO, 10.7o of GO is disappeared and shifted to high 2θ value, shows the reduction of oxygen containing functional groups and the incorporation of nitrogen atoms onto graphene surface. The Raman spectra of GO and NAHGO display two major bands as shown in Fig. 1f. The C=C stretching vibration of sp2 hybridised carbon systems illustrate the G band at about 1592 cm-1 and D band at 1352 cm-1 corresponds to the distortions in the edge-centred band structure of carbon framework of GO29. In NAHGO, the positions of these two peaks shift to 1581 cm-1 and 1341 cm-1, confirming the reduction in hybrid GO30. The intensity ratio ID/IG of GO and NAHGO are 0.96 and 1.03 respectively. The increased intensity ratio of ID /IG in NAHGO, which attributed to defects present after the functionalization of GO31.
The IR spectra of GO and NAHGO are shown in Fig. 2a. The broad peak at 3375 cm-1 is attributed to the stretching vibration of carboxyl -OH for GO. The C=O, C-OH, and C-O of carboxylic and epoxy groups represent peaks at 1721 cm-1, 1215 cm-1, and 1030 cm-1, respectively32,33. Three new peaks were observed in the spectra of NAHGO. They are at 1625 cm-1, correspond to the amide carbonyl stretch; 1561 cm-1, is due to the N-H of amide band; 1445 cm-1, the C-N stretching of amide bond34,35. This is in accordance with XPS, support the nitrogen doping of GO by nicotinic acid hydrazide. Thermal stability of the GO and NAHGO were investigated by TGA, and are shown in Fig. 2b. The slow degradation of NAHGO with GO indicates the high stability which acquire by the reduction of GO and the presence of nitrogen atoms in NAHGO. The XPS data of GO and NAHGO also supports the same. The GO and NAHGO were examined by SEM to understand more about the morphology and microstructure of these compounds. As seen in Fig. 2c-d, GO demonstrated the extended sheet-like structures and NAHGO results the deformation of sheet GO owing to the functionalization of NAH on GO. In NAHGO, the small tube like structures are distributed among the layered sheets of GO. The incorporation of nitrogen moieties on GO delaminate the sheet structure leads to more voids with crumpled structure. This justifies the significant disorder in the graphitic structure on NAHGO.
2.2 Electrochemical activity of NAHGO
The surface area of GO and NAHGO modified working electrodes were probed using 0.001Molar K4Fe(CN)6 in 0.1 Molar KCl at scan rate of 50 mV s-1 in potential window of -0.2 to 0.8 V, and are shown in Fig. 3a. The poor electrochemical response of K4Fe(CN)6 in 0.1 M KCl at bare GCE implies slow transfer of electrons at the surface. The slight increase of redox peak currents at GCE modified NAHGO than GO possibly owing to fast electron transfer and the large electroactive surface area at NAHGO. This is effective to offer more active sites for caffeine oxidation. The scan rates of NAHGO in a mixture of 0.001Molar K4Fe(CN)6 and 0.1 Molar KCl and plots of anodic current (Ipa) and cathodic current (Ipc) vs square root of the scan rate (υ1/2(mVs-1)1/2 (Fig S1 and S2a and b) were observed. The electrochemically active surface areas of all the GCE modified electrodes were calculated with the individual slopes of the plot using Randles Sevick equation
Ip = (2.69´105) n3/2D1/2υ1/2AC* (1)
which relates the peak current of the active species with surface area for a reversible process36. Ip refers to the anodic peak current, n is the total number of electrons transferred (n = 1), A is the effective surface area of the electrode, D is the diffusion coefficient of K4Fe(CN)6, C* is the concentration of K4Fe(CN)6 and υ is the scan rate. The calculated electroactive surface areas for bare GCE, GO, and NAHGO modified electrodes are 0.078 cm2, 0.126 cm2 and 0.29 cm2, respectively. NAHGO seemed to provide an enhanced active surface area presumably owing to the defects in GO with electron rich nitrogen moieties37.
2.3 Electrocatalytic characteristics of NAHGO for oxidation of caffeine
The electrochemical activities of bare GCE, GO, and NAHGO were further investigated in the presence of caffeine. Fig. 3b shows CVs of bare GCE, GO, and NAHGO modified GCE electrodes, at pH 2 with 100 μM of caffeine, at a scan rate of 50 mV s-1. The electrocatalytic behaviour of the NAHGO was also examined in the absence and presence of caffeine and is shown in Fig. 3c The occurrence of the caffeine oxidation peak in 1.5 V indicates the oxidation of caffeine on NAHGO modified GCE which is in accordance with the previous reported results19,38. Caffeine has tertiary amine group with three alkyl groups and an amide group in which a carboxyl group is connected to a nitrogen atom. In addition to these groups, an imine, carboxyl, methyl, and alkene groups are also present in caffeine. The carbon-oxygen and carbon-nitrogen covalent bonds make caffeine more polar due to its molecular geometry. The distinct positive and negative parts that are created by the overall molecular dipole of caffeine have strong affinity to the ends of other oppositely charged ends of the NAHGO by intermolecular attractions. Though the three nitrogen atoms of caffeine are methylated (N1, N3, and N7) as seen in Fig. S3, caffeine serves as a hydrogen bond acceptor with the ring nitrogen atom (N9) and carbonyl oxygen atoms (O2 and O6)39. Therefore the interaction of hydrogen is highly dominated on these positions of caffeine. The effective probes of interactions in NAHGO are epoxide, alcohol, carboxylic acid, pyrrolic, and pyridinic nitrogen moieties40. The alcohol and carboxylic acid oxygen containing functional groups and the NH of pyrrolic nitrogen can act as both hydrogen bond donor and acceptor. In this case, they play the role of hydrogen donor and make electrostatic interactions with caffeine. Hence, the increased electrochemical response of NAHGO arise from the efficient interactions of highly polar functional groups of caffeine and NAHGO, which generate charge accumulations on highly polar functional groups turned out to create electrostatic interactions. These interactions are less in GO due to the less number of heteroatoms which lead to not as much current response of NAHGO.
2.4 Effect of supporting electrolytes
To understand the effect of supporting electrolyte on the cyclic voltammetric behaviour of caffeine, different electrolyte solutions of pH were used. The effect of pH on the peak currents and peak potentials of caffeine electro-oxidation (100 μM) by NAHGO electrode was studied, and are represented in Fig. 3d. As the pH changes from 3 to 8 the oxidation peak current weakens, and strongest current was obtained at pH 2. In pH 2, the electrostatic interactions are more likely to occur with caffeine as mentioned earlier. So, it was considered as the sensible pH for further experiments. It is also clear that the peak potential shifted negatively as the increase of pH value of the supporting electrolyte. The linear dependence of anodic current and anodic peak potential with pH for caffeine are shown in Fig. 3e and f respectively. The corresponding regression equations are: Ipa = 0.214pH + 0.307 (R=0.9963) and Epa=2.56-0.0491pH (R=0.9955). This implies that the electrochemical oxidation occurs with equal number of proton and electron transfer process41. In the presence of acidic pH, the exchange of electrons between the π bonds of graphene sheet and the lone pair of electrons bring the polarization effect at the electrode surface. The interactions of NAHGO with acidic protons are shown in Fig. 4.
Cyclic voltammograms of NAHGO (c) with and without caffeine, (d) at pH values of: 2, 3, 5, 6, 7, 8, in 100 μM of caffeine, at a scan rate of 50 mVs-1, Calibration plot of (e) pH vs anodic current (Ipa), and (f) pH vs anodic peak potential (Epa).
2.5 Effect of scan rate
In order to obtain more knowledge about the adsorption properties of NAHGO the CVs of 100 μM caffeine on NAHGO at various scan rates (Fig. S4) were studied for the electro-oxidation process of caffeine. As displayed in Fig. 5a the anodic peak current of caffeine increases linearly with the increase of scan rate from 10 to 60 mV s-1. The linear relationship of the oxidation current and the square root of the scan rates present that the electrochemical reactions occur on the proposed electrode is diffusion controlled. The oxidation peak potential (Epa) shifts to more positive potential as the gradual increase of scan rates. Laviron model was used to describe the kinetic parameters of the electrochemical oxidation of caffeine42. The linear relationship of anodic oxidation potential with logarithms of scan rate is illustrated by this method. According to Laviron, the correlation between the scan rate and the oxidation peak potential is explained as follows:
Epa=E0+2.3RT/[(1-∞)nF]$logυ (2)
As shown in Fig. 5b, Epa exhibits the linear dependence with logarithms of scan rate (log υ) and the equation for linearity is Epa=1.4326+0.0507logυ with R2=0.9965. The electrochemical transmission coefficient of an irreversible process is about 0.4-0.6 and for those processes the value is considered to be 0.543. Accordingly, the electronic transmission number (n) of electrochemical oxidation of caffeine is calculated as 3.85, that is in good agreement with the previous results44. The mechanism of electrochemical oxidation of caffeine is shown in Fig. 6 as reported earlier45. Substituted uric acid is formed in the first step, by the oxidation of two electrons and two protons at C-8-N-9 position, followed by formation of 4,5 diol of uric acid analogue through the immediate two electron and two proton oxidation.
2.6 Analytical performance
In the present work, the differential pulse voltammetric method was found to be a more sensitive method, as it produced better electrochemical performance indicators, than the CV technique. Therefore, in order to achieve the favourable analytical performance using differential pulse voltammetric (DPV) method, the parameters such as, pulse amplitude (from 0.05 V to 0.15 V), pulse width (from 0.005 to 0.5 s) and pulse period (0.1 to 3 s) is optimised. The best oxidation peak of caffeine was recorded as follows: pulse amplitude 0.05 V, pulse width 0.06 s and pulse period of 0.5 s with NAHGO modified electrode in 0.1 M H2SO4 at pH 2 in caffeine. Different concentrations from 0.01 μM to 40 μM of caffeine solutions were prepared to examine the relationship between the peak current and the concentration of caffeine. Fig. 5c displays the linear increase of the anodic peak current (Ipa) with concentrations of caffeine. The calibration graph (Fig. 5d) for caffeine at NAHGO obeys the linear regression equation: Ipa(mA)=6.883+0.0989C; R2=0.9906. The limit of detection (LOD) of caffeine is calculated to estimate this method for the determination of caffeine using the formula 3.3s/S, where s represents residual standard deviation and S is the slope of the calibration curve38 and the LOD for NAHGO modified electrode is found to be 8.7 nM, the best reported so far. Moreover, it is compared with other reported values for the determination of caffeine using electrochemical techniques. They are presented in Table 1 and which shows the effectiveness of the proposed NAHGO sensor.
Table 1 An overview of electrochemical methods for the determination of caffeine
Modified electrodes
|
Technique
|
Linear range (μM)
|
Detection limit
(μM)
|
References
|
BQ/CPE
|
SWV
|
500-8000
|
51.00
|
46
|
Bi-CNT/CPE
|
SWV
|
51.03-1026
|
0.182
|
47
|
DNA/CNT/CPE
|
SWV
|
0.5120-61.70
|
0.350
|
48
|
MIP/CPE
|
DPV
|
0.0600-25.00
|
0.015
|
49
|
CTAB/GR/GCE
|
DPV
|
0.300-100.0
|
0.091
|
50
|
SWCNT/CPE
|
DPV
|
0.250-100.0
|
0.120
|
51
|
PT/TiO2-GR/GCE
|
DPV
|
25.00-200.0
|
0.500
|
52
|
Nitrogen doped graphene
|
SWV
|
0.060-50.00
|
0.020
|
53
|
Poly(Alizarin Red S)/GCE
|
SWV
|
0.500-250.0
|
0.060
|
54
|
Attapulgite/Nafion/GCE
|
DPV
|
1.000-4.000
|
0.040
|
16
|
NAHGO/GCE
|
DPV
|
0.010-40.00
|
0.008
|
This work
|
2.7 Determination of caffeine in energy drink and pharmaceutical sample
Finally, in order to estimate the practical application of this method, NAHGO/GCE was used to determine caffeine in imol plus and O’cean one8 energy drink. Energy drink was degasified by ultrasonication for 10 minutes before the analysis and diluted with supporting electrolyte (100 μM). The tablet was grounded with mortar to a fine powder, dissolved in supporting electrolyte (100 μM) and sonicated for 10 minutes. Subsequently, the appropriate amount of the resulting dispersion was selected, and then was centrifuged. 9 cm3 of 0.1 mol L-1 pH 2 solution was added to 1 cm3 of the sample solution and recorded the DPVs (Fig.5e and f). In the mean time to calculate the recovery standard solutions of caffeine is added and it is represented in Table S1. The recovery range of 94.9-101.9 % suggests that NAHGO is adequate for practical sensing applications.