Introduction

Tartrazine (Tz) is a synthetic colorant commonly used in soft drinks due to its low cost and great stability under visible light, and in acidic media [1]. However, this synthetic dye must be strictly controlled by government laws due to potential risks to human health given the presence of azo groups (R1-N = N-R2, where R1 and R1 can be either aryl or alkyl functional groups) susceptible to reductive cleavage leading to primary aromatic amines [2, 3]. Then, the maximum permitted level for Tz in food is 100 ppm [2, 4].

Several methods have been developed for the determination of Tz in water, such as absorption[5], photocatalytic degradation [6], HPLC [7], electroluminescence [8], etc. These methods are expensive and present a complex sample handling. Therefore, it is still a challenge to develop a sensitive and simple method for Tz determination.

In recent years, electrochemical sensors have been raising attention for dye detection due to their low cost, suitability to detect different molecules, robustness, high sensitivity, selectivity, and reproducibility [9], where the drop in the price of electrodes manufacturing can be associated to the development of new technologies involving the use of carbon nanomaterials [10] and metallic nanoparticles [11, 12]. Currently, metal nanoparticles are used to increase sensitivity in electrochemical systems taking advantage of their high surface/volume ratio, which increases the electrochemically active surface area [13]. As a support for the metallic nanoparticles, several carbon-based materials have been used. Particularly, for the detection of azo dyes, multi-walled [10, 14] and single walled carbon nanotubes [15], graphene oxide [16], reduced graphene oxide [17], graphitic carbon [18], glassy carbon [19], graphite [20], hierarchical porous carbon [21], and carbon quantum dots [22] have been employed. However, to the best of our knowledge, carbon nanofibers (CNF) have not yet been used as supporting material for the electrochemical detection of food dyes.

CNF are interesting one-dimensional carbon materials widely used in solar cells and supercapacitors [23, 24]. They exhibit a unique nanofiber network structure, good conductivity and chemical stability [25]. However, they display a small surface area and a small number of active sites, resulting in a lower detection sensitivity [26, 27]. In this regard, gold nanoparticles (AuNP) are used to improve the Tz electrochemical detection of CNFs. The use of these composite nanomaterial could help to obtain low detection limits, due to the high conductivity value of the AuNP (4 × 105 S cm−1) [28, 29], as well as the adsorbent capacity and the high surface area of CNF [30,31,32]. In addition, in recent years, AuNP have attracted increasing attention due to their electrocatalytic behavior leading to an improved sensitivity [33]. Moreover, AuNP-based sensors do not require the use of membranes, polymers or other materials for the formation and growth of the nanoparticle on the CNF electrode surface [34,35,36].

In this work a composite material based on AuNP supported on CNF is developed for the Tz electrochemical detection. This electrochemical sensor enables a rapid and high sensitivity detection of Tz in commercial drinks without a sample pretreatment, making it an appealing device for quality monitoring.

Experimental

Reagents and Solutions

All solutions were freshly prepared using deionized water (resistivity ≥ 18.2 MΩ cm at 25 °C) obtained from a Milli-Q purification system (Millipore). A standard solution containing 1000 mg L−1 of Au (HAuCl4) in HCl 2 mol L−1) and potassium chloride KCl were obtained from Merck. Tz, Sodium Borohydride (NaBH4), and iron salts 5.0 mmol L−1 K4[Fe(CN)6]/K3[Fe(CN)6] were obtained from Sigma-Aldrich. Sunset yellow, sodium benzoate, NaNO3, NaHCO3, glucose and citric acid were obtained from Sigma-Aldrich and Merck, respectively. Carbon nanofibers (graphitized (iron-free), composed of conical platelets, D × L 100 nm × 20 – 200 μm) were obtained from Sigma Aldrich.

Phosphate buffer solutions of 0.1 mol L−1 at pH 6, 7, 8 and 9 were prepared using the salts (NaH2PO4.H2O/Na2HPO4.7H2O, Merck). Stock solutions of 1.0 × 10–2 mol L−1 of Tz were prepared, dissolving 53.43 mg of the compound in 0.1 mol L−1 of phosphate buffer at pH 8.

CNF/AuNP nanocomposite synthesis

250 mg of CNF was weighed and mixed with 200 mL of 1 × 10–4 mol L−1 gold precursor standard solution at pH 9.5 (the pH was adjusted with NaOH). Then, AuNP were synthesized in situ using the procedure described in literature [37]. In the mixture, 5 mL of 0.01 mol L−1 NaBH4 were added at room temperature for the formation and growth of AuNP on the CNF surface. Finally, the solution was heated to remove unreacted borohydride ions, and the composite obtained was centrifuged 3 times at 10,000 rpm in NaOH solution at pH 9.5. It was dried using an oven at 60 °C for 24 h. The composite obtained was CNF/AuNP.

Preparation of CNF/AuNP paste electrode

To obtain the carbon paste electrode, 1 mL of 0.05 mol L−1 KCl was added and carefully mixed with 100 mg of CNF/AuNP or CNF in a mortar and dried at ~ 50 °C for 15 min. Afterward, 80 µL of mineral oil were added to obtain a paste. The prepared paste was then packed into the cavity of the electrode (3 mm diameter and 1.5 mm deep). The modified carbon paste electrodes (CPE) were named CNF-CPE and CNF/AuNP-CPE, respectively.

Physicochemical and Electrochemical measurements

The morphology of CNF and CNF/AuNP composite were examined using a TESCAN LYRA3 FIB-SEM. The conditions were: 25 kV, work distance of 8.5 mm and spot size of 60 µm. Energy-dispersive X-ray spectroscopy (EDS) was used to detect the AuNP in the composite at the same SEM conditions.

X-ray diffraction was used to characterize the crystalline phase and crystallite size of the samples. A PANalytical instrument Empyrean model was used to perform the measurements. The conditions were: 20° ≤ 2θ ≤ 70°, step size of 0.026° per each 250 s, slits of ½ (tube) and ¼ (detector) and detector model PIXcel 3D. To determine the crystalline sizes and lattice strain of AuNP, the Williamson-Hall (W–H) method has been applied [38].

Electrochemical measurements were performed by cyclic voltammetry (CV), differential pulse voltammetry (DPV) and electrochemical impedance spectroscopy (EIS) on a Potentiostat–Galvanostat PGSTAT204 Autolab (Eco–Chemie, the Netherlands) using Nova software. Measurements were made using a glass electrochemical cell with a three-electrode setup. A graphite bar was used as a counter electrode, and Ag/AgCl (3 mol L−1 KCl) as a reference electrode.

Evaluation of CNF/AuNP/CPE sensor and the electrochemical detection of Tz

The electrochemical characterization of CNF/AuNP-CPE was performed at different pH (between 6 and 9) in presence of Tz using 0.1 mol L−1 phosphate buffer solutions, by CV at a scan rate of 25 mV s−1 and DPV using the following parameters: 75 mV of amplitude, 25 mV s−1 of scan rate and 30 s of conditioning time. EIS analyzes were carried out in a solution of Fe2+/Fe3+ redox couple, in an electrolyte of 5.0 mmol L−1 and KCl 0.1 mol L−1, using a frequency range between 0.1 and 100 kHz and an amplitude of 10 mV. The Chronoamperometric and Chronocoulometric analyzes were performed by an applied potential of 0.91 V for 120 s in the absence and in the presence of Tz. CV and DPV techniques were used to evaluate the stability, reproducibility, repeatability, sensitivity, limits of detection (LOD), limits of quantification (LOQ), and selectivity of the sensor in the presence of various interferents such as: sunset yellow, sodium benzoate, NaNO3, NaHCO3, glucose and citric acid, during the electrochemical detection of Tz. Finally, the Tz concentration in peruvian soft drinks Inca Kola® and Cifrut® was determined by the optimized sensor, and it was validated by the UV–visible spectrophotometric technique.

Computational study of the gold-tartrazine interaction

A density functional theory (DFT) computational study has been conducted to determine the most stable array of the tartrazine molecules interacting with the gold nanoparticles at the electrode/electrolyte interface, and also to estimate the intensity of the interactions involved. The DFT calculations were carried out by using the Gaussian16 software [39], tightening self-consistent field convergence thresholds (10–10 a.u.), and considering a 6-311G(d)/LANL2DZ basis set and the M06-2X functional [40]. For the Au20 cluster we took as a starting point the tetrahedral symmetry structure proposed by Hensen and coworkers obtained from a DFT-based genetic algorithm [41]. This structure agrees with previous quantum-chemical calculations as well as experimental IR data used to resolve the structure of neutral gas-phase gold clusters [42, 43]. The interaction energy is obtained by a fragment calculation where the gold cluster and Tz energies are subtracted from the energy of the overall system. For this calculation the counterpoise correction for the basis set superposition error is considered [44]. The solvent effects of water were included according to the Polarizable Continuum Model (PCM) [45] where the solute is placed in a cavity of molecular shape, and the solvent dielectric response is projected on the surface of the cavity.

Results and discussions

Morphological and structural evaluation of CNF/AuNP sensor

The morphologies of CNF and CNF/AuNP were explored using SEM images at different scales. Figure 1a and b show the CNF morphology, which exhibits similar features when compared with carbon nanotubes. The diameter of CNF varies between 80-110 nm, approximately. The CNF/AuNP shows a morphology similar to pure CNF (Fig. 1c and d). Through SEM, small agglomerations of AuNP between 15–25 nm of sizes were identified (Fig. 1d). Furthermore, the AuNP clusters are not distributed evenly, enabling only a qualitative elemental identification by EDS (indicated by arrows in Fig. S1).

Fig. 1
figure 1

SEM images for the CNF at (a) and (b) and the CNF/AuNP at (c) and (d). The AuNP are identified qualitatively from EDS as indicated with arrows at (d) image

Full size image

Figure 2 shows the X-ray diffractogram of the CNF/AuNP composites. The CNF diffractogram (Fig. S2) shows a diffraction peak at 26.49° and another at 54.5°, which are consistent with the (002) and (004) plane reflections of graphite structure [46]. The crystalline structure of CNF/AuNP composite showed well-defined Bragg peaks corresponding to (111), (002) and (202) plane reflections of the Au material (face centered cubic (fcc) structure). Besides, the crystalline size of AuNP was approximately 7.5 nm, which showed a low crystalline order and low agglomeration when it is impregnated on the CNF surface (Fig. 1d) [47].

Fig. 2
figure 2

X-Ray Diffraction pattern of CNF/AuNP

Full size image

Electrochemical characterization of CNF/AuNP-CPE sensor

EIS and CV analysis were performed on CNF-CPE and CNF/AuNP-CPE (Fig. 3a and b), using the potassium hexacyanoferrate solution [Fe(CN)6]3/4− redox couple 5.0 mmol L−1 and KCl 0.1 mol L−1 as supporting electrolyte. EIS was evaluated to determine the electron transfer behavior of Tz. The electrochemical results are shown in Fig. 3a. The inset displays the equivalent circuit model with the following values: electrolytic support resistance (Rs), charge transfer resistance (Rct), Warburg impedance (W1 and W2), and a phase element constant (CPE). From the Randles model it is obtained Rs = 33.5 Ω and Rct = 127.6 Ω; Rs = 22.4 Ω and Rct = 79.8 Ω for CNF-CPE and CNF/AuNP-CPE, respectively (Table S1). Regarding the proposed equivalent circuit, the second resistance was identified as Rct. The decrease in Rct is due to the presence of AuNP on the CNF surface, indicating the enhancement of electron transfer between the CNF/AuNP-CPE sensor surface and the electrolyte solution [48]. This component is confirmed by the Tafel measurement displayed in Fig. S3a. For CNF and CNF-AuNPs, 185.9 and 99.8 Ω resistances were respectively obtained. The values fit the data from impedance analysis and the low error percentage values in the selection of components confirm our equivalent circuit. Figure 3b shows the oxidation/reduction of the Fe2+/Fe3+ redox couple signal by CV, being sharper for the CNF/AuNP-CPE sensor. This indicates an improvement in the electron transfer and also indicates that the oxidative process between the sensor surface and the electrolyte solution is favored [49, 50].

Fig. 3
figure 3

a) Nyquist plot (OCP: 0.25 V), b) CV at 25 mV s−1 of CNF-CPE and CNF/AuNP-CPE in 5.0 mmol L−1 [Fe(CN)6]3/4− and 0.1 mol L−1 KCl electrolyte

Full size image

To evaluate the transfer of electrons (k0) on the surface of the CNF-CPE and CNF/AuNP-CPE electrodes, the estimation of the constant k0 was made, since it is of great interest when examining the performance of electrode materials. The rate constant provides an indication of the rate of electron transfer between an electroactive species and an electrode surface, whether the electrode material determines the overall rate of the electrochemical reaction. The value of k is obtained from Eq. 1 and Fig. S3a, where i0 is the exchange current, n is the number of transferred electrons, F is the Faraday constant 96,485 C mol−1, A is the electrode electroactive area in cm2 and C the concentration of the species [51]. The value of k0 is 4.03 × 10–6 and 7.50 × 10–6 cm s−1 for CNF-CPE and CNF/AuNP-CPE electrodes, respectively. These values suggest that the CNF modified with the AuNP presents a better electron transfer. The results obtained here may have a relationship with the sensitivity during Tz dye detection.

$${i}_{0}=nFA{k}^{0}C$$
(1)

The pH value of the supporting electrolyte has a significant influence on the electrochemical behavior of Tz [16]. Therefore, it is important to evaluate the optimal pH value for Tz detection. The oxidation of Tz was evaluated at different pH values in the range of 6 to 9 by DPV (Fig. 4a). The results show a potential shift to negative values when the pH for CNF/AuNP-CPE increases, as expected [16, 17]. Furthermore, the oxidation peak potential (Eoxid) is linear in the pH range 6–9 (Supplementary Fig. S3). The linear equation is Eoxid = 1.09 − 0.030pH (r2 = 0.9594), and the value slope is − 30.02 ± 3.5 mV/pH [52]. These results are in good agreement to the Nernst equation (Eq. 2) of Epc = E − 0.059/n pH, where n is the number of transfer electrons [53, 54].

$${E}_{pc}=E-\frac{0.059}{n} pH$$
(2)

where it can be deduced that the slope is different from the theoretical value (30 mV \(\ne\) 59 mV). These values indicate that the Tz oxidation reaction occurs via a proton with 2 transferred electrons in the process (reaction 2) [55, 56]. Furthermore, the slope -30 mV/pH indicated in Fig. S3b verifies that two electrons are transferred n = 1.96 ~ 2. [55, 56]. In addition, the highest Tz current oxidation intensity is observed when pH is 8. Therefore, this pH was chosen for the following experiments.

Fig. 4
figure 4

DPV analysis of a) CNF/AuNP-CPE in phosphate buffer 0.1 mol L−1 as supporting electrolyte at different pH values (6–9) in the presence of Tz 4 × 10–4 mol L−1, b) CNF-CPE and CNF/AuNP-CPE in the presence of 9.9 × 10–5 mol L−1 Tz at pH 8 at 75 mV and 25 mV s.−1

Full size image

The Fig. 4b presents the electrochemical evaluation of the CNF-CPE and CNF/AuNP-CPE compounds in the presence of 9.9 × 10–5 mol L−1 Tz, with an oxidation peak at 0.86 V, for the CNF-CPE sensor modified with the AuNP, which improves the Tz electrochemical detection. After obtaining the optimal electrolyte, the DPV parameters were optimized for Tz detection. Then, the amplitude (Eamp = 150 mV), scan rate (ʋ = 25 mV s−1), accumulation potential (Ea = 1.5 V) and accumulation time (ta = 60 s) were defined. [11, 57] (see Supplementary Fig. S4). The studies of the optimal parameters were carried out in the following way: The amplitude was evaluated in a range from 5 to 200 mV using a 9.9 × 10–5 mol L−1 solution of Tz while the other conditions remained constant, the results are observed in Fig S4a, where it is observed that the highest oxidation current is obtained at 150 mV of amplitude. The same procedure was performed to obtain the optimal parameters leading to an improvement of the Tz electrochemical detection by the CNF/AuNP-CPE sensor [58]. The results are shown in Fig. S4.

Electrochemical determination of Tz

The CVs of the CNF/AuNP-CPE sensor were recorded in absence and presence of Tz at pH = 8 in the potential range from -0.2 V to 1.3 V vs Ag/AgCl (Fig. 5). Figure 5a presents the CV of the CNF without modifying, where the oxidation potential at 1.1 V (reaction 1) and the reduction potential at 0.5 V are observed. These peaks correspond to the redox reaction of AuNP on the CNF surface according to [59, 60]:

$${Au}^{0} \leftrightarrow {Au}^{III}+3{e}^{-}$$
(Reaction 1)
Fig. 5
figure 5

Electrochemical evaluation of CNF/AuNP-CPE and CNF-CPE in a 0.1 mol L−1 phosphate buffer pH 8. Analysis by cyclic voltammetry a) In the absence of Tz and b) in the presence of Tz at 25 mV s−1; c) chronoamperometry analysis and d) chronocoulometry analysis in the absence and in the presence of 9.9 × 10–5 mol L−1 Tz at 0.91 V during 120 s

Full size image

Figure 5b shows a Tz oxidation peak at 0.9 V detected by CNF-CPE and CNF/AuNP-CPE materials [56]. The CNF-CPE material detects the Tz, through oxidation measurements. In contrast, the CNF/AuNP-CPE sensor provides well-defined signals for the anodic current of Tz. This demonstrates the synergistic effect of AuNP on CNF surface, due to the good conductivity and rapid electron transfer attributed by the AuNP, which is related to the decreasing in the overpotential reaction on the modified electrode surface [34, 61]. The electrochemical process involved in Tz presents only one oxidation signal implying two electrons and one proton in the reaction, thus depicting an irreversible process. The electrochemical oxidation reaction of Tz is shown in reaction (2) [34, 62].

(Reaction 2)

Chronoamperometry is one of the electrochemical methods most commonly used to study processes in chemically modified electrodes. For example, our CNF/AuNP-CPE sensor in the detection of Tz, exhibits a potential of 0.91 V and t = 120 s, generating a noticeable increase in current due to the Tz molecule oxidation (Fig. 5c) [22, 34]. The inset in Fig. 5c shows the modification of those straight lines, increasing their intercept from 0.005 to 3.503 μA, due to detection of Tz by the CNF/AuNP-CPE sensor [22, 61].

Figure 5d shows the adsorbed charge by Tz on CNF/AuNP-CPE surface, determined by chronocoulometric analysis. The extrapolation of the linear slope, from the chronocoulometric test, does not pass through the origin due to the double layer charge and the oxidation reaction on Tz (Eq. 3)

$$Q=\frac{2nFAD_O^{1/2}C_O^\ast t^{1/2}}{\pi^{1/2}}+Q_{dl}+nFA\Gamma _O$$
(3)

where n is the stoichiometric number of electrons involved in the electrode reaction F is the Faraday constant 96,485 C mol−1, A is the electrode electroactive area in cm2, \({D}_{o}\) is the diffusion coefficient in cm2 s−1, \({C}_{o}^{*}\) is the concentration of the species in the solution in mol cm−3, t is the time in s, \({Q}_{dl}\) is the capacitive charge and \({\Gamma}_{O}\) (mol cm−2) is the faradaic component associated with the reduction and oxidation of the molecule adsorbed on the surface, \(\Gamma_O\) [63]. The charge present on the CNF/AuNP-CPE was 9.91 μC and the amount of Tz charge adsorbed on the surface of CNF/AuNP-CPE was 77.95 μC [64].

Studies of scan rate effects

The effect of different scan rates was evaluated to observe the height variation of the anodic current for Tz detected by the CNF/AuNP-CPE sensor. Fig. S5 shows the electrochemical analysis of Tz at 9.0 × 10–4 mol L−1 by CV at different scan rates from 0.01 to 0.2 V s−1. It is observed that the scan rate linearly increases with the oxidation current (ipa) (Fig. 6S). In Fig. S6a, the same behavior is observed between the maximum Tz oxidation current (ipa) and the scan rate (υ), which is expressed as ipa (μA) = 3.35 + 83.70υ (r2 = 0.9956), showing that the oxidation currents of Tz were enhanced with increasing scan rate from 0.01 to 0.2 V s−1. A shift to more positive potentials was also observed, suggesting that the reaction mechanism of Tz at the electrode gradually changed from an absorption process to a diffusion one [11, 16]. Fig. S6b, indicates that ipa is proportional to the square root of the scan rate in the range of 0.01 to 0.2 V s−1, ipa(μA) = -1.75 + 4.55υ1/2 (r2 = 0.9416), suggesting a diffusion controlled mechanism [61]. Fig. S6c presents the linear relationship of Log(ipa) with Log(υ). The linear equation is Log(ipa) = -4.405 + 0.510Log(υ) (r2 = 0.9319). The slope of the graph (Log(ipa) vs Log(υ)) is 0.51, suggesting a mixed diffusion-adsorption controlled system (slope between 0.5 and 1.0) [34, 65]. In addition, Fig. S6d presents the linear relationship between Log(ipa) and Log(Epa). The linear equation is Log(ipa) = -3.54 + 26.58Log(Epa) (r2 = 0.9761) which indicates a mixed diffusion-adsorption controlled system.

Fig. 6
figure 6

a) DPV analysis of the CNF/AuNP-CPE sensor in the presence of Tz at different concentrations, in a phosphate buffer 0.1 mol L−1 pH = 8, at 150 mV, 25 mV s−1, ta = 60 s and Ea = 1.5 V b) Calibration curve of the anodic current heights as a function of Tz concentrations

Full size image

Fig. S6e shows the linear relationship between Epa and the logarithm of the scan rate, (Log(υ)). The linear equation for this relationship is: Epa = 0.925 + 0.038Log(υ), (r2 = 0.9022). This indicates the irreversible nature of the electrochemical process of Tz [59]. Since the Tz reaction is an irreversible process, the number of electrons transferred in the reaction can be obtained from Eq. (4) through the difference between the Epa and E1/2, which corresponds to the peak current, and the half of the peak current (α = 0.50), respectively [61].

$${E}_{pa}- {E}_\frac{1}{2}= \frac{0.0477}{\alpha .n}$$
(4)

The results suggest an electron transfer process (n = 1.94 ~ 2) in the rate-determining step of the Tz oxidation [59].

Calibration curve and Tz detection

The electrochemical detection of Tz by the CNF/AuNP-CPE sensor was examined using DPV in a 0.1 mol L−1 phosphate buffer solution at pH 8.0 under optimized experimental conditions. The results are shown in Fig. 6a, where the addition of the Tz increases the current proportionally. The results in Fig. 6b show that the oxidation current linearly increases with the concentration of Tz [66]. The calibration curve obtained for the Tz response was Δi(µA) = -0.179 + 50110[Tz], with a correlation coefficient r2 = 0.9946, a linear range between 3.29—49.5 µmol L−1, LOD of 0.80 ± 0.07 µmol L−1 and LOQ of 2.64 ± 0.23 µmol L−1. These measurements were performed by triplicate. The LOD of the developed sensor was compared with different previous studies (Table 1), showing that our CNF/AuNP-CPE sensor is within the working range and within the maximum permissible limits [16].

Table 1 Comparison of different modifications of working electrodes based on carbon for Tz detection
Full size table

DFT Study of the interactions between Tz and the gold nanoparticles

The tetrahedral geometry attributed to the Au20 cluster has been described as a structure of face-centered cubic bulk gold, with the twenty gold atoms on its surface and each of its four faces depicting a (111) plane [40, 62]. The constructed polyhedra is a tetrahedral pyramid with its four corners being low coordination sites. Four configurations were considered for the Tz-AuNP assemblies (See Fig. 7). The lowest energy (-46.6 kcal mol−1) corresponds to a Tz parallel orientation relative to an Au20 face (Fig. 7a) where phenyl and pyrazole moieties interact with the gold atoms displaying averaged interatomic distances of 3.35 Å, and 3.37 Å, respectively. In this interacting assembly the nitrogen atoms of the azo group are the closest ones to Au at a distance of 3.32 Å, larger than the sum of van der Waals radii (3.21 Å) suggesting noncovalent interactions between the fragments, in line with a physisorption phenomenon. The second configuration (Fig. 7b) with the phenyl and pyrazole rings interacting with the AuNP exhibits a lower magnitude interaction energy (-41.9 kcal mol−1), as a consequence of the Tz bent-shaped geometry-given the contacts involving the sulfonate and carboxyl groups toward the low coordinate sites of the Au20 cluster- which increases the distance between phenyl and the cluster, resulting in a less efffective interaction between the fragments. As expected, lower interaction energies (about 11 kcal mol-1) are found for the side-to-face binding modes (Fig. 7c and d), where the Tz sulfonate groups interact with the central Au atoms of the cluster face. In summary, DFT calculations reveal that a parallel orientation of the Tz molecule with respect to the surface of Au20 is favored, with interatomic distances suggesting a physisorption at the electrode-solution interface of the CNF/AuNP-CPE sensor.

Fig. 7
figure 7

Optimized geometries of Tz-AuNP systems calculated at the M06-2X/6-311G(d)/LANL2DZ level of theory. Interaction energies in kcal mol−1 and interatomic distances in Angstroms

Full size image

Study of the reproducibility, repeatability and selectivity of the CNF/AuNP-CPE sensor

Repeatability and reproducibility analysis of the CNF/AuNP-CPE sensor by DPV using a Tz concentration of 4.75 × 10–4 mol L−1 in a phosphate buffer solution of 0.1 mol L−1 at pH 8.0 and using 3 different electrodes, were performed by 10 consecutive measurements (n = 10) in 4 different days. The application of the analytical method yielded a relative standard deviation (RSD) lower than 10%, as shown in Table 2. The results obtained showed that the proposed method exhibits an excellent precision when applied for Tz dye detection.

Table 2 Results of the RSD obtained by the reproducibility and repeatability of the sensor
Full size table

The selectivity analysis of the proposed sensor was performed in the presence of the following interferents: sunset yellow (SY), sodium benzoate (SB), NaNO3, NaHCO3, glucose (Glu) and citric acid (CA) at a concentration of 1 × 10–4 mol L−1 in a phosphate buffer 0.1 mol L−1 of pH 8 in presence of 4.75 × 10–4 mol L−1 of Tz. Figure 8 presents the current variation percentage due to the interfering species presence. The percentage of relative standard deviation (%RSD) varies from 94.8 to 113.5%, implying that the CNF/AuNP-CPE sensor is selective. This demonstrates that the CNF/AuNP-CPE sensor can be successfully applied for Tz dye detection in the presence of the aforementioned interfering compounds.

Fig. 8
figure 8

DPV analysis of the CNF/AuNP-CPE sensor in the presence of different interferents in phosphate buffer 0.1 mol L−1 pH = 8, at 150 mV, 25 mV s−1, ta = 60 s and Ea = 1.5 V

Full size image

Validation of the electrochemical method by UV–visible spectrophotometry

Our sensor was validated using an UV–Vis analysis, and it was used as a comparative method for Tz quantification in synthetic and real samples. First, a calibration curve for Tz was constructed by the UV–visible technique in the same linear concentration range used for the DPV electrochemical technique (Fig. 9 and Table 3).

Fig. 9
figure 9

Analysis of Tz in synthetic samples using a), b) the UV–visible spectrophotometric method and calibration curve; c), d) the electrochemical method and calibration curve and e) Validation of the electrochemical method

Full size image
Table 3 Analysis of Tz in synthetic samples using the electrochemical method and the UV–visible spectrophotometric method
Full size table

Figure 9a presents a maximum absorbance around 430 nm of Tz [50, 74, 75]. From this UV–vis spectrum, an analytical curve was obtained, which exhibits a linear behavior with r2 = 0.9995 in the concentration range between 10.0 to 50.0 μmol L−1 of Tz, from the following equation: A = -0.00303 + 0.02017[Tz] (Fig. 9b). In the same way, as observed in Fig. 9c, the electrode was analyzed by DPV with a concentration range between 10.0 to 50.0 μmol L−1 of Tz, obtaining the following equation: Δi = -2.61 × 10–7 + 2.97 × 10–7[Tz], and a linear behavior with r2 = 0.9977 (Fig. 9d).

Since the DPV method was evaluated at different levels of Tz concentration in synthetic samples. These results are compared with those obtained by the spectrophotometric method (Fig. 9e). The DPV validation by the spectrophotometric method was carried out by simple linear regression, obtaining a high coefficient of determination value (r2) of 0.998, where the high percentage of the data is explained with the adjusted linear regression model: CDPV = 0.002 + 0.999 CUV-vis. In addition, the linear regression slope presents an upper and lower confidence limits of + 1.071 and -0.928, respectively, where the obtained slope value is in good agreement with the ideal value of 1.0 for this model. On the other hand, the upper and lower confidence limits for the intercept are + 2.361 and -2.357, respectively. The values include the ideal value of zero, with a confidence level of 95%. Therefore, from the results is concluded that the spectrometric method is a useful analytical technique to validate the DPV analysis. The CNF/AuNP-CPE electrochemical sensor provides reliability and good accuracy for a wide range of Tz concentrations [49, 76]. The results of the analysis of the real sample are indicated in Table 4, evaluating the Inca Kola® and Cifrut® soft drinks.

Table 4 Analysis of Tz in soft drinks using the electrochemical method and the UV–visible spectrophotometric method
Full size table

Conclusion

A new electrochemical sensor based on carbon nanofibers modified with AuNP was developed for Tz dye detection. The morphology of the AuNP on the surface of the CNF was observed by SEM. The crystallite size of the AuNP (face centered cubic structure) ranges between 15 and 20 nm. The CNF/AuNP-CPE presents excellent stability, reproducibility and repeatability, with linear a range between 3.29—49.5 µmol L−1, with r2 = 0.9946, with a low detection limit of 0.80 ± 0.07 µmol L−1 and a limit of quantification of 2.64 ± 0.23 µmol L−1. The robustness of the sensor selectivity in the presence of various interfering species (sunset yellow, sodium benzoate, NaNO3, NaHCO3, glucose and citric acid) has been demonstrated. DFT calculations in a PCM solvent model reveal that a parallel orientation of Tz relative to the Au20 surface is favored, where the interatomic distances involved suggest a physisorption phenomenon at the electrode-solution interface. Finally, the CNF/AuNP-CPE sensor exhibits a rapid response, making it an appealing analytical tool for the detection and quantification of Tz dye in soft drinks.