3.1. Characterization of TiO2/Co/NCNTs, ZIF-67@TiO2 composites
Morphological features of TiO2/Co/NCNTs and ZIF-67@TiO2 have been well characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) in Fig. 1. As shown in Fig. 1a and b, the SEM and TEM images exhibit the rough surface of dodecahedral shapes with the average size of 500 nm after hydrolysis process, displaying the coating of TiO2 via TBOT hydrolyzes on the surface of rhombic dodecahedral ZIF-67. Figure 1c and d shows the morphology of TiO2/Co/NCNTs by carbonization of ZIF-67@TiO2 precursor at 800°C. SEM image (Fig. 1c) exhibits the structure shrinkage and inner-invaginated surface of polyhedral compared to ZIF-67@TiO2 owing to the pyrolysis of organic ligand. TEM (Fig. 1d) of TiO2/Co/NCNTs dodecahedron reveals that Co nanoparticles (black dots) are uniformly embedded in the carbon framework (gray matrix). Notably, a typical in-situ growth of N-doped CNTs with adjustable size were found wrapped the polyhedra.
Figure 2 display the high-resolution transmission electron microscope (HRTEM) images. The NCNTs were distributed on the edge of the composites and Ti elements at the head of NCNTs and Co at the center of the dodecahedron. As displayed in Fig. 2b and c, the inter plane fringes distance of Co nanoparticles is 0.205 nm consistent with the (111) crystal plane of the Co crystal (Jin et al. 2015). The lattice fringe of the head of NCNTs is consistent with rutile TiO2 having (110) lattice fringes with an interspacing of 0.325 nm (Xu et al. 2018). The energy dispersive spectrometer (EDS) mappings (Fig. 1e − k) reveal the existence of C, N, O, Co and Ti elements in as-prepared TiO2/Co/NCNTs composites. These results confirmed that carbonized ZIF-67@TiO2 consists of the N-doped nanotubes with TiO2 and CoNPs located at the upper and lower ends around the surface of the rhombohedral dodecahedron, respectively.
The phase composition of ZIF-67@TiO2 and TiO2/Co/NCNTs are investigated by X-ray diffraction (XRD) patterns (Fig. 3a). All diffraction peaks of ZIF-67@TiO2 ranging from 5–80° are indexed to the simulated ZIF-67 and rutile TiO2 phase (JCPDS No. 21-1276), respectively, proving the successfully synthesis of ZIF-67@TiO2. Furthermore, the diffraction peaks of TiO2/Co/NCNTs composites situate at 44.2°, 51.5° and 75.8° corresponding to (111), (200) and (220) planes match well with a cubic body-centered structure of Co (JCPDS 15–0806). Correspondingly, the XRD peaks at 27°, 36°, 41°, and 54° correspond to rutile TiO2 with the JCPDS card no. 21-1276, which corresponds to (110), (101), (111) and (211) planes of rutile TiO2.
The quantitative atomic composition of the TiO2/Co/NCNTs material surface is typically investigated by XPS. The XPS survey spectrum depicted in Fig. 3b shows the elemental compositions of cobalt (Co, ~ 780 eV), titanium (Ti, ~ 458 eV), carbon (C, ~ 284 eV), nitrogen (N, ~ 400 eV) and oxygen (O, ~ 530 eV), which are matched with the EDS mappings in Fig. 1. Figure 2c
depicts the high-resolution spectrum of Co 2p. The core-level Co 2p3/2 spectrum consists of five peaks centered at 778.8, 781.5, 793.3 and 796.5 eV, which is assigned to Co metal (Yin et al. 2016). As shown in Fig. 3d, the Ti spectrum of the TiO2/Co/NCNTs can be fitted into the Ti4+ (456.9 eV, 456.9 eV, 458.3 eV, 458.7 eV and 464.3 eV) (Erdem et al. 2001).
N2 adsorption/desorption isotherm of TiO2/Co/NCNTs belong to type-IV isotherm with a slight hysteresis loops, manifesting mesoporous structure of this material (Fig. 3e). The Brunauer-Emmett-Teller (BET) surface area is 283.2 m2·g− 1 of TiO2/Co/NCNTs. The high surface area of the TiO2/Co/NCNTs composites would facilitate electron transfer and expose more active sites to improve electrocatalysis performance. In addition, the pore size distribution curve (Fig. 3f) illustrate that TiO2/Co/NCNTs have abundant mesopores with an average pore size of ~ 4.7 nm, which are formed during pyrolysis.
3.2. Electrochemical Impedance Spectroscopy (EIS) and electroactive surface area
The electron transfer resistance (Rct) is the measurement standard to investigate electron transport characteristics. Diameter of the semicircle represents Rct. As shown in Fig. 4a, ZIF-67@TiO2/GCE, bare GCE, TiO2/Co/NCNTs/GCE demonstrate gradually decreasing in semicircle part. Compared to bare GCE, the Rct value of ZIF-67@TiO2 is much larger, it might be due to the poor conductivity of ZIF-67 although TiO2 is a semiconductor. After carbonization of ZIF-67@TiO2, TiO2/Co/NCNTs has the smaller Rct value, revealed that TiO2/Co/NCNTs possess a smaller charge transfer resistance and higher charge transfer rate. This result mainly attributes to the special structure that Ti exists in the NCNTs head and Co presents in the central dodecahedron, and superior conductivity of NCNTs.
The electroactive surface area of the modified electrode is another important parameter for electrochemical performance. Further cyclic voltammetry (CV) of bare GCE, ZIF-67@TiO2/GCE and TiO2/Co/NCNTs/GCE in 5.0 mM [Fe(CN)6]3−/4− were studied, and depicted in Fig. 4b. A pair of redox peaks observe at TiO2/Co/NCNTs/GCE with anodic peak currents of 2.6 and 1.6 times larger than those of bare GCE, and ZIF-67@TiO2/GCE, respectively. The TiO2/Co/NCNTs sensor has a faster electron transfer velocity due to the synergistic effect of enhance active sites, large electroactive surface area and high conductivity.
According to inset of Fig. 4b, anodic peak currents of the probe have a good linear relationship with v1/2, the electrode effective area was achieved by Randles–Sevcik equation. According to equation:
where Ip, n, v, A, D0, C0 refer to the anode peak current, the number of electron transfer, the scanning rate, the surface area of the electrode, the diffusion coefficient (7.6×10− 6 cm2·s− 1) and the concentration of the probe, respectively. From the slope of Ip vs ν1/2 curves, the electroactive areas are calculated as 0.016 (bare GCE), 0.026 (ZIF-67@TiO2/GCE) and 0.062 cm2 (TiO2/Co/NCNTs/GCE), respectively. The highly efficient surface in terms of available surface area at the fabricated sensor manifest that TiO2/Co/NCNTs is an ideal material towards TBHQ detection.
3.3. Electrochemical behaviors of TBHQ at TiO2/Co/NCNTs/GCE
The electrochemical parameter of GCE on modification with ZIF-67@TiO2 and TiO2/Co/NCNTs is investigated through cyclic voltammetry technique (CV) in 0.1 M phosphate buffer solution (PBS, pH 7.0) with 100 µM TBHQ. CV profiles are displayed in Fig. 5a, exhibits a pair of distinct redox peaks for all the electrodes due to the electro-oxidation of tert-butylhydroquinone (TBHQ) to tert-butylquinone (TBQ). Analyzing Fig. 5a, bare GCE shows small current response towards TBHQ with oxidation and reduction peaks at 0.281 and − 0.262 V. However, the CV of ZIF-67@TiO2/GCE resulted in well-defined redox peaks at 0.155 (EOx) and − 0.159 V (ERed), respectively, and the current response is almost 4.5-fold than that of bare GCE. Because of the synergistic catalysis effect between TiO2, Co nanoparticles and NCNTs, TiO2/Co/NCNTs/GCE displays prominent increase in redox peak current response with minimum peak-to-peak separation (ΔEp=60 mV). The results demonstrated that the fabricated TiO2/Co/NCNTs/GCE could provide a sensitive and accurate way for TBHQ detection.
Additionally, the electrochemical response of TBHQ (100 µM) at the different electrodes was examined by DPV technique (Fig. 5b). On bare GCE, an oxidation peak of TBHQ is observed at around 0.152 V with slight current signal of 0.464 µA. On ZIF-67@TiO2/GCE, the anodic current value of 7.789 µA is enhanced clearly and the potential is slightly sifted negative to -0.036 V. Interestingly, the response current of TiO2/Co/NCNTs/GCE is the largest (57.84 µA), which is almost 124-fold than that of the bare GCE. This may be due to their unique structural and compositional features. The excellent performance of TBHQ may be due to: (1) TiO2 with superior physical and chemical properties could facile electron transfer. (2) Co nanoparticles exists in the central dodecahedron is beneficial for accelerating electron transfer and improving active sites (Wang et al. 2013). (3) NCNTs with good conductivity and large surface area are benefit to improve the mass transport of electrochemical reaction (Xiong et al. 2010).
3.4. Effect of sweep rate
The electron transfer kinetics in electrochemical detection of TBHQ based on TiO2/Co/NCNTs/GCE was examined by CV technique for analyzing the effect about scan rate variation vs. redox peak currents. Figure 6a depicts the CV curves of TiO2/Co/NCNTs/GCE containing 100 µM TBHQ by changing the sweep rate from 20 mV/s to 400 mV/s. Clearly, the CV profile displays the gradual linear increase in anodic and cathodic peak currents with scan rate. The proportionality of the peak currents against the root of scan rates is plotted in Fig. 6b. The respective regression equations are obtained as: Ipa (µA) = 4.522v1/2 (mV1/2·s− 1/2) − 5.001 (R2 = 0.998) and Ipc (µA) = − 4.655v1/2 (mV1/2·s− 1/2) + 5.947 (R2 = 0.997), respectively. Based on this evaluation, the redox behavior of TBHQ at the TiO2/Co/NCNTs sensor is a typical diffusion-controlled (Karthikeyan et al. 2019).
Moreover, it is noteworthy that the redox peak potential of TBHQ progressively shift to the positive direction with the rise scanning rate (Fig. 6c). The oxidation peak potential (Epa) and reduction peak potential (Epc) are linear to the logarithm of scan rate (lgv) ranging from 200–400 mV s− 1, which are expressed as Epa (V) = 0.1693 lgv (mV s− 1) – 0.3121 (R2 = 0.995) and Epc (V) =–0.1642 lgv (mV s− 1) − 0.2520 (R2 = 0.988), respectively. Based on the Laviron's equation, the relationship between the redox peak potentials (Ep) and scan rate (v) is described as the following equations (Laviron E, 1979).
$${E_{{\text{pa}}}}{\text{=}}{E^{0'}}+\frac{{2.3RT}}{{\left( {1 - \alpha } \right)nF}}\lg \nu$$
2
$${E_{{\text{pc}}}}{\text{=}}{E^{0'}} - \frac{{2.3RT}}{{\alpha nF}}\lg \nu$$
3
$$\lg {k_s}=\alpha \lg (1 - \alpha )+(1 - \alpha )\lg \alpha - \lg \frac{{RT}}{{nF\nu }} - \alpha (1 - \alpha )\frac{{nF\Delta {E_{\text{p}}}}}{{{\text{2}}{\text{.303}}RT}}$$
4
wherein v, R, T, F, ks, α and n represent the scan rate, gas constant, thermodynamic temperature, Faraday’s constant, electron transfer rate constant, electron transfer coefficient, electron transfer number. The calculate values of α, n and ks for TBHQ are 0.51, 1.69 (approximate 2) and 0.76 s− 1, respectively, implying that TBHQ at TiO2/Co/NCNTs/GCE undergoes a two-electron electrochemical redox process (see Scheme 1) (Fan et al. 2018). The higher ks value obtained at TiO2/Co/NCNTs sensor conform that the electron transfer is faster than other electrodes (Ma et al. 2021).
3.5. Effect of pH and TiO2/Co/NCNTs amount
The influence of buffer solution pH for TBHQ detection in various pH ranges from 4.0 to 8.0 are depicted in Fig. 6d. The anodic peak potential gradually shifts to the negative potential with the pH increases, suggesting that the redox reaction of TBHQ is influenced by proton. A good linear relationship is observed between buffer solution pH and peak potential (Fig. 6e), and the fitted regression equation is Epa (V) = − 0.050 pH + 0.255 (R2 = 0.997). Obviously, the slope of the equation is − 50 mV/pH, close to the theoretical slope of − 59 mV per pH at 298K, indicating that the ratio of protons to electrons is equal for TBHQ on the TiO2/Co/NCNTs/GCE surface. Furthermore, TiO2/Co/NCNTs has the high current response as the pH value is 7.0. At high pH, the electrochemical reaction of TBHQ becomes more difficult because of the lack of protons. At low pH, the electrode might be unstable. Thus, 7.0 was set as the optimized pH for the determination of TBHQ in this experiment.
The TiO2/Co/NCNTs amount on the GCE, as a vital parameter for the electrochemical signal of detecting TBHQ, is investigated by DPV ranging from 3 µL to 11 µL (Fig. 6f). It can be clearly seen that, when 7 µL is dripped, the anodic peak current for TBHQ reached largest value. The higher amounts may hinder electron transfer, resulting in prolong the electron diffusion time, thus reducing the electrocatalytic activity of TBHQ on TiO2/Co/NCNTs film. Therefore, the optimal TiO2/Co/NCNTs amount could be selected as 7 µL.
3.6. TBHQ detection at TiO2/Co/NCNTs/GCE
Under optimal experimental conditions, the current responses of the TiO2/Co/NCNTs for TBHQ quantitative detection was examined by DPV with the higher sensitivity and lower background signal. Apparently, Fig. 7a displays that as the TBHQ concentration increases, the anodic current responses also increase linearly with no significant shift in potential. The dynamic linear range is 0.04–10.0 µM and 10.0–100.0 µM, respectively (Fig. 7b). The two linear regression equation of the equivalent plots are obtained for Ipa (µA) = 1.526C (µM) – 0.083 (R2 = 0.993, 0.04–10.0 µM) and Ipa (µA) = 0.486C (µM) + 13.01 (R2 = 0.997, 10.0–100.0 µM), respectively. The limit of detection (LOD) at TiO2/Co/NCNTs/GCE is given by 4 nM (signal-to-noise (S/N) of 3). Table 1 states other reported TBHQ sensor, and the as-prepared sensor outperforms them in terms of wider linear range and lower LOD.
Table 1
Comparison of the analytical parameters obtained using different modified electrodes for the determination of TBHQ
Sensor | Linear range(µM) | LOD(µM) | Ref. |
ZIF-8/MWCNTs- COOH/GCE | 0.01–0.1; 0.1–2.5 | 0.009 | Wang et al. 2023 |
MIP/AuNPs/GCE | 0.08–100 | 0.07 | Fan et al. 2018 |
MIP/ZC/GCE | 1–75 | 0.42 | Ma et al. 2021 |
PEDOT/CNT/GCE | 0.5–820 | 0.12 | Wang et al. 2021 |
Co3O4NRs/FCB/SPCE | 0.12–62.2 | 0.001 | Balram et al. 2022 |
MnO2/ERGO/GCE | 1-300 | 0.8 | Cao et al. 2019 |
TiO2/Co/NCNTs | 0.04-10; 10–100 | 0.004 | This work |
MIP: Molecularly imprinted polymer. PEDOT: poly(3,4-ethylene dioxythiophene). MIP/ZC: molecularly imprinted ZIF-8 derived nanoporous carbon. FCB: chemically oxidized carbon black. ERGO: reduced graphene oxide.
3.7. Interferences, repeatability, reproducibility and stability
In order to assess the selectivity of TiO2/Co/NCNTs-based sensor for TBHQ detection, the DPV current signal of TBHQ was applied to analyze in existence of interferents. A few organic and inorganic interfering chemicals (including structural analogues) were selected as potential interferents for detection. Figure 8a. compared peak response of TBHQ (20 µM) and other compounds (20 µM of butylated hydroxytoluene (BHT), butylated hydroxyanisole (BHA), propyl gallate (PG), ascorbic acid, hydroquinone (HQ) and 100 µM of glucose, 1000 µM of K+, Mg2+, NO3− and Ca2+) in 0.1 M PBS buffer (pH 7.0). And the interferences have barely influence on the current response, the oxidation peak current of TBHQ don’t vary largely, indicating that this TiO2/Co/NCNTs sensor has strong anti-interference ability for TBHQ detection.
Six different TiO2/Co/NCNTs-based sensors were fabricated in parallel and examined in 0.1 M PBS buffer with 20 µM TBHQ. As illustrated in Fig. 8b, the relative standard deviation (RSD, n = 3) of DPV current of six paralleled TiO2/Co/NCNTs/GCE is just 2.3%, proving the proposed sensor have sustained reactivity in preparation. As repeatability is a vital factor to be considered especially in sensing applications, DPV response of TiO2/Co/NCNTs/GCE in existence of TBHQ (20 µM) was measured 10 times consecutively. As depicted in Fig. 8c, one TiO2/Co/NCNTs/GCE electrode has an RSD of 2.6%, approving outstanding repeatability property at this electrode. After storing at room temperature for 20 days, the oxidation current at the TBHQ concentration of 20 µM shows above 92% retention (Fig. 8d). All the interpretations from the above measurements demonstrated that the constructed TiO2/Co/NCNTs/GCE had favorable selectivity, reproducibility, repeatability and stability.
3.8. Analysis in real sample
To verify the validity of TiO2/Co/NCNTs/GCE, TBHQ in the edible oil (corn oil, rapeseed oil and blend oil) samples was detected via the standard addition method. In brief, the oil samples (5 mL) was mixed with alcohol and stirred continuously for 1.5 hour. Then, the mixture was centrifugation at 5000 rpm for 30 min, and the supernatant was extracted for 3 times as extract solvent. Typically, 400 µL of oil samples were respectively diluted to 20 mL with 0.1 M PBS by DPV. Good recovery results were obtained in Table 2, displaying the recovery range of 98.0–103.4%. The values evaluated from the provided method match with the HPLC method, demonstrating that the TiO2/Co/NCNTs/GCE sensor can supply an available analysis method for TBHQ in food samples.
Table 2
Recoveries of TBHQ from commercial edible oil samples using the standard addition method (n = 3)
Sample | Detected (µM) | Spiked (µM) | Found (µM) | Recovery (%) | RSDb (%) | HPLC (µM) | |
1a | 1.7 | 5 | 6.87 ± 0.18 | 103.4b | 2.62 | 6.94 ± 0.22 |
| 1.7 | 10 | 11.5 ± 0.12 | 98.0 | 1.04 | 11.2 ± 0.37 |
2 | 0 | 5 | 4.92 ± 0.14 | 98.4 | 2.84 | 5.05 ± 0.21 |
| 0 | 10 | 9.93 ± 0.29 | 99.3 | 2.92 | 9.86 ± 0.34 |
3 | 2.3 | 5 | 7.41 ± 0.30 | 102.2 | 4.05 | 7.27 ± 0.28 |
| 2.3 | 10 | 12.1 ± 0.47 | 98.0 | 3.89 | 12.6 ± 0.38 |
a Samples: 1, Corn oil; 2, Rapeseed oil; 3, Blend oil.
b %Recovery = 100×(cFound −cDetected) / cSpiked