The morphologies of CeO2/eGr, CeO2, and eGr were observed by SEM and TEM. The eGr/CeO2 nanocomposite used here was 1:1 in weight ratio (1:1 wt.). From Fig. 2a, it shows that CeO2 uniformly loaded on the surface of eGr, and the corresponding elemental mapping (Fig. 2b, c, d and e) illustrated the elements of C, O and Ce are existed in CeO2/eGr nanocomposite. From Fig. 2f, the synthesized CeO2 presents nano-cubic structure and agglomerated together due to the nanosize effects. The diffraction rings in the selected area electron diffraction (SAED) suggests the as-synthesized CeO2 is polycrystalline and mainly exists (111), (200), (220) and (311) crystallite planes (Fig. 2g). The lattice space distance of 0.314 nm belongs to (111) crystallite plane of CeO2 (Fig. 2h). The eGr displays layered structure (Fig. 2i), and SAED indicates that the eGr mainly presents (002), (004) crystallite planes (Fig. 2j), the lattice space distance of 0.34 nm is the theoretical value of graphene (002) crystallite plane, it verifies the high quality of the prepared eGr.
The crystallite structure and composition of CeO2, eGr and eGr/CeO2 composite (1:1 wt.) were evaluated by Raman spectra, XRD and XPS. In Raman spectra of Fig. 3a, it shows a characteristic peak located at 461 cm− 1 for CeO2 sample, which stems from the symmetrical stretching of Ce-O vibrational and originates from the F2g vibrational mode of CeO2 phase (Tan 2015). For eGr, the peaks at 1356, 1580 and 2710 cm− 1 were assigned to D, G and 2G bands of graphene, respectively (Yu 2018). The D band at ~ 1356 cm− 1 is deriving from the defects and structural disorder in the sp2-carbon nanomaterials. The G band at 1580 cm− 1 is relating to the in-plane vibrations of the 2D hexagonal graphene lattice. eGr/CeO2 composite sample possesses both Raman characteristics of CeO2 and eGr. XRD was using to analyze the structure of the prepared materials. In Fig. 3b, the strong and sharp diffraction peaks indicate all the samples are in good crystallinity. For eGr, two diffraction peaks at 26.4°and 54.5° are observed, which related to (002) and (004) planes of graphene, this is in accordance with the SEAD result (Fig. 2j). For CeO2, the diffraction peaks located at 28.5°, 33.1°, 47.5°, 56.3°, 59.1°, 69.4°, 76.7°, 79.1°and 88.4° are corresponding to (111), (200), (220), (311), (222), (400), (331), (420), (422) planes (JCPDS 81–0792). The eGr/CeO2 composite contains all the characteristic peaks of CeO2 and eGr as well. Base on the eGr/CeO2 composite contains all the features of CeO2 and eGr, the chemical composition of eGr/CeO2 composite was further characterized by XPS. The XPS survey spectrum (Fig. 3c) reveals the existence of Ce, O, C elements in eGr/CeO2 composite. The Ce3d electron core line was analyzed and depicted in Fig. 3d, it can be deconvoluted into 8 peaks and labeled as v0, v1, v2, v3 (3d3/2 region) and u0, u1, u2, u3 (3d5/2 region). Peaks v0, v2, v3 and u0, u2, u3 are characteristics of Ce(IV) 3d final states, while, v1 and u1 are Ce(III) 3d final states (Mullins 1998). Therefore, the as-prepared CeO2 contains part of Ce(III), and the percentage of Ce(III) was calculated by Eq. (1), which based on the fitted areas of the corresponding peaks of Ce(III) and Ce(IV).
$${\left( {C{{\text{e}}^{3+}}} \right)_{surf}}=\frac{{Ce(III)}}{{Ce(III)+Ce(IV)}}$$
1
The calculated percentage of Ce(III) is ~ 20%, which is similar to the previous reported CeO2 nanomaterials (Lyu 2022). The presence of Ce(III) indicates that the formation of oxygen vacancies, which can provide catalytically active sites for the sensor. The O1s spectrum can be separated into three peaks and illustrated in Fig. 3e, the peak locates at ~ 529.9 eV corresponding to the crystal lattice oxygen in Cerium oxide CeO2. The peak located 532.2 eV and 533.4 eV could be respectively related to the oxygen vacancy and the adsorbed oxygen on the composite (Meng 2013; Yang 2014). The C1s spectrum can be separated into three peaks (Fig. 3f). The peak placed at 284.6 eV corresponds to the sp2 carbon atoms or attribute to C = C (Parvez 2014). The other small peaks at 286.1 and 287.9 eV correspond to C–O and C = O on the surface of the composite, respectively.
The prepared eGr, CeO2 and eGr/CeO2 composite (1:1 wt.) were respectively casted on glass carbon electrode (eGr/GCE, CeO2/GCE and eGr/CeO2/GCE), and their electrochemical performances were firstly estimated by Cyclic voltammetry (CV) at 50 mV s− 1 in the solution of 5 mM [Fe(CN)6]3−/4− and 0.1 M KCl. The bare GCE electrode was conducting as control sample. From Fig. 4a, all electrodes show different levels of electrochemical activities, after evaluated the redox peak current densities and CV curve area, the electrochemical activity follows the order of eGr/CeO2/GCE > CeO2/GCE > eGr/GCE > GCE. This suggests that the eGr/CeO2/GCE has the largest specific surface area, the best electrochemical kinetics and activities, which could arise from the synergistic effects of excellent catalytic active sites of CeO2 and good electron transference of eGr. In addition, the standard heterogeneous rate constant (k0) for bare GCE, eGr/CeO2/GCE, eGr/GCE, CeO2/GCE were calculated by Nicholson’s equation (Nicholson 1965) and the values are respectively 0.0041 cm·s− 1, 0.0077 cm·s− 1, 0.0045 cm·s− 1, 0.0049 cm·s− 1. The eGr/CeO2/GCE has the highest value of 0.0077 cm s− 1 that verifies eGr/CeO2 composite has the best electrochemical activity.
The electro-active surface area is a critical factor for electrochemical sensor, which was estimated in [Fe(CN)6]3−/4− solution with scan rates ranging from 0.03 to 0.45 V s− 1 via Randles-Sevcik equation (Eq. (2)) (Sha 2019).
$${I_{{\text{pa}}}}=(2.69 \times {10^5}){n^{3/2}}A{D^{1/2}}C{v^{1/2}}$$
2
Where n refers to electron transfer number, A is the active surface area, C is the concertation of [Fe(CN)6]3−/4−, v is the scan rate and D is the diffusion coefficient. Here, n = 1, D = 6.6⋅10− 6 cm2 s− 1 [27] for 5 mM K3[Fe(CN)6] solution containing 0.1 M KCl. For eGr/CeO2/GCE (Fig. 4b), the active surface area of eGr/CeO2/GCE was calculated to be 0.097cm2 (Fig. 4c), which is the highest than that of the electroactive surface areas of GCE (0.04 cm2), eGr/GCE (0.08 cm2), and CeO2/GCE (0.065 cm2), which displayed in Supporting Information, Fig. S1. This result agrees well with the electrochemical activity order of the prepared sensors. The roughness factor (fr) of the electrochemical sensors was calculated to evaluate the actual active surface area by comparing the oxidation peak current (Ipa) of the prepared sensor to bare GCE for [Fe(CN)6]3−/4− reaction (Krzyczmonik 2014). The ratio of the oxidation peaks current for two electrodes is equal to the electrode surface area (Eq. (3)):
$${f_r}=\frac{{{I_{P2}}}}{{{I_{P1}}}}=\frac{{{A_2}}}{{{A_1}}}$$
3
The fr determined by electrochemical methods depends not only on the size of the electrode (the actual surface), but also on the number of redox centers that can be reached on the surface. Therefore, the fr was calculated to 2.425, 2, and 1.625 for eGr/CeO2/GCE, eGr/GCE, and CeO2/GCE, respectively.
The GCE, eGr/GCE, CeO2/GCE and eGr/CeO2/GCE (1:1 wt.) for 4-BPA detection were characterized by CV in the electrolyte of absence and presence 50 µmol L− 1 4-BPA in 0.1 mol L− 1 phosphate buffer (pH = 3). As displayed in Fig. 5a, when presence 50 µmol L− 1 4-BPA, all electrodes present one oxidation peak, which indicates the 4-BPA is electrochemically detectable and the reaction of 4-BPA is irreversible. The eGr/CeO2/GCE (1:1 wt.) shows the highest oxidation peak current (Ipa) and the lowest onset potential, this verified that eGr/CeO2 composite has higher sensitivity for electrochemical detection of 4-BPA, which should be attributed to the synergetic effect of the catalytic properties of CeO2 and the fast electron transference of eGr. The ratios between CeO2 and eGr have further been measured and shown in Fig. 5b. With CeO2: eGr ratio increasing from 0:4 to 1:1, the oxidation peak current of 4-BPA increases and reaches the maximum at the ratio of 1:1, then the peak current drops with further increasing CeO2 content. The reason could be CeO2 is a semiconductor, and it provides electrocatalytic activity sites. When CeO2 content is too low, and it will not create enough activity sites. While the content is higher than 1: 1, the conductivity and electron transference of the electrode will decrease. Therefore, the optimum ratio was 1: 1 for 4-BPA detection and selected in the following study.
The different loading amounts of eGr/CeO2 composite on GCE were measured with 10 µmol L− 1 4-BPA (Supporting information, Fig.S2). It found that the oxidation peak current of 4-BPA increases with the loading volume increasing to 6 µL, while the response decreased when the loading amount further increased (Fig. 5c). This could be arising from the electrode surface was not covering enough when the loading composites below 6 µL, while too much loading amount will hinder the activity sites that cause the decrease on the response peak current (Zamiri 2019).
The PBS, Britton-Robison (B-R), and acetic acid-sodium acetate buffer solutions were evaluated as supporting electrolytes for 4-BPA detection (Supporting information, Fig.S3). Among these supporting electrolytes, the PBS buffer solution shows more sensitivity for 4-BPA detection. Therefore, PBS was selected as supporting electrolyte. Moreover, pH is another key impact factor for electrochemical analysis. The pH values range from 3 to 6.5 were evaluated in 0.1 M PBS buffer solution (Supporting information, Fig.S4). It can be seen that the optimum response pH for 4-BPA was 3, and the response gradually decreases as the pH increasing (Fig. 5d). This phenomenon could be due to the conductivity loss and presence of carboxyl groups with pH increasing (Chen 2020).
The oxidation process of 4-BPA on eGr/CeO2/GCE was further studied by linear sweep voltammograms (LSVs), different scan rates (50 mV s− 1 to 450 mV s− 1) were conducted and 20 µmol L− 1 4-BPA was used. As exhibited in Fig. 6a, the Ipa increases with scan rates increased. Moreover, the oxidation peaks positively shifted. More importantly, the oxidation peak current increased linearly with the square root of scan rates (Fig. 6b), and the linear regression is Ipa = 3.936v1/2-9.318, R2 = 0.999, which indicates that the electrochemical oxidation of 4-BPA on eGr/CeO2/GCE was controlled by diffusion process (Huang 2020). The relationship between Epa and lnv was presented by Laviron’s theory (Sequoia 1979):
$${E_{{\text{pa}}}}={E^0}+\left( {\frac{{RT}}{{\alpha nF}}} \right)\ln \left( {\frac{{RT{k^0}}}{{\alpha nF}}} \right)+\left( {\frac{{RT}}{{\alpha nF}}} \right)\ln v$$
4
where α is the charge transfer coefficient, E0 is the apparent potential, n is the number of electron, v is the scan rate, the value of R, T, F is 8.314 J K− 1 mol− 1, 298 K, 96485 C mol− 1, respectively. Therefore, the number of electron can be calculated via the linear equations of Epa - lnv (Supporting information, Fig.S5). Generally, for a totally irreversible electrode process, the value of α is assumed to be 0.5. Hence, the value of n is calculated to be 2.
DPV shows sensitive response to low concentrations as compared to LSV. Therefore, DPV was using to detect 4-BPA in PBS solution with different concentrations. As illustrated in Fig. 6c, the peak current increases linearly with the concentrations of 4-BPA varying from 0.3 to 150 µM. However, there are two linear relationships obtained. From Fig. 6d, in the range of 0.3 to 20 µM, the linear regression equation is Ipa = 0.75c + 0.08, (R2 = 0.991), and from 20 to 150 µM, the linear relationship is Ipa = 0.199c + 11.24, (R2 = 0.993). Moreover, the lowest detection limit (LOD) was calculated to be 0.06 µmol L− 1 according to the following equation of 3s/m, where m is slope of the regression equation and s is the stand deviation of the response.
The repeatability of the eGr/CeO2/GCE was carried out with 10 µmol L− 1 4-BPA by the means of LSV (Supporting information, Fig.S6). After 10 continuous measurements (Fig.S7a), the relative standard deviation (RSD) of the oxidation peak currents was found to be 2.35% for 4-BPA. After storing the electrode at 4 ℃ for 15 days, the electroactive oxidation of 4-BPA was just reducing 3.21% compared to the original currents. To investigate the reproducibility of the sensor, six replicated measurements for 10 µmol L− 1 4-BPA were carried out using 7 different eGr/CeO2/GCEs that made same method and same materials. The results showed excellent reproducibility with relative standard deviations of 3.6%. These results indicate that the proposed sensor has good repeatability and reproducibility.
To estimate the ability of anti-interference of the eGr/CeO2/GCE, some regular interfering species were tested. From Fig.S7b, no considerable interferences were observed from the following compounds: K+, Na+, Mg2+, rutin, quercetin, fenitrothion, imidacloprid, clothianidin, IAA, SA, glucose, sucrose (peak current change < 6%). However, when adding the homologs of phenoxyacetic acid (4-chlorophenoxyacetic acid, 4-fluorophenoxyacetic acid), there exist some interferences. Therefore, we need to explore more sensitive, stable materials to solve the problem in the future study.
To evaluate the practicability of eGr/CeO2/GCE, the sensor was using to detect 4-BPA in real apple samples. It was found that no response of 4-BPA in the apple sample, and the recoveries were then evaluated by standard addition method and the analytical results are listed in Table 1.
Table 1
Results of the recovery analysis of 4-BPA in apple sample (n = 3)
Sample
|
Added value(µM)
|
Determined value(µM)
|
Recovery(%)
|
1
|
0
|
Not detected (a, b)
|
-
|
2
|
0.5
|
0.45 ± 0.01
|
90%
|
3
|
1
|
1.08 ± 0.04
|
108%
|
4
|
3
|
2.78 ± 0.03
|
93%
|
5
|
5
|
4.86 ± 0.04
|
97%
|
a The 4-BPA level determined by proposed eGr/CeO2/GCE |
b The 4-BPA level determined by HPLC system |