An Electrochemical Sensor for Diphenylamine Detection Based on Reduced Graphene Oxide/Fe3O4-Molecularly Imprinted Polymer with 1,4-Butanediyl-3,3′-bis-l-vinylimidazolium Dihexafluorophosphate Ionic Liquid as Cross-Linker

In this paper, we report a new composite of reduced graphene oxide/Fe3O4-ionic liquid based molecularly imprinted polymer (RGO/Fe3O4-IL-MIP) fabricated for diphenylamine (DPA) detection. RGO/Fe3O4-IL-MIP was prepared with RGO/Fe3O4 as supporter, ionic liquid 1-vinyl-3-butylimidazolium hexafluorophosphate ([VC4mim][PF6]) as functional monomer, ionic liquid 1,4-butanediyl-3,3’-bis-l-vinylimidazolium dihexafluorophosphate ([V2C4(mim)2][(PF6)2]) as cross-linker, and diphenylamine (DPA) as template molecule. Fourier transform infrared spectroscopy, thermal gravimetric analysis, scanning electron microscopy, and vibrating sample magnetometer were employed to characterize the RGO/Fe3O4-IL-MIP composite. RGO/Fe3O4-IL-MIP was then drop-cast onto a glassy carbon electrode to construct an electrochemical sensor for DPA. The differential pulse voltammetry (DPV) peak current response for 20 μM DPA of RGO/Fe3O4-IL-MIP modified glassy carbon electrode (GCE) was 3.24 and 1.68 times that of RGO/Fe3O4-IL-NIP and RGO/Fe3O4-EGDMA-MIP modified GCEs, respectively, indicating the advantage of RGO/Fe3O4-IL-MIP based on ionic liquid (IL) as a cross-linker. The RGO/Fe3O4-IL-MIP sensor demonstrated good recognition for DPA. Under the optimized conditions, the RGO/Fe3O4-IL-MIP sensor exhibited a DPA detection limit of 0.05 μM (S/N = 3) with a linear range of 0.1–30 μM. Moreover, the new RGO/Fe3O4-IL-MIP based sensor detected DPA in real samples with satisfactory results.


Introduction
Diphenylamine (DPA) is used as a pre-or post-harvest scald inhibitor for some fruits, a rubber antioxidant, and a solid fuel rocket propellant [1]. Residues of DPA are found in fruit and environmental water samples [2]. The presence of DPA residues in fruits [3] and the environment pose a hazard to human health since it is classified as a probable human carcinogen. To monitor and control the overuse of this harmful compound, accurate detection of DPA in fruit and environmental water samples is important and desirable. Different analytical methods, such as high-performance liquid (DPA), 1-naphthylamine, 1,4-phenylenediamine, phenol, and 2,2'-azobisisobutyronitrile (AIBN) were purchased from Aladdin Industrial Corporation (Shanghai, China). Chitosan was obtained from Sinopharm Group Chemical Regent Co., Ltd. (Shanghai, China). RGO/Fe 3 O 4 (45:55, w/w) was purchased from Xianfeng Nanotechnology Co., Ltd. (Nanjing, China). [VC 4 6 ] was provided by Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences (Lanzhou, China).

mim][PF
[V 2 C 4 (mim) 2 ][(PF 6 ) 2 ] was prepared in our laboratory and characterized by 1 H NMR and 13 C NMR. Phosphate buffer was prepared using NaH 2 PO 4 and Na 2 HPO 4 . Deionized water of 18 MΩ cm was used throughout the experiments.

Instrumentation
Surface morphology measurements were carried out using a JSM-7500F scanning electron microscope (SEM) (JEOL, Tokyo, Japan). Fourier transform infrared (FT-IR) spectra were recorded on Nicolet Nexus-470 FT-IR spectrometer (Waltham, MA, USA). Thermogravimetric analysis (TGA) was conducted using an STA-449F3 instrument (Netzsch, Selb, Germany). A vibrating sample magnetometer (MPMS3) was used to investigate the magnetic properties of samples. The 1 HNMR and 13 CNMR spectra were recorded with a Varian 400-MR spectrometer (Palo Alto, CA, USA). All electrochemical experiments were implemented on a CHI660D electrochemical workstation (CHI Instruments Co., Shanghai, China) with a conventional three electrode system comprising of a platinum wire as the auxiliary electrode, a saturated calomel electrode (SCE) as the reference electrode, and a modified GCE (3 mm diameter) as the working electrode.

Characterization of RGO/Fe3O4-IL-MIP
FT-IR spectra of RGO/Fe3O4, RGO/Fe3O4-IL-MIP before and after extracting the template molecule, and RGO/Fe3O4-IL-NIP are shown in Figure 1a. The peaks at 3435 and 1635 cm −1 in the spectrum of RGO/Fe3O4 are assigned to the stretching vibrations of O-H and C=C [14,32], respectively. The peak at 562 cm −1 for RGO/Fe3O4 corresponds to the stretching vibration of Fe-O of Fe3O4 [33]. While comparing the spectra of RGO/Fe3O4-IL-MIP before extracting the template molecule with RGO/Fe3O4, peaks at 1164 and 837 cm −1 were observed, corresponding to the imidazolium cation vibration of the C-N bond and the P-F of PF6 − stretching vibration [34], respectively. The results suggest the functionalization of PF6 − based imidazolium ILs onto RGO/Fe3O4. The spectra of RGO/Fe3O4-IL-MIP before and after extracting the template molecule and RGO/Fe3O4-IL-NIP are similar. It is worthy to point out that DPA peaks are not obvious from the spectra of RGO/Fe3O4-IL-MIP before extracting the template molecule. This phenomenon may be explained by the fact that the DPA peaks were covered by RGO/Fe3O4-IL-MIP before extracting the template molecule. These results demonstrate the successful synthesis of RGO/Fe3O4-IL-MIP. Figure 1b

Characterization of RGO/Fe 3 O 4 -IL-MIP
FT-IR spectra of RGO/Fe 3 O 4 , RGO/Fe 3 O 4 -IL-MIP before and after extracting the template molecule, and RGO/Fe 3 O 4 -IL-NIP are shown in Figure 1a. The peaks at 3435 and 1635 cm −1 in the spectrum of RGO/Fe 3 O 4 are assigned to the stretching vibrations of O-H and C=C [14,32]

Electrochemical Behavior of RGO/Fe3O4-IL-MIP/GCE
To investigate the electrochemical behavior of RGO/Fe3O4-IL-MIP, we performed cyclic voltammetry (CV) experiments in 5 mM [Fe(CN)6] 3−/4− solution containing 0.1 M KCl (Figure 4). Bare GCE exhibited the lowest peak currents, while we observed the highest peak currents for RGO/Fe3O4/GCE, ascribed to the good electrocatalytic property of RGO/Fe3O4. The result also indicates the advantage of RGO/Fe3O4 as supporter. We observed higher peak currents for RGO/Fe3O4-IL-MIP/GCE compared with that of RGO/Fe3O4-IL-NIP/GCE. We attribute the higher peak current exhibited by the MIP to the cavities, as [Fe(CN)6] 3−/4− could pass through these cavities and reach the surface of the electrode more easily.

Electrochemical Behavior of RGO/Fe3O4-IL-MIP/GCE
To investigate the electrochemical behavior of RGO/Fe3O4-IL-MIP, we performed cyclic voltammetry (CV) experiments in 5 mM [Fe(CN)6] 3−/4− solution containing 0.1 M KCl (Figure 4). Bare GCE exhibited the lowest peak currents, while we observed the highest peak currents for RGO/Fe3O4/GCE, ascribed to the good electrocatalytic property of RGO/Fe3O4. The result also indicates the advantage of RGO/Fe3O4 as supporter. We observed higher peak currents for RGO/Fe3O4-IL-MIP/GCE compared with that of RGO/Fe3O4-IL-NIP/GCE. We attribute the higher peak current exhibited by the MIP to the cavities, as [Fe(CN)6] 3−/4− could pass through these cavities and reach the surface of the electrode more easily.  We studied the electrochemical behavior of 0 μM DPA on RGO/Fe3O4-IL-MIP/GCE using CV and DPV. There was no peak of CV and DPV observed at the RGO/Fe3O4-IL-MIP/GCE in 0.1 M phosphate buffer. We investigated the electrochemical behavior of 20 μM DPA on RGO/Fe3O4-IL-MIP/GCE using CV and observed an oxidation peak at 0.657 V without any reduction peak between 0.2 and 1.0 V, indicating DPA oxidation is an irreversible process [35,36] (Figure 5a). To examine the electrochemical responses for DPA, RGO/Fe3O4-IL-MIP/GCE and RGO/Fe3O4-IL-NIP/GCE were incubated in 20 μM DPA solution in phosphate buffer (0.1 M, pH 5.0) for 4 min, and the responses were measured by DPV (Figure 5b). RGO/Fe3O4-IL-MIP/GCE showed higher DPV peak current response for DPA than RGO/Fe3O4-IL-NIP/GCE. The DPV peak current response for 20 μM DPA of RGO/Fe3O4-IL-MIP/GCE was 3.24 times that of RGO/Fe3O4-IL-NIP/GCE. We again attribute this higher current response of RGO/Fe3O4-IL-MIP to the imprinted cavities. In addition, the DPV peak current response for 20 μM DPA of RGO/Fe3O4-IL-MIP/GCE was 1.68 times that of RGO/Fe3O4-EGDMA-MIP/GCE, indicating the superiority of RGO/Fe3O4-IL-MIP based on IL as a cross-linker.

Optimization of Experimental Conditions
Optimization of experimental conditions is important to achieve maximum effect by any system. We optimized the following experimental parameters for DPA detection: pH of the incubation solution, incubation time, and the concentration of RGO/Fe3O4-IL-MIP. While studying the influence of the incubation solution pH on DPV peak current of DPA, we observed a gradual increase in the oxidation peak current alongside an increase in pH from 4.0 to 5.0, which later decreased with a further increase in pH from 5.0 to 7.0 ( Figure 6a). Hence, we selected pH 5.0 as the optimum pH for further experiments. Figure 6b shows a sharp increase in the peak current within the first 4 min before reaching a plateau. Based on this, we selected 4 min as the incubation time for all subsequent experiments. With an increase in the concentration of RGO/Fe3O4-IL-MIP in HAc (1 M) containing chitosan (0.5 wt %) from 1.0 to 9.0 mg·mL −1 , the DPV peak current for DPA increased (Figure 7). We attribute this increase in the peak current to the expansion of the conductive electrode area to accumulate more DPA. When RGO/Fe3O4-IL-MIP concentration reached beyond 5.0 mg·mL −1 , the peak current decreased. This decrease can be explained in terms of the thickness of RGO/Fe3O4-IL-MIP restricting the transfer of DPA molecules to the GCE surface. Hence, we chose a concentration of 5.0 mg·mL −1 for RGO/Fe3O4-IL-MIP as the optimized value to construct our DPA sensor.

Optimization of Experimental Conditions
Optimization of experimental conditions is important to achieve maximum effect by any system. We optimized the following experimental parameters for DPA detection: pH of the incubation solution, incubation time, and the concentration of RGO/Fe 3 O 4 -IL-MIP. While studying the influence of the incubation solution pH on DPV peak current of DPA, we observed a gradual increase in the oxidation peak current alongside an increase in pH from 4.0 to 5.0, which later decreased with a further increase in pH from 5.0 to 7.0 ( Figure 6a). Hence, we selected pH 5.0 as the optimum pH for further experiments. Figure 6b

Analytical Performance and Selectivity of RGO/Fe3O4-IL-MIP Sensor
Under the optimized conditions, we examined the detection performance of the RGO/Fe3O4-IL-MIP/GCE (RGO/Fe3O4-IL-MIP sensor) towards DPA using DPV. Figure 8a shows a gradual increase in the oxidation peak currents of DPA with increasing DPA concentration, and the oxidation peak currents display a good linear dependence on DPA concentrations in the range of 0.

Analytical Performance and Selectivity of RGO/Fe3O4-IL-MIP Sensor
Under the optimized conditions, we examined the detection performance of the RGO/Fe3O4-IL-MIP/GCE (RGO/Fe3O4-IL-MIP sensor) towards DPA using DPV. Figure 8a shows a gradual increase in the oxidation peak currents of DPA with increasing DPA concentration, and the oxidation peak currents display a good linear dependence on DPA concentrations in the range of 0.

Analytical Performance and Selectivity of RGO/Fe 3 O 4 -IL-MIP Sensor
Under the optimized conditions, we examined the detection performance of the RGO/Fe 3 O 4 -IL-MIP/GCE (RGO/Fe 3 O 4 -IL-MIP sensor) towards DPA using DPV. Figure 8a shows a gradual increase in the oxidation peak currents of DPA with increasing DPA concentration, and the oxidation peak currents display a good linear dependence on DPA concentrations in the range of 0.1-30 µM (Figure 8b  To evaluate the selectivity of the RGO/Fe3O4-IL-MIP sensor, we studied the electrochemical responses of 20 μM DPA, 20 μM DPA analogs, and 20 μM DPA in the presence of some analogs. Figure 9a shows that the DPV peak current responses of the RGO/Fe3O4-IL-MIP sensor for DPA (a') are higher than other analogs (b', c', from d to k, DPV peak current responses were obtained from 0.2 to 1.0 V). In addition, we did not find any interferences for detecting 20 μM DPA (DPV signal change < 5%) in the presence of 10-fold 2-nitroaniline (b'), 3-nitroaniline (c'), 4-nitroaniline (d), benzene (e), 1,2-dimethylbenzene (f), 5-fold 1,4-phenylenediamine (g), hydroquinone (h), 1-fold catechol (i), 1naphthylamine (j), and phenol (k) (Figure 9b). Figure 9c shows the chemical structures of DPA and its analogs. The results suggest good imprinting efficacy achieved by RGO/Fe3O4-IL-MIP composites, enabling specific recognition ability towards the template DPA. Table 1 summarizes a comparative list of the results of DPA detection by RGO/Fe3O4-IL-MIP/GCE and other published electrochemical methods [6,[37][38][39]. The comparative results indicate that RGO/Fe3O4-IL-MIP/GCE with MIP material exhibits a low detection limit and high selectivity, which are attributed to the combined effect of good conductivity and imprinting effect.   To evaluate the selectivity of the RGO/Fe3O4-IL-MIP sensor, we studied the electrochemical responses of 20 μM DPA, 20 μM DPA analogs, and 20 μM DPA in the presence of some analogs. Figure 9a shows that the DPV peak current responses of the RGO/Fe3O4-IL-MIP sensor for DPA (a') are higher than other analogs (b', c', from d to k, DPV peak current responses were obtained from 0.2 to 1.0 V). In addition, we did not find any interferences for detecting 20 μM DPA (DPV signal change < 5%) in the presence of 10-fold 2-nitroaniline (b'), 3-nitroaniline (c'), 4-nitroaniline (d), benzene (e), 1,2-dimethylbenzene (f), 5-fold 1,4-phenylenediamine (g), hydroquinone (h), 1-fold catechol (i), 1naphthylamine (j), and phenol (k) (Figure 9b). Figure 9c shows the chemical structures of DPA and its analogs. The results suggest good imprinting efficacy achieved by RGO/Fe3O4-IL-MIP composites, enabling specific recognition ability towards the template DPA. Table 1 summarizes a comparative list of the results of DPA detection by RGO/Fe3O4-IL-MIP/GCE and other published electrochemical methods [6,[37][38][39]. The comparative results indicate that RGO/Fe3O4-IL-MIP/GCE with MIP material exhibits a low detection limit and high selectivity, which are attributed to the combined effect of good conductivity and imprinting effect.

Reproducibility and Stability of RGO/Fe 3 O 4 -IL-MIP Sensor
Reproducibility and stability are the two important parameters for applicability of any sensor. To examine the reproducibility of the imprinted sensor, we studied the current responses of DPA solution (20 µM) six times using the same RGO/Fe 3 O 4 -IL-MIP sensor. The current response showed a relative standard deviation (RSD) of 4.1%. Furthermore, after storing for two weeks in a refrigerator, the RGO/Fe 3 O 4 -IL-MIP sensor retained 94.7% of its initial current response for DPA (20 µM). These results certainly indicate the potential applicability of RGO/Fe 3 O 4 -IL-MIP sensor in DPA detection.

Analytical Application
To evaluate the feasibility of the RGO/Fe 3 O 4 -IL-MIP sensor towards the detection of DPA in real samples, we collected lake water samples from Jiaxing University and filtered through 0.45 µm filters. We also prepared food samples including pear and apple peels by grinding them to slurries followed by centrifuging to obtain supernatants. No DPA was detected in those real samples by the RGO/Fe 3 O 4 -IL-MIP sensor. Therefore, we employed the spiking method to demonstrate DPA detection by the RGO/Fe 3 O 4 -IL-MIP sensor. As summarized in Table 2, the recoveries of detection results ranged from 95.6% to 115.2%, demonstrating the reliable detection of DPA in water samples by the RGO/Fe 3 O 4 -IL-MIP sensor.

Conflicts of Interest:
The authors declare no conflicts of interest.