Electrochemical Detection of Nitrite in Food Based on Poly (3,4-ethylenedioxythiophene) Doped with Fe3O4 Nanoparticles Loaded Carboxylated Nanocrystalline Cellulose

Carboxylated nanocrystalline cellulose (CNCC) was prepared by oxidation degradation of microcrystalline cellulose (MCC) using ammonium peroxydisulfate and modified with Fe3O4 nanoparticles to form Fe3O4-CNCC nanocomposite via simple refluxing process. The Fe3O4-CNCC nanocomposite doped poly(3,4-ethylenedioxythiophene) (PEDOT) was successfully decorated on the glassy carbon electrode (GCE) by electrochemical deposition. The PEDOT/Fe3O4-CNCC modified GCE with enlarged real electrochemical surface area was used to determine nitrite with high selectivity, sensitivity and outstanding reproducibility. Using amperometric current-time (i-t) curve, the proposed sensor provided a wider linear range (0.5–2500 μM) and a lower detection limit (0.1 μM) towards nitrite compared with the method of differential pulse voltammetry (DPV). This analytical method gave good selectivity in the practical measurement of nitrite in pickles.


Introduction
Nanocellulose (NCC), one kind of sustainable functional nanomaterials, is derived from native cellulose. [1][2][3][4][5] Their morphology mainly depends on the source of precursor cellulose and the conditions of preparation. NCC exhibits remarkable properties, such as high surface area, thermally stable, renewable, biodegradable, non-toxic [6][7][8] and accessible industrially in large scale. 9 NCC has been extensively researched as key components in the design of super capacitors, 10-12 aerogels, 13 sensors, 14 pharmaceuticals, 15 chiral materials 16 and catalysts. 17 They are generated by the removal of amorphous regions of different sources of cellulose using acid hydrolysis, enzymatic or mechanical treatments. As a functional derivative of NCC, carboxylated nanocrystalline cellulose (CNCC) possesses better dispersibility and stability compared to NCC because of high density of carboxyl groups on its surface. In this work, CNCC was prepared using a strong oxidant, ammonium peroxydisulfate, instead of conventional acid hydrolysis. 18 As an important magnetic nanomaterial, Fe 3 O 4 magnetic nanoparticle received much attention and extensive investigation in the past years for its excellent electrocatalytic activity, biocompatibility and absorption ability. [19][20][21][22][23] For example, a sensitive electrochemical biosensor for the detection of H 2 O 2 from living cells has been developed based on graphene blended with Fe 3 O 4 nanoparticles. 24 Fe 3 O 4 nanosphere decorated with Au nanoparticles was used as a kind of catalyst for the detection of As(III) in water. 25 However, Fe 3 O 4 nanoparticles are thermodynamically unstable and tend to aggregate to form bulk particle. To overcome this problem, Fe 3 O 4 -CNCC was synthesized by dispersing Fe 3 O 4 nanoparticles on the surface of CNCC in this work. Fe 3 O 4 -CNCC can be doped into poly (3,4-ethylenedioxythiophene) (PE-DOT) to enhance its electrical conductivity. 26 Electrochemically polymerized PEDOT composites with differ-ent dopant have been widely used in various kinds of electrochemical sensors. [27][28][29][30] Nitrite, a kind of food additive, is widely used in the food industry. However, excessive nitrite is detrimental to human health. 31,32 In human body, nitrite can lead to the irreversible oxidation reaction of hemoglobin to methemoglobin and also can be converted into nitrosamine, which would cause hypertension and cancer. 33 Therefore, the accurate detection of nitrite becomes a pressing issue both in the human body and food industries. Till now, various analytical techniques have been used to detect nitrite. For example, Zhang used chemiluminescence to detect nitrite with the detection limit of 0.024 μM. 34 Huang adopted spectrofluorimetry to determine nitrite in human saliva with the detection limit of 0.5 μM. 35 However, most of these methods are lengthy, expensive, require complicated procedure and expert knowledge that make them unsuitable for routine analysis. To the contrary, electrochemical techniques 36,37 provided a highly sensitive and rapid nitrite determination. Besides, the electrochemical approach was an environmentally friendly method because no extra chemical loading is required.
In this paper, Fe 3 O 4 -CNCC was synthesized through a simple refluxing process and then doped PEDOT onto a glass carbon electrode by electrochemical deposition. Combining the advantages of CNCC, Fe 3 O 4 , and PEDOT, the fabricated sensor can determine nitrite with high selectivity, sensitivity and outstanding reproducibility. This method gave good selectivity in the practical measurement of nitrite in pickles.

Apparatus and Measurements
All electrochemical experiments were performed using a conventional three-electrode system containing a bare or PEDOT/CNCC, PEDOT/Fe 3 O 4 -CNCC modified glassy carbon working electrode (3 mm diameter), a platinum wire counter electrode and a saturated calomel reference electrode on a CHI660D (Shanghai CH Instruments Co., China) electrochemical work station. The structure and morphology of the samples were characterized by transmission electron microscopy (TEM; Hitachi High-Technology Co., Ltd., Japan) and field emission scanning electron microscopy (SEM; Hitachi High-Technology Co., Ltd., Japan). XRD spectra were recorded on an X-ray diffractometer (Shimadzu7000S, Shimadzu Analytical, Japan) equipped with CuKα radiation (λ = 0.154 nm). Zeta potentials of samples were measured with the Malvern Zetasizer Nano ZS90 (Malvern InstrumentsLtd., UK ).

3. Synthesis of CNCC
CNCC was prepared by a one-pot green procedure treatment of the mixture of MCC and APS. In brief, 2.5 g MCC and 57.05 g APS were dissolved in 250 mL of deionized water to form white suspension. The mixture was heated at 70 °C under stirring for 12 h to give a suspension of CNCC and then cooled to room temperature naturally. The suspension was centrifuged and washed with deionized water for several times until the pH of the suspension was close to that of deionized water. Finally, the product was lyophilized to yield a white powder.

4. Synthesis of Fe 3 O 4 -CNCC Composite
Fe 3 O 4 -CNCC was synthesized through a simple refluxing process. Typically, 100 mL of the as-synthesized CNCC (8 mg mL -1 ), 0.1127 g of FeSO 4 7H 2 O and 0.1918 g FeCl 3 6H 2 O were added to the three-necked flask. Then the pH value of solution was buffered close to 10 using 25 % NH 3 H 2 O, and the solution was heated at 60 °C in flow of N 2 for 2 h. The prepared reaction products were put into the reaction still and heated at 80 °C for 2 h. The obtained solution was isolated in a centrifugal machine at the rate of 10000 rpm for 5 min and then washed thoroughly and dispersed with deionized water. Eventually, the Fe 3 O 4 -CNCC was composed following the processes discussed previously.

5. Preparation of the PEDOT/Fe 3 O 4 -CNCC Modified Electrode
Before the modification, GCE was polished with 0.3 μm, 0.05 μm alumina slurries in sequence and then cleaned by ultrasonication in deionized water, ethanol and deionized water for 3 min, respectively.

6. Electrochemical Measurements
Electrochemical impedance spectroscopy (EIS) measurements were recorded in 5.0 mM [Fe(CN) 6 ] 3-/4solution containing 0.1 M KCl within a frequency range of 1-100,000 Hz. The amplitude of the applied sine wave was 5 mV, and the direct current potential was set at 0.80 V. Both cyclic voltammetry (CV) and differential pulse voltammetry (DPV) used to study the electrochemical reaction of nitrite were carried out in 0.2 M phosphate buffered saline (PBS, pH 7.4, containing 0.9% NaCl) containing 5 mM nitrite. The CV was recorded from 0.4-1.1 V at a scan rate of 100 mV s −1 , while the DPV was performed from 0.4-1.4 V with the following parameters: pulse amplitude 50 mV; pulse width 0.2 s and pulse period 0.5 s. Amperometric current-time (i-t) curve was performed in stirring PBS (0.2 M, pH 7.4) with the potential set at 0.80 V. All experiments were conducted at ambient temperature.

7. HPLC Measurements
The HPLC technique was employed as a reference method to measure nitrite in real samples. The measurement was based on the national standards of China (GB 5009.33-2016 [38] ).

1. Working Mechanism of PEDOT/ Fe 3 O 4 -CNCC Modified Electrode
The electrode modification and electrocatalytic mechanism are shown in Fig 1. CNCC was prepared by degradation of MCC using ammonium peroxydisulfate, and then Fe 3 O 4 -CNCC was synthesized by a simple refluxing process. The obtained Fe 3 O 4 -CNCC nanocomposite doped poly(3,4-ethylenedioxythiophene) (PEDOT) was successfully decorated on the GCE by electrochemical deposition. The PEDOT/Fe 3 O 4 -CNCC modified electrode was used to determine nitrite in PBS. Based on the following results of the electrochemical tests, the corresponding transformation is shown as follows:

2. Characterization of the Fe 3 O 4 -CNCC
Fig S1 shows the zeta potentials of different materials. The zeta potentials of NCC and CNCC were measured as -20.46 mV and -30.05 mV, which were ascribed to the high density of the carboxyl groups on the CNCC surface. Therefore hydrophilic CNCC has good dispersiveness and stability. The zeta potential of Fe 3 O 4 -CNCC was -34.05 mV, which made it easily to be doped in the polymerizing EDOT to neutralize the positive charges on the backbone of PEDOT.
The The TEM images of the CNCC and Fe 3 O 4 -CNCC are shown in Fig 2. Clearly, the CNCC was rod-like and the average diameter of CNCC was about 10 nm (Fig 2a). When modified on CNCC, Fe 3 O 4 nanoparticles formed uniformly on the surface of CNCC (Fig 2b). PEDOT/ CNCC (Fig 2c) exhibited a network-like wrinkled surface morphology. Compared to PEDOT/CNCC, the obtained PEDOT/Fe 3 O 4 -CNCC (Fig 2d) showed different morphology and the diameter of the PEDOT/Fe 3 O 4 -CNCC was significantly wider than that of the PEDOT/CNCC.

Electrochemical Characterization of PEDOT/Fe 3 O 4 -CNCC/GCE
CV and EIS were useful tools to research the nature of the modified electrodes and the influence of nanoparticles on electron-transfer kinetics. 40 Fig 3A shows CVs recorded in a 0.1 M KCl aqueous solution containing 5 mM Fe(CN) 6 3-/4-. The bare GCE (curve a) showed poor electron-transfer kinetics for the Fe(CN) 6 3-/4redox couple with an obvious peak-to-peak separation (ΔE p ). Both the PEDOT/CNCC/GCE (curve b) and PEDOT/Fe 3 O 4 -CNCC/GCE (curve c) exhibited a reversible redox response. The reversible redox peak currents at PEDOT/ Fe 3 O 4 -CNCC/GCE (curve c) was larger than that at bare GCE and PEDOT/CNCC/GCE, indicating that the introduction of Fe 3 O 4 could markedly increase the electron transfer ability of the electrode. The EIS for Fe(CN) 6 3-/4at different electrodes are shown in Fig 3B. The diameter of semicircle could be used to estimate the electron-transfer resistance (Ret) [37]. The Ret value obtained at bare GCE (curve a, 197.8 Ω) was significantly higher than that of the PEDOT/CNCC/GCE (curve b, 53.54 Ω). The PEDOT/ Fe 3 O 4 -CNCC/GCE (curve c) was linear curve, implying that the Ret was close to zero, which may be ascribed to PEDOT/Fe 3 O 4 -CNCC composite film being highly conductive and the unique microstructure of the PEDOT/ Fe 3 O 4 -CNCC can increase the effective surface area of the modified electrode. The PEDOT/Fe 3 O 4 -CNCC/GCE showed much lower electron-transfer resistance than PE-DOT/CNCC/GCE and bare GCE, which was quite in agreement with the results obtained from CV.

4. Electrochemical Detection of Nitrite at the PEDOT/Fe 3 O 4 -CNCC/GCE
The electrochemical response of the different electrodes toward nitrite was examined by CV, as shown in also investigated. As shown in Fig S4A, the oxidation current increased with the increasing of the scan rate and a liner relationship can be obtained between the peak current versus square root of scan rate in the range from 0.05 to 0.19 V/s (R 2 = 0.997). The diffusion-controlled irreversible electrochemical process of nitrite oxidation on the PE-DOT/Fe 3 O 4 -CNCC/GCE was determined. Fig S4B also confirmed that sodium nitrite did not adsorb to the electrode surface and therefore did not affect the subsequent experimental determination.

5. Amperometric Response to Nitrite Detection
For better nitrite detection, the effect of deposition time was investigated. As exhibited in Fig 6, when the deposition time reached 180 s, the oxidation peak current of nitrite reached the maximum. With a shorter deposition time, the deposited PEDOT/Fe 3 O 4 -CNCC was not sufficient, and with a longer deposition time, the thicker PEDOT/Fe 3 O 4 -CNCC composite will peel off during the reaction. Therefore, 180 s was selected as the optimal deposition time.   (Fig 7). The maximum peak current was obtained at pH 7 and then decreased slowly. We chose pH 7.4 for further studies because it was closely related to our physiological pH ranges.
The detection of nitrite with different concentrations at PEDOT/Fe 3 O 4 -CNCC/GCE was performed under the optimized deposition time. A well-defined anodic peak was observed at about 0.8 V (vs SCE), and the peak current increased with the increase of nitrite concentration (Fig 8).   μM, R 2 = 0.996) The LOD was calculated to be 0.1 μM. We achieved to develop a nitrite sensor with low LOD and wide liner range that is superior to previously reported sensors, as listed in Table 1.

6. Selectivity, Stability, Repeatability, and Reproducibility
The selectivity of a sensor is a crucial factor for its practical application. Some possible interfering substances in electrochemical detection of nitrite, such as ascorbic acid (AA), dopamine (DA), uric acid (UA), sodium citrate and sodium benzoate were added to nitrite solution ( Fig  10). Clearly, a large oxidation peak current was obtained in A) B) PBS (0.2 M, pH 7.4) containing 0.1 mM nitrite (Fig 10a). After 0.5 mM AA, 0.5 mM DA, 1 mM UA, 1 mM sodium citrate and 1 mM sodium benzoate were added into 0.1 mM nitrite solution, the nitrite oxidation peak current declined by 4% but oxidation peak potential was not changed (Fig 10b).

7. Determination of Nitrite in Pickles
To demonstrate the practical feasibility of such an electrochemical sensor, the PEDOT/Fe 3 O 4 -CNCC/GCE was used to detect nitrite in pickles. The pickle samples were purchased from the local market in Qingdao, Shandong, and were pretreated as follows. The samples were ground adequately in a juicer, a portion of 10 g of puree was mixed with 50 mL of water, followed by sonication for 30 min in a beaker. Then the resulting mixture was centrifuged for 10 min at 4000 rpm. Finally, the supernatant was diluted to 100 mL using 0.2 PBS (pH 7.4). The recovery of nitrite was calculated using the standard addition method. The standard HPLC method was employed for the comparison. Table 2 lists the results obtained by these two  methods. The recoveries from 96.5% to 102% suggest that results from the proposed method are comparable with those from HPLC. Noticeably, without the use of extraction columns which are required for HPLC, the simple and rapid pretreatment of the proposed method is suitable for on-site detection. These results imply the good accuracy, reliability, and feasibility of PEDOT/Fe 3 O 4 -CNCC/ GCE for the detection of nitrite in pickle samples.

Conclusions
A simple electrochemical sensor was developed based on PEDOT/Fe 3 O 4 -CNCC composite and used to determine nitrite. It is confirmed that rod-like CNCC works as a substrate and the Fe 3 O 4 magnetite nanoparticles grow on the surface of CNCC. The prepared Fe 3 O 4 -CNCC nanocomposite can be doped into PEDOT to enhance its electrical conductivity. As a key element of the nitrite sensor, the PEDOT/Fe 3 O 4 -CNCC nanocomposite combines the advantages of CNCC and Fe 3 O 4 , as well as PEDOT, and can enhance the measurement of nitrite. The PEDOT/Fe 3 O 4 -CNCC/GCE has been applied to determine nitrite with high selectivity, low detection limit (0.1 mM), large linear range (0.5-2450 μM) and outstanding stability and reproducibility. More importantly, the electrochemical sensor based on PEDOT/Fe 3 O 4 -CNCC has shown great potential application for nitrite detection in food safety analysis. Table 2. Results of the determination of nitrite in pickles (n = 4).

Samples
The