A Long-Term Stable Sensor Based on Fe@PCN-224 for Rapid and Quantitative Detection of H2O2 in Fishery Products

Hydrogen peroxide (H2O2) has been reported to be used for the illegal treatment of fishery products in order to obtain “fake” freshness. Residues of H2O2 in food may be of toxicology concern. In this study, a nonenzymatic sensor was developed based on Fe@PCN-224 metal–organic frameworks wrapped by Nafion to detect H2O2 concentration. The hybrid structure of Fe@PCN-224 was fabricated by incorporated free FeIII ions into the center of PCN-224, which was ultra-stable due to the strong interactions between Zr6 and the carboxyl group. Scanning electron spectroscopy images exhibited that Nafion sheets crossed together on the surface of Fe@PCN-224 nanoparticles to form a hierarchical and coherent structure for efficient electron transfer. Electrochemical investigations showed that the Fe@PCN-224/Nafion/GCE possessed good linearity from 2 to 13,000 μM (including four orders of magnitude), low detection limits (0.7 μM), high stability in continuous monitoring (current remained nearly stable over 2300 s) and in long-term measurement (current decreased 3.4% for 30 days). The prepared nanohybrid modified electrode was effectively applied to H2O2 detection in three different fishery products. The results were comparable to those measured using photometrical methods. The developed electrochemical method has a great potential in detecting the illegal management of fishery products with H2O2.


Introduction
Nowadays, food quality control has become increasingly important due to the growing demand for high-quality and sanitary food [1,2]. Among a variety of foods, the freshness of fishery products is the most crucial commercial quality factor for consumers [3][4][5]. The Regulation (EC) N 2406/96 of the European Parliament and the Council defines four categories for fresh fish products. The fishery products classified as the last category must be judged as not suitable for humans consumption and withdrawn from the market [6].
Because of the actions of many endogenous and exogenous enzymes, fishery products are perishable and easy to have appearance changes with off-flavors [7]. However, some illegal treatments on these products may simulate "fake" freshness, one of which is to use hydrogen peroxide (H 2 O 2 ). Illegal treatment with 0.5-0.8% H 2 O 2 aqueous solution has been reported [8], which causes whitening and "fresh" effects on fishery products due to the oxidation properties of H 2 O 2 [9,10].
In fact, H 2 O 2, with ready availability and an affordable price, can convert trimethylamine (TMA, a kind of degradation product) to trimethylamine-N-oxide (TMAO, amine oxide in living fishes) [11,12]. If the amount of TMAO is increased by H 2 O 2 treatment, the protein of muscle tissue will be stabilized [13]. Moreover, the decrease of main glycoproteins on the fish skin will reduce viscosity and slow down the appearance of off-flavor [14]. Consuming these foods containing excessive H 2 O 2 can cause nausea, headaches and potential risks of cancer [15][16][17]. The residual H 2 O 2 must be removed from dairy foods during the processing of foods, according to US FDA regulations [18]. Consequently, a rapid and

Instruments
A scanning electron microscope (Zeiss Sigma 500, Oberkochen, Germany) was used to analyze morphology. A UV-visible spectrophotometer (Evolution 300, Waltham, MA, USA) was applied to detect H 2 O 2 concentration. Cyclic voltammetry (CV) and amperometric measurements were performed using a PARSTAT 4000 electrochemical workstation (AME-TEK, Princeton, NJ, USA). All electrochemical measurements were carried out in a typical three-electrode system with saturated calomel electrode (SCE) as the reference electrode, platinum (Pt) disk electrode as the counter electrode and modified glassy carbon electrode (GCE) as the working electrode.

Fabrication of PCN-224 and Fe@PCN-224
PCN-224 nanoparticles were manufactured according to the previously reported procedure [48]. In this experiment, 50 mg of H 2 TCPP, 150 mg of ZrOCl 2 ·8H 2 O, and 1.4 g of benzoic acid were first dissolved in 50 mL of DMF. The solution was heated evenly at 90 • C for 5 h. After the reaction was completed, PCN-224 nanoparticles were collected by centrifugation and then washed three times with fresh DMF.
As for Fe@PCN-224, 60 mg of PCN-224 and 80 mg of FeCl 3 were dispersed in 20 mL DMF. The solution was stirred for 30 min at room temperature and then heated at 120 • C under stirring (300 rpm) for 8 h. Finally, the Fe@PCN-224 were obtained by centrifugation and washed three times with DMF and were stored in fresh DMF for further analysis.

Preparation of Fe@PCN-224-Modified Electrodes
The above Fe@PCN-224 was washed three times with water and 100 µL 1wt% Nafion were mixed under ultrasonication for 1 h, so the Fe@PCN-224/Nafion composites were obtained. Before fabrication of the electrodes, the treatments of bare glassy carbon electrodes were required. Glassy carbon electrode was first polished with 1.5 µm, 0.5 µm, and 50 nm alumina slurries to create a mirror finish, and then sonicated with ultra-pure water, 1:1 nitric acid, ethanol, and ultra-pure water successively. After the glassy carbon electrode was dried by nitrogen gas, 5 µL of Fe@PCN-224/Nafion suspension was coated on the glassy carbon electrode surface, forming Fe@PCN-224/Nafion/GCE upon drying overnight under room temperature.

Sample Treatment for H 2 O 2 Determination
Preparation of fresh fishery products: 2 g muscle samples of the fresh fishery products were chopped and extracted with 30 mL 0.1 M phosphate buffer. The samples were obtained by centrifugation (6000 rpm, 15 min) and filtration with a 0.45 µm filter membrane.
Preparation of some samples with H 2 O 2 to simulate illegal treatment: 2 g muscle samples of the fresh fishery products were completely immersed in 10 mL H 2 O 2 solution (0.8%) for 2 min [7]. The samples were rinsed 3 times with 50 mL of fresh water after removal of the liquid. The same method as above was used to extract the treated fishery products.

Electrochemical Determination of H 2 O 2
All electrochemical measurements were carried out in a three-electrode electrochemical cell at room temperature of 25 • C. CVs were obtained with a potential window of −1.5-1.5 V at a scan rate of 50 mV/s in 0.1 M phosphate buffer (pH 7.0). All amperometric measurements were carried out at an applied potential of 1.0 V in 0.1 M phosphate buffer (pH 7.0) without specific description, requiring the transient background to decrease to a steady-state value. Magnetic stirring was applied to the solution during amperometric measurements to maintain convective mass transfer characteristics.

Spectrophotometric Determination of H 2 O 2
We used the photometrical method described by Bader et al. [49], which was slightly modified by Drabkova et al. [50]. The buffer stock solution was prepared by mixing 1.2 mL 0.5 M phosphate buffer (pH 7.0) and 10.8 mL sample. 20 µL of DPD reagent (0.1 g DPD diluted in 10 mL 0.5 M H 2 SO 4 solution) and 20 µL of HRP reagent (10 mg HRP diluted in 10 mL water) were added into the buffered sample, while continually stirred. The developed color was measured at a wavelength of 551 nm. The absorbance of the whole mixture without HRP addition was measured as a blank. The concentration of hydrogen peroxide was calculated according to the following equation: (1) in which ∆A 551 is absorbance after subtracting the value of the blank, V final is the final volume of the measured mixture, ε is 21,000 M −1 cm −1 , L is the length of the optical cell, and V sample is the volume of the original sample.

Characterization of Composites
The preparation of Fe@PCN-224/Nafion/GCE was a four-step process, as presented in Figure 1. First, prior to the preparation of Fe@PCN-224, PCN-224 was synthesized using typical methods [48]. Due to the strong interaction between Zr 6 and the carboxyl group, PCN-224 has super stability. Second, free Fe III ions were incorporated into the center of the porphyrin unit. The obtained Fe@PCN-224 owns high open metal site density for electrochemical applications due to the special structure of MOFs [51]. Then, by mixing the Fe@PCN-224 particles in Nafion solution in a fast ultrasonic process, the well-mixed Fe@PCN-224/Nafion composites were obtained. Nafion, as a proton-conducting membrane in electrochemical sensors, has the ability to block the anionic oxidant and reductant, which is expected to avoid interference in real samples [52]. Finally, Fe@PCN-224/Nafion suspension was coated on the GCE surface, forming Fe@PCN-224/Nafion/GCE upon drying. Surface structure and response properties were obtained by subsequent characterization analysis and electrochemical experiments.   Figure S1 shows corresponding elemental mapping images of Fe@PCN-224/Nafion. Through element map identification, Fe III was successfully incorporated into PCN-224, and the Fe@PCN-224 was evenly mixed with Nafion.  Figure S1 shows corresponding elemental mapping images of Fe@PCN-224/Nafion. Through element map identification, Fe III was successfully incorporated into PCN-224, and the Fe@PCN-224 was evenly mixed with Nafion.

Cyclic Voltammetry of the H2O2 Sensor
The sensing properties of Fe@PCN-224/Nafion/GCE for the electrochemical detection of H2O2 were studied preliminarily using cyclic voltammetry (CV). Figure 3a shows the CVs of the bare GCE, the Nafion/GCE, the Fe@PCN-224/GCE and the Fe@PCN-224/Nafion/GCE with 2 mM H2O2 at a scan rate of 50 mV s -1 . For the bare GCE and Nafion-

Cyclic Voltammetry of the H 2 O 2 Sensor
The sensing properties of Fe@PCN-224/Nafion/GCE for the electrochemical detection of H 2 O 2 were studied preliminarily using cyclic voltammetry (CV). Figure 3a shows the CVs of the bare GCE, the Nafion/GCE, the Fe@PCN-224/GCE and the Fe@PCN-224/Nafion/GCE with 2 mM H 2 O 2 at a scan rate of 50 mV s −1 . For the bare GCE and Nafion-coated electrodes, a very weak oxidation peak at about 1.2 V is observed, which suggests that GCE and Nafion have no catalytic effects on the reaction of H 2 O 2 . A similar phenomenon was observed in the H 2 O 2 sensor (Cu-TDPAT/GCE) developed by Zhang et al. [53], which was attributed to the slow electron transfer kinetics of the H 2 O 2 oxidation process. In contrast, CVs of Fe@PCN-224/GCE and the Fe@PCN-224/Nafion/GCE exhibit a remarkable oxidation current peak at about 1.1 V, which are approximately 4.7 and 2.5 times higher than the Nafion/GCE, respectively, indicating that Fe@PCN-224 has an efficient electrocatalytic activity for H 2 O 2 oxidation. Comparing the CVs of Fe@PCN-224/Nafion/GCE and Fe@PCN-224/GCE, the oxidation current of Fe@PCN-224/Nafion/GCE is almost 1.9 times higher than that of Fe@PCN-224/GCE, which is rational because of better conductivity caused by Nafion. Moreover, the synergistic effect of Nafion and Fe@PCN-224 could result in an amplified oxidation current. The CV of Fe@PCN-224/Nafion/GCE in blank solution, shown in Figure 3b, exhibits one pair of redox peaks with extremely weak oxidation currents. After adding 2 mM H 2 O 2 in solution, a pair of enhanced redox peaks are observed. The oxidation peak current increases 6.8 times higher than in the blank solution, demonstrating appreciable electrocatalytic activity of the Fe@PCN-224/Nafion/GCE toward H 2 O 2 oxidation.  Fe@PCN-224 shows nearly no emission peak by TA (terephthalic acid) probing method, which demonstrates that Fe@PCN-224 does not produce • OH [48]. TA could react with • OH to form TA-OH, which is fluorescent at an excitation wavelength of 315 nm [54,55]. Some researchers suggest that the Fenton reaction produces not only • OH but also the ferryl ion (Fe 4+ =O), which is dependent on the nature of the chelator [56]. Fe@PCN-224/Nafion probably produces ferryl ion in the presence of H2O2 to exhibit peroxidase-like Fe@PCN-224 shows nearly no emission peak by TA (terephthalic acid) probing method, which demonstrates that Fe@PCN-224 does not produce • OH [48]. TA could react with • OH to form TA-OH, which is fluorescent at an excitation wavelength of 315 nm [54,55]. Some researchers suggest that the Fenton reaction produces not only • OH but also the ferryl ion (Fe 4+ =O), which is dependent on the nature of the chelator [56]. Fe@PCN-224/Nafion probably produces ferryl ion in the presence of H 2 O 2 to exhibit peroxidase-like activity [48,57,58], which needs further studies to demonstrate. Based on the above results, the pertinent reaction mechanism could be proposed as two procedures: In the first step, the catalytic center Fe III @PCN-224/Nafion is oxidized electrochemically to Fe IV =O@PCN-224/Nafion. The second procedure is the progress of chemical recognition. H 2 O 2 can be absorbed to the pores and surfaces of Fe IV =O@PCN-224/Nafion, then Fe IV =O@PCN-224/Nafion reacts simultaneously with H 2 O 2 and is reduced to Fe III @PCN-224/Nafion. H 2 O 2 loses electrons and is oxidized to produce oxygen. The reaction mechanism could be described as follows: The effect of the scan rate versus the current in 2 mM H 2 O 2 solution was detected. Referring to Figure 3c, the oxidation peak current (I pa ) increases with the scan rate (v) in the range of 40 to 400 mV s −1 . There is a good linear relationship between I pa and the square root of v with R 2 = 0.999 (Figure 3d). The relationship could be expressed as I pa (µA) = 0.911+7.926v 1/2 (mV s −1 ), which indicates that the electrochemical reaction of H 2 O 2 on Fe@PCN-224/Nafion/GCE is a diffusion-controlled irreversible process in the investigated potential range.

Amperometric Measurement of H 2 O 2
As shown in Figure S2, the oxidation current response of the sensor gradually increases with the increase of applied potential (0.8-1.0 V) and decreases after this (1.0-1.2 V), reaching a maximum value at 1.0 V. Therefore, 1.0 V was selected as the optimal applied potential in subsequent measurements. Typical current-time dynamic response of the Fe@PCN-224/Nafion/GCE towards H 2 O 2 is shown in Figure 4a. The electrode responds quickly to the change of H 2 O 2 concentration. The current is stable within 10 s after adding different concentrations of H 2 O 2 . The linear plot of H 2 O 2 concentration versus amperometric currents demonstrates two corresponding linear regions of 2 to 1500 µM and 1500 to 13,000 µM, which covers four orders of magnitude of H 2 O 2 concentrations. As illustrated in Figure 4b, the corresponding calibration curve in range from 2 to 1500 µM exhibits regression equation I pa (µA) = (0.05 ± 0.01) + (4.37 ± 0.03) C (mM), R 2 = 0.999. In range from 1500 to 13,000 µM, the corresponding calibration curve could be expressed as regression equation I pa (µA) = (5.34 ± 0.25) + (1.75 ± 0.03) C (mM), R 2 = 0.993. The detection limit is 0.7 µM with a signal-to-noise ratio of three (S/N = 3).
A comparison of linear range and detection limit for Fe@PCN-224/Nafion/GCE with other H 2 O 2 sensors reported in the literature is shown in Table 1. The proposed electrode has a wider range than traditional sensors, especially some horseradish peroxidase sensors. The wider linear range, including four orders of magnitude, allows the electrode to monitor a broader range of H 2 O 2 concentrations. It can be seen that Fe@PCN-224/Nafion/GCE is able to present satisfactory sensing performance with a wide linear range and a comparable detection limit.
The good performance of the Fe@PCN-224/Nafion/GCE may be attributed to two main reasons. First of all, Nafion can block the anionic oxidant and reductant, which helps to attenuate their interference and extend the service life of the H 2 O 2 sensor. Second, Fe@PCN-224 is highly porous and provides a microenvironment for H 2  ric currents demonstrates two corresponding linear regions of 2 to 1500 μM and 1500 to 13,000 μM, which covers four orders of magnitude of H2O2 concentrations. As illustrated in Figure 4b, the corresponding calibration curve in range from 2 to 1500 μM exhibits regression equation Ipa (μA) = (0.05 ± 0.01) + (4.37 ± 0.03) C (mM), R 2 = 0.999. In range from 1500 to 13,000 μM, the corresponding calibration curve could be expressed as regression equation Ipa (μA) = (5.34 ± 0.25) + (1.75 ± 0.03) C (mM), R 2 = 0.993. The detection limit is 0.7 μM with a signal-to-noise ratio of three (S/N = 3). A comparison of linear range and detection limit for Fe@PCN-224/Nafion/GCE with other H2O2 sensors reported in the literature is shown in Table 1. The proposed electrode has a wider range than traditional sensors, especially some horseradish peroxidase sensors. The wider linear range, including four orders of magnitude, allows the electrode to monitor a broader range of H2O2 concentrations. It can be seen that Fe@PCN-224/Nafion/GCE is able to present satisfactory sensing performance with a wide linear range and a comparable detection limit.   However, fluctuation could be observed during the detection process, which influences the detection limit of the electrode. The possible reason for the fluctuation of Fe@PCN-224/Nafion/GCE could be related to the relatively poor conductivity of the material. On one hand, the intrinsic insulating characteristics of the carboxyl groups utilized to form MOFs results in a low electrical conductivity, and electrons are obstructed from migrating along or accessing the skeleton of MOFs material. On the other hand, MOFs with micro size often have poor contact with the smooth surface of the electrode, making it difficult to transfer interfacial electrons from MOFs to the electrode surface.

Selectivity, Stability, and Reproducibility
Selectivity. Investigations of the selectivity of the Fe@PCN-224/Nafion/GCE to potential interferents were essential for practical applications. The public interferents of common H 2 O 2 electrodes were chosen. Biological samples often contain electroactive reducing agents, which produce corresponding oxidation currents during the detection of H 2 O 2 and seriously interfere with the determination. Figure 5a shows the response curve of Fe@PCN-224/Nafion/GCE to H 2 O 2 , Glc (glucose), L-Tyr (L-tyrosine), L-Phe (L-Phenylalanine), AA (Ascorbic Acid), H 2 C 2 O 4 , KH 2 PO 4 , MgSO 4 , Na 2 HPO 4 , CaCl 2 and Li 2 CO 3 . After adding 100 µM H 2 O 2 , an obvious current response could be observed. The current does not change significantly with the subsequent addition of ten 100 µM interfering species, indicating a good selectivity for H 2 O 2 sensing. Therefore, the high selectivity of the sensor makes it a potential candidate for H 2 O 2 determination in complex media. agents, which produce corresponding oxidation currents during the detection of H2O2 and seriously interfere with the determination. Figure 5a shows the response curve of Fe@PCN-224/Nafion/GCE to H2O2, Glc (glucose), L-Tyr (L-tyrosine), L-Phe (L-Phenylalanine), AA (Ascorbic Acid), H2C2O4, KH2PO4, MgSO4, Na2HPO4, CaCl2 and Li2CO3. After adding 100 μM H2O2, an obvious current response could be observed. The current does not change significantly with the subsequent addition of ten 100 μM interfering species, indicating a good selectivity for H2O2 sensing. Therefore, the high selectivity of the sensor makes it a potential candidate for H2O2 determination in complex media.  Stability. The current-time curve was continuously recorded to examine the stability of the modified electrode. As shown in Figure 5b, the current signal nearly remains unchanged for a long period of 2300 s, which suggests good stability of the sensor. Longterm measurement was also performed to confirm electrode stability. After storing the electrode in the air for 30 days, the current response decreases to 96.6% of the original response, indicating its good long-term stability.
Reproducibility. The reproducibility of the Fe@PCN-224/Nafion/GCE sensors prepared in different batches was also explored. Within seven Fe@PCN-224/Nafion/GCE sensors prepared, all of them can give a stable response in 100 µM H 2 O 2 solution. Five sensors exhibit an average of 0.49 µA response current with a relative standard deviation (RSD) of 3.83%. However, two sensors exhibit a much lower measuring current compared to the other five, which may be related to the base GCE electrode. Further research should be carried out to find it.

Application of the H 2 O 2 Sensor in Real Samples
As a cheap and effective preservative and bleaching agent, H 2 O 2 is generally used excessively by some illegal vendors. Superfluous H 2 O 2 residues in seafood and other foods pose a huge threat to consumer health. In order to testify the feasibility of the proposed Fe@PCN-224/Nafion/GCE for practical analysis, it was used to measure the accuracy of H 2 O 2 concentration in Todarodes pacificus, Larimichthys polyactis and Pennahia argentata. We added the above three kinds of chopped fresh fishery products samples to the buffer and filtered them, then used the standard addition method to make calibration curves of H 2 O 2 concentrations and amperometric currents, respectively. Figure S3 was the typical currenttime dynamic curves and linear relationships, suggesting that Fe@PCN-224/Nafion/GCE has a good linear relationship with H 2 O 2 in the range of 10-1500 µM in different fishery products samples. Then the concentration of H 2 O 2 in illegally treated fishery products samples could be calculated by substituting the measured current values into the standard curves. A photometrical method was also applied to measure the H 2 O 2 concentrations (considered as true values), which could help to assure the accuracy of H 2 O 2 quantitative detection by the sensor. As shown in Table 2, the H 2 O 2 concentrations detected by the electrochemical method are about 7% lower than those detected by the photometrical method. The difference between the two methods may be ascribed to the applied constant voltage in the electrochemical detection process. H 2 O 2 could react with fishery products samples to bleach and prevent corrosion, thereby reducing the concentration of H 2 O 2 remaining in the solution. During the electrode measurement, the applied constant voltage may accelerate the reaction of residual H 2 O 2 and samples, which results in a decrease in H 2 O 2 concentration. Consequently, the measured concentration by the electrochemical method is lower than that measured by the photometrical method. In general, the accuracies of different fishery product samples are between 91% and 95% in Table 2. The comparable deviations indicate that this electrode could effectively detect H 2 O 2 and resist the interference in real sample analysis. In addition to the advantage of fast and facile H 2 O 2 detection, other chemical reagents are not required for using the prepared nanohybrid-modified electrode. Therefore, Fe@PCN-224/Nafion/GCE is expected to be used in real sample research such as seafood.

Conclusions
In summary, Fe@PCN-224 with excellent performance was successfully fabricated by incorporating FeIII into the center of PCN-224, and it was applied to fabricate a novel electrochemical sensor for the determination of H 2 O 2 concentration. The sensor shows a high electrocatalytic ability to H 2 O 2 oxidation in a wide linear range and exhibits outstanding anti-interference ability, splendid stability. The prepared nanohybrid-modified electrode can be used to determine H 2 O 2 concentration in three kinds of fresh fishery products samples. In addition to the advantage of rapidness and briefness, simplicity of the device and easy application of Fe@PCN-224/Nafion/GCE open up new opportunities for in situ H 2 O 2 detection in foods.