A Paper-Based Electrochemical Sensor Based on PtNP/COFTFPB−DHzDS@rGO for Sensitive Detection of Furazolidone

Herein, a paper-based electrochemical sensor based on PtNP/COFTFPB−DHzDS@rGO was developed for the sensitive detection of furazolidone. A cluster-like covalent organic framework (COFTFPB−DHzDS) was successfully grown on the surface of amino-functional reduced graphene oxide (rGO-NH2) to avoid serious self-aggregation, which was further loaded with platinum nanoparticles (PtNPs) with high catalytic activity as nanozyme to obtain PtNP/COFTFPB−DHzDS@rGO nanocomposites. The morphology of PtNP/COFTFPB−DHzDS@rGO nanocomposites was characterized, and the results showed that the smooth rGO surface became extremely rough after the modification of COFTFPB−DHzDS. Meanwhile, ultra-small PtNPs with sizes of around 1 nm were precisely anchored on COFTFPB−DHzDS to maintain their excellent catalytic activity. The conventional electrodes were used to detect furazolidone and showed a detection limit as low as 5 nM and a linear range from 15 nM to 110 μM. In contrast, the detection limit for the paper-based electrode was 0.23 μM, and the linear range was 0.69–110 μM. The results showed that the paper-based electrode can be used to detect furazolidone. This sensor is a potential candidate for the detection of furazolidone residue in human serum and fish samples.


Introduction
Antibiotics are widely used in medicine and aquaculture. However, the harm caused by the use of antibiotics is irreversible. In addition, antibiotic drugs are harmful to the body's immune system, the liver, and the kidneys. The abuse of antibiotics is extraordinarily serious, and about 700,000 people die directly or indirectly every year. Furazolidone is a nitro-containing antibiotic drug, which has been widely used in aquaculture and veterinary medicine to minimize the acute effects of Escherichia coli, Shigella, Salmonella, and other infections [1][2][3]. Owing to the accumulation of furazolidone through the food chain, it is detrimental to the human body's immune system and even causes mutation and carcinogenesis [4,5]. Therefore, it is urgent to develop a rapid method to realize the ultrasensitive detection of furazolidone.
Covalent organic frameworks (COFs) are crystalline organic porous polymers assembled by covalent bonding of C, N, O, and other light elements, whose structure can be regulated by the periodic arrangement of organic structural units [6,7]. COFs have stable networks and open channels, which are extraordinarily conducive to the transport of electrolyte ions and have great potential in various applications. It has been reported that some COFs or COF-based sensors have better performance [8][9][10]. However, COFs are highly conjugated, and the π-π interaction between layers would lead to a decrease in specific surface area and poor electrical conductivity.
The development of COF composite materials to enhance electrical conductivity and specific surface area has been proven to be a feasible route. The sp 2 hybridization of reduced graphene oxide (rGO) and its extremely thin atomic thickness enable rGO to be used in many fields [11][12][13][14]. The rGO-based composites are beneficial to improve the overall conductivity of the composites. COFs can be covalently connected with rGO-NH 2 to improve the electrical conductivity of the composites. In addition, COFs are also ideal nanospacers that can reduce the stacking of adjacent rGO nanosheets and improve ion transport [15][16][17]. Furthermore, the presence of numerous chelating sites in the COF structure could be employed to immobilize metal nanoparticles through coordination effects. The precise anchoring of nanoparticles in the ordered holes of COFs can also avoid self-aggregation [18].
Although chromatography, immunoassay, and fluorometry exhibit high sensitivity and accuracy in the detection of antibiotic residues in water and meat products, their limitations are the complex and time-consuming sample-handling procedures, and the corresponding expensive instruments and equipment. The electrochemical method is not only simple to operate but also has high sensitivity, accuracy, and portability of equipment, which effectively overcomes the above shortcomings [19]. A paper-based analytical device was developed by Whitesides for the first time. Electrochemical paper-based analytical devices usually consist of a three-electrode setup integrated into a paper substrate, offering several benefits, such as reduced consumption of reagents and samples, portability, low cost, and availability of the raw material. They are widely used in various fields [20][21][22][23][24].
In this work, rGO was converted into rGO-NH 2 through amino functionalization, which was employed as the substrate material. Then, COF TFPB−DHzDS with abundant chelating sites were prepared through the amine-aldehyde condensation reaction of 2, 5-bis (3-(ethyl thiol) propoxy) p-benzoyl hydrazine and 1,3,5-tri(p-formylphenyl) benzene. The -NH 2 is beneficial for the COF TFPB−DHzDS to grow on rGO-NH 2 uniformly to form COF TFPB−DHzDS @rGO. COF TFPB−DHzDS has abundant N and S atoms; therefore, Pt 4+ can be precisely doped in its surface and internal structure through coordination and adsorption. Lastly, PtNP/COF TFPB−DHzDS @rGO was prepared by using an in situ reduction method. The PtNP/COF TFPB−DHzDS @rGO exposes tremendous catalytic activity. A paper-based electrochemical device based on PtNP/COF TFPB−DHzDS @rGO composites on flexible peeled graphite paper for the detection of furazolidone was proposed. It can effectively avoid the time-consuming polishing and cleaning work of the conventional electrode, and the test results showed good performance. The device was expected to be an all-in-one electrochemical platform for the detection of antibiotics. This work provides a reference for the design and fabrication of integrated paper-based electrodes and their application in electrochemical sensing.

Instruments
Transmission electron microscopy (TEM) was performed using A JEM-2010 (HR) instrument. Scanning electron microscopy (SEM) images were collected by using a HITACHI S-3400N instrument, and the breakdown voltage was set to 15 kV. X-ray powder diffraction (XRD) analysis was performed using a D/Max 2500 V/PC instrument with Cu Kα radiation from 2 • to 35 • at a scanning rate of 1 • /min. N 2 adsorption/desorption isotherm tests were carried out with Autosorb-iQ (Quantachrome) under 77 K. All electrochemical tests were performed at the CHI760D (Shanghai, China) Electrochemical Workstation in a conventional three-electrode system (platinum wire as an auxiliary electrode, saturated calomel (SCE) as a reference electrode, and different modified GCEs as working electrodes for routine testing, or homemade paper-based electrodes including graphite-like foam electrodes as working electrodes and counter-electrodes and Ag/AgCl as the reference electrode). Cyclic voltammetry (CVs) and differential pulse voltammetry (DPV) tests were carried out in 0.2 M static N 2 phosphate-buffered solutions. The control frequency range of electrochemical impedance (EIS) was 0.01-105 Hz, and the amplitude was 5 mV, using 5.0 mM Fe (CN) 6 3−/4− as the signal probe. The data plotted for the calibration curve are the averages of the five experiments, and the length of the error bar indicates the magnitude of the relative deviation. Preparation of rGO-NH 2 : First, 12 mg rGO was dissolved in a 10 mL anhydrous acetonitrile solution, and the solution was ultrasonic for 1 h. Then, 24 mg tetrafluoroborate diazonium salt and 329 mg tetrabutyltetrafluoroborate ammonium were added, and all the reagents were dissolved and stirred in the dark for 20 h. The solution was dried through centrifugation, and then 10 mg precipitate was added into a 40 mL ethanol-water solution (ethanol: water = 3:2). Next, 104 mg zinc powder, 2 mg ammonium chloride, and 2.3 mL glacial acetic acid were added to the solution, and the mixed solution was heated in 60 • C water bath for 3 h. After cooling down to room temperature, the rGO-NH 2 was obtained by alternate cleaning with ethanol and water and drying in a 60 • C oven.
Preparation of COF TFPB−DHzDS : Briefly, 23 mg TFPB and 8.1 mg DHzDS were dissolved in a 3 mL mixed solution composed of mesitylene and 1,4-dioxane in a ratio of 3:1. Then, 600 µL of 6 M acetic acid was added to the solution and reacted at 120 • C for 3 days. Finally, it was cleaned 5 times with THF and dried for 6 h in a 60 • coven.
Preparation of COF TFPB−DHzDS @rGO: The steps were similar to COF TFPB−DHzDS . Briefly, 1.62 mg TFPB and 4.6 mg DHzDS were dissolved in a 3 mL mixed solution, then 10 mg rGO-NH 2 was added, and the same concentration and volume of acetic acid solution were added after the mixture was evenly mixed. The follow-up procedure was the same as the preceding procedure.
Preparation of PtNP/COF TFPB−DHzDS @rGO: The dried 4 mg COF TFPB−DHzDS @rGO was dispersed into a 3 mL methanol solution, and then 2 mL 0.012 mM potassium chloride platinate solution was added and stirred at room temperature for 12 h. Next, 20 µL 0.25 M sodium borohydride solution was added, stirred for 6 h, and washed with ethanol and methylene chloride until the upper solution became colorless and transparent. Finally, the product was dried in a freeze-dryer for 4 h. The preparation process of PtNP/COF TFPB−DHzDS @rGO materials is shown in Scheme 1.

Preparation of Paper-Based Electrodes (ePADs)
First, the commercial carbon paper was cut into long strips of 3 mm in width and 3 cm in length by using a regular paper knife. Then, two sides of a white cardboard sheet 2 cm in width and 3 cm in length were painted with white nail polish. Next, three long carbon paper strips were pasted onto a side of the white cardboard with an interval of about 0.5 cm. The middle sections of the long carbon paper strips (about one-third of the long strips) were painted with white nail polish. The bottom sections of the long carbon paper strips were peeled off using acrylic transparent tape to obtain a new surface with graphite-like foam as electrodes (working electrode, reference electrode, and counter electrode). Then, the reference electrode was coated with a layer of conductive silver powder, and HCl was added dropwise to form AgCl/Ag. (Scheme 2). Scheme 2. Schematic illustration of the preparation of paper-based electrodes.

Preparation of Paper-Based Electrodes (ePADs)
First, the commercial carbon paper was cut into long strips of 3 mm in width and 3 cm in length by using a regular paper knife. Then, two sides of a white cardboard sheet 2 cm in width and 3 cm in length were painted with white nail polish. Next, three long carbon paper strips were pasted onto a side of the white cardboard with an interval of about 0.5 cm. The middle sections of the long carbon paper strips (about one-third of the long strips) were painted with white nail polish. The bottom sections of the long carbon paper strips were peeled off using acrylic transparent tape to obtain a new surface with graphite-like foam as electrodes (working electrode, reference electrode, and counter electrode). Then, the reference electrode was coated with a layer of conductive silver powder, and HCl was added dropwise to form AgCl/Ag. (Scheme 2).

Preparation of Paper-Based Electrodes (ePADs)
First, the commercial carbon paper was cut into long strips of 3 mm in width and 3 cm in length by using a regular paper knife. Then, two sides of a white cardboard sheet 2 cm in width and 3 cm in length were painted with white nail polish. Next, three long carbon paper strips were pasted onto a side of the white cardboard with an interval of about 0.5 cm. The middle sections of the long carbon paper strips (about one-third of the long strips) were painted with white nail polish. The bottom sections of the long carbon paper strips were peeled off using acrylic transparent tape to obtain a new surface with graphite-like foam as electrodes (working electrode, reference electrode, and counter electrode). Then, the reference electrode was coated with a layer of conductive silver powder, and HCl was added dropwise to form AgCl/Ag. (Scheme 2). Scheme 2. Schematic illustration of the preparation of paper-based electrodes. Scheme 2. Schematic illustration of the preparation of paper-based electrodes.

Preparation of PtNP/COF TFPB−DHzDS @rGO/ePAD
The procedures were similar to PtNP/COF TFPB−DHzDS @rGO/GCE. Briefly, 2 mg PtNP/COF TFPB−DHzDS @ rGO was dispersed in a 1 mL DMF solution, dissolved through sonication, then 10 µL suspensions were dropped on the lower 1/3 of the working electrodes of the paper-based electrodes. Finally, the electrode surface was dried under a tungsten lamp for 5 min.

Characterization of COF TFPB−DHzDs
The SEM image ( Figure S1a) showed that COF TFPB−DHzDs was filamentous and could stack into clusters under the interaction of the Van der Waals force. FTIR spectrum ( Figure S1b) demonstrated the successful synthesis of COF TFPB−DHzDs . The stretching vibration peak of C=N at 1620 cm −1 appeared in COF TFPB−DHzDs , proving the formation of an imine bond and confirming that COF TFPB−DHzDs was formed through an aminealdehyde condensation reaction between monomer TFPB and DHzDS [25]. The XRD pattern ( Figure S1c) showed strong diffraction peaks at 4.77 • and 26.3 • , which belong to (100) and (001) crystal planes, respectively, indicating that COF TFPB−DHzDs has a good crystal structure [26,27]. According to the N 2 adsorption and desorption isotherms ( Figure S1d) of COF TFPB−DHzDs , the specific surface area of COF TFPB−DHzDs was about 153.76 m 2 g −1 , and the average pore size was about 1.9 nm. Its specific surface area was large, suggesting it could be used as an excellent support material [28].

Characterization of PtNP/COF TFPB−DHzDS @rGO
The SEM ( Figure 1a) and TEM (Figure 1d) images showed that rGO was a large twodimensional slice with slight folds [29,30]. After the amino functionalization, COF TFPB−DHzDS could uniformly grow on the surface of rGO (Figure 1b). Compared with rGO ( Figure 1a,d), the surface of COF TFPB−DHzDS @rGO became rough. Further investigation of the surface morphology using TEM showed that the thickness of COF TFPB−DHzDS @rGO increased, compared with the rGO, confirming the synthesis of COF TFPB−DHzDS @rGO (Figure 1e). Figure 1c shows the SEM image of PtNP/COF TFPB−DHzDS @rGO, whose morphology was not significantly different from that of COF TFPB−DHzDS @rGO. Figure 1f shows that a large number of PtNPs with diameters of about 1 nm were uniformly loaded on COF TFPB−DHzDS @ rGO ( Figure S2). The formation of small PtNPs can be ascribed to the fact that Pt 4+ uniformly adsorbed on the interlaminar structure or channel of COF TFPB−DHzDS during the synthesis process because COF TFPB−DHzDS had double chelating sites of N or S atoms [31].   The FTIR spectrum (Figure 2a) showed that rGO was basically a straight line without obvious adsorption peaks. After the amino functionalization, it could be observed that the stretching vibration peak of -NH 2 appeared in the spectrum of around 3453 cm −1 , which strongly proved that rGO was successfully functionalized. From the spectrum of COF TFPB−DHzDS @rGO, it could be found that, in addition to the stretching vibration peak of -NH 2 at about 3442 cm −1 , there was also a vibration peak at 1630 cm −1 , which was attributed to C=N vibration peaks. The C=N vibration peaks were observed after the aldehyde condensation (Figure 2a), indicating that COF TFPB−DHzDS was grown on the surface of rGO-NH 2 successfully [32]. In the X-ray photoelectron spectrum (Figure 2b) of PtNP/COF TFPB−DHzDS @rGO, it could be observed that the material was composed of elements such as C, N, O, and Pt. As shown in Figure 2c, 72.45 eV and 75.80 eV corresponded to the peaks of Pt 0 4f 7/2 and Pt 0 4f 5/2 , respectively, which proved that the material contained PtNPs. In Figure 2d, a strong peak mainly appeared at 284.01 eV in the C 1s fine spectrum, which belonged to C−C or C=C, while the three strong peaks appearing at 397.98 eV, 399.06 eV, and 399.88 eV in Figure 3e corresponded to =NH, −NH− and −N=N−, respectively [33][34][35]. In the XPS fine spectrum of O 1s (Figure 2f), the peak at 530.70 eV corresponded to C−O, while the peak at 532.60 eV corresponded to C=O. In summary, taking advantage of the large specific surface area and good electrical conductivity of rGO, and the porous framework structure and double chelation sites of COF TFPB−DHzDS , Pt 4+ was in situ reduced through a chemical reduction method, and a large amount of small-sized PtNPs were uniformly loaded on COF TFPB−DHzDS @rGO. This work provides a new reference for the construction of composite materials.
Biosensors 2022, 12, x FOR PEER REVIEW 6 of 13 taking advantage of the large specific surface area and good electrical conductivity of rGO, and the porous framework structure and double chelation sites of COFTFPB−DHzDS, Pt 4+ was in situ reduced through a chemical reduction method, and a large amount of small-sized PtNPs were uniformly loaded on COFTFPB−DHzDS@rGO. This work provides a new reference for the construction of composite materials.    The bare GCE had a pair of reversible redox peaks with a peak-to-peak potential difference (ΔEp) of 85 mV. After the modification with COFTFPB−DHzDS, the peak current was obviously reduced, and ΔEp was increased to 118 mV, indicating that COFTFPB−DHzDS inhibited the electron transfer. The PtNP/COFTFPB−DHzDS@rGO-modified electrode had a larger peak current and smaller ΔEp (about 98 mV). This suggested that PtNPs and rGO enhanced the reversibility of the reaction and significantly improved its electrochemical performance. This result is mainly attributed to the excellent electrical conductivity of rGO and PtNPs. Figure 3b shows the EIS of COF (curve a), COFTFPB−DHzDS@rGO (curve b), and PtNP/COFTFPB−DHzDS@rGO. The results indicated that PtNP/COFTFPB−DHzDS@rGO has the smallest charge transfer resistance, similar to GCE. The CVs of PtNP/COFTFPB−DHzDS@rGO/GCE in a 0.1 M N2-statured phosphate-buffered solution (pH = 7.0) with 10 μM furazolidone showed an obvious reduction of furazolidone and an indistinct oxidation peak. According to a previous report [36], the redox mechanism is speculated to be that the nitrogroup contained in the structure of furazolidone is reduced under the synergistic catalysis of Pt and rGO. With the increase in the scanning rate, the peak current density of furazolidone increased (Figure 3c), and a good linear relationship was presented (Figure 3d), indicating that the reaction process was a typical surface control process [37,38].

Optimization of the Experimental Conditions
The amount of COFTFPB−DHzDS growing on the rGO surface, the amount of PtNPs immobilized on COFTFPB−DHzDS@rGO, the pH value of electrolyte solution, and the concentration of PtNP/COFTFPB−DHzDS@rGO dispersion had significant effects on the performance of the constructed electrochemical sensor, and accordingly, these conditions were optimized in order to realize the sensitive detection of furazolidone (Figure 3e-h). Figure 3e indicates that with the amount of COFTFPB−DHzDS on rGO, the current responses gradually increased. When the amount exceeded 10 μM, the current responses  The bare GCE had a pair of reversible redox peaks with a peak-to-peak potential difference (∆Ep) of 85 mV. After the modification with COF TFPB−DHzDS , the peak current was obviously reduced, and ∆Ep was increased to 118 mV, indicating that COF TFPB−DHzDS inhibited the electron transfer. The PtNP/COF TFPB−DHzDS @rGO-modified electrode had a larger peak current and smaller ∆Ep (about 98 mV). This suggested that PtNPs and rGO enhanced the reversibility of the reaction and significantly improved its electrochemical performance. This result is mainly attributed to the excellent electrical conductivity of rGO and PtNPs. Figure 3b shows the EIS of COF (curve a), COF TFPB−DHzDS @rGO (curve b), and PtNP/COF TFPB−DHzDS @rGO. The results indicated that PtNP/COF TFPB−DHzDS @rGO has the smallest charge transfer resistance, similar to GCE. The CVs of PtNP/COF TFPB−DHzDS @rGO/GCE in a 0.1 M N 2 -statured phosphate-buffered solution (pH = 7.0) with 10 µM furazolidone showed an obvious reduction of furazolidone and an indistinct oxidation peak. According to a previous report [36], the redox mechanism is speculated to be that the nitrogroup contained in the structure of furazolidone is reduced under the synergistic catalysis of Pt and rGO. With the increase in the scanning rate, the peak current density of furazolidone increased (Figure 3c), and a good linear relationship was presented (Figure 3d), indicating that the reaction process was a typical surface control process [37,38].

Optimization of the Experimental Conditions
The amount of COF TFPB−DHzDS growing on the rGO surface, the amount of PtNPs immobilized on COF TFPB−DHzDS @rGO, the pH value of electrolyte solution, and the concentration of PtNP/COF TFPB−DHzDS @rGO dispersion had significant effects on the performance of the constructed electrochemical sensor, and accordingly, these conditions were optimized in order to realize the sensitive detection of furazolidone (Figure 3e-h). Figure 3e indicates that with the amount of COF TFPB−DHzDS on rGO, the current responses gradually increased. When the amount exceeded 10 µM, the current responses decreased because COF TFPB−DHzDS seriously accumulated. The amount of COF TFPB−DHzDS @rGO was maintained at 3 mg, and the current response of PtNP/COF TFPB−DHzDS @rGO prepared at different concentrations of potassium chloroplatinate to furazolidone was investigated. It could be seen intuitively from Figure 3f that, when the concentration of potassium chloroplatinate was 12 µM, the resulting sensor had the optimum current response to furazolidone. When the concentration of potassium chloroplatinate was further increased to 16 µM, the peak current density of furazolidone decreased instead. The reason might be that the ion concentration was too high, thus leading to the larger size of the formed PtNPs with decreased catalytic ability. The detection of furazolidone in a 0.2 M phosphate-buffered solution with different pH levels by using PtNP/COF TFPB−DHzDS @rGO/GCE was studied, and the results are shown in Figure 3g. It could be seen that the peak current density value reached the maximum at pH = 7.0. Therefore, a phosphate-buffered solution with pH = 7.0 was used as the optimal electrolyte solution for the detection of furazolidone by using PtNP/COF TFPB−DHzDS @rGO/GCE. Finally, the amount of PtNP/COF TFPB−DHzDS @rGO modified on GCE was optimized, and the results are shown in Figure 3h. As the concentration of PtNP/COF TFPB−DHzDS @rGO increased, the catalytic current of furazolidone gradually increased and reached the maximum value at 2 mg mL −1 . When the concentration of PtNP/COF TFPB−DHzDS @rGO exceeded 2 mg mL −1 , the PtNP/COF TFPB−DHzDS @rGO accumulated on the surface of GCE, and the modified layer became thicker, which hindered the mass transfer, resulting in a decrease in the peak current density value. Therefore, 2 mg mL −1 of PtNP/COF TFPB−DHzDS @rGO was selected as the optimal modification concentration.

Electrochemical Sensors Based on PtNP/COF TFPB−DHzDS @rGO for Furazolidone
The DPV curves of GCE, COF TFPB−DHzDS /GCE, and PtNP/COF TFPB−DHzDS @rGO/GCE in a 0.2 M N 2 -saturated phosphate-buffered solution (pH = 7.0) with 30 µM furazolidone showed that the PtNP/COF TFPB−DHzDS @rGO/GCE exhibited excellent furazolidone reduction performance with a large reduction peak current and a positive reduction potential, thus providing a sensitive method for furazolidone determination (Figure 4a). The highly sensitive response of PtNP/COF TFPB−DHzDS @rGO/GCE to furazolidone was attributed to the high catalytic activity of PtNP/COF TFPB−DHzDS @rGO and the strong interaction between the PtNP/COF TFPB−DHzDS @rGO and furazolidone [39,40]. The high electrical conductivity of PtNP/COF TFPB−DHzDS @rGO accelerated the electron transfer to furazolidone during the redox process, and the rich aromatic system in COF TFPB−DHzDS facilitated π-π stacking interactions with conjugated molecules such as furazolidone. Furthermore, the large surface area of the PtNP/COF TFPB−DHzDS @rGO provided abundant sites for analyte binding and reduction. PtNPs as nanozymes catalyzed the reduction of furazolidone, and rGO was beneficial to improve the conductivity. Under the synergistic effect, the peak current became larger. Due to the existence of PtNPs, the overpotential of the reaction was reduced, making furazolidone easier to reduce, so the more positive the potential was, the more the peak potential shifted to the right.
Next, the current responses of PtNP/COF TFPB−DHzDS @rGO/GCE in furazolidone solutions with different concentrations were measured. Figure 4b shows the response of DPV to furazolidone with different concentrations. The results showed that the peak current response increased as the concentration of furazolidone increased, which maintained a linear relationship over a wide range (15.0 nM−110 µM). Figure 4c shows the corresponding linear relationship between the peak current density and the concentration of furazolidone. Each value was the average of the resulting values that were repeated five times. The linear equation was j = -84.6511c-3.2922 (R 2 = 0.99), where j and c were the peak current density and the concentration of furazolidone solution, respectively. The detection limit of PtNP/COF TFPB−DHzDS @rGO/GCE for furazolidone was 5.0 nM.
The current responses of PtNP/COF TFPB−DHzDS @rGO/ePAD in furazolidone solutions with different concentrations were measured. Figure 5b shows the response of the paper-based electrochemical sensor to the different concentrations of furazolidone. The results showed that the peak current response increased as the concentration of furazolidone increased, which maintained a linear relationship over a wide range (0.69 µM −100 µM). Figure 5c shows a linear relationship between the peak current density and the concentration of furazolidone. The linear equation was j = -14.24c-2170.91 (R 2 = 0.99), where j and c are the peak current density and the concentration of furazolidone solution, respectively. The detection limit of the paper-based electrochemical sensor was 0.23 µM. The paper-based electrodes had good linear ranges, which proved that the preparation of paper-based electrodes was successful. The bare ePAD had a pair of reversible redox peaks. After the modification of the target material, the peak current slightly decreased, and the impedance slightly increased. These experimental results were similar to those of GCE, which proved that the paper-based electrode was successfully prepared. The current responses of PtNP/COFTFPB−DHzDS@rGO/ePAD in furazolidone solutions with different concentrations were measured. Figure 5b shows the response of the paper-based electrochemical sensor to the different concentrations of furazolidone. The results showed that the peak current response increased as the concentration of furazolidone increased, which maintained a linear relationship over a wide range (0.69 μM −100 μM). Figure 5c shows a linear relationship between the peak current density and the concentration of furazolidone. The linear equation was j = -14.24c-2170.91 (R 2 = 0.99), where j and c are the peak current density and the concentration of furazolidone solution, respectively. The detection limit of the paper-based electrochemical sensor was 0.23 μM. The paper-based electrodes had good linear ranges, which proved that the preparation of paper-based electrodes was successful.

Determination of Furazolidone in Human Serum and Fish Sample
The diluted human serum was used as actual samples, and the results showed that the paper-based electrode sensor had a good recovery rate, which proved that the PtNP/COF TFPB−DHzDS@rGO/ePAD sensor has the potential to detect furazolidone in real examples (Table S1). In addition, fish were raised for three days and given different amounts of furazolidone in the water ( Figure S3). The water in the tank was used for the actual sample testing, and the results are shown in Table S2, which indicates that a part of furazolidone was absorbed by the fish.

Determination of Furazolidone in Human Serum and Fish Sample
The diluted human serum was used as actual samples, and the results showed that the paper-based electrode sensor had a good recovery rate, which proved that the PtNP/COF TFPB−DHzDS @rGO/ePAD sensor has the potential to detect furazolidone in real examples (Table S1). In addition, fish were raised for three days and given different amounts of furazolidone in the water ( Figure S3). The water in the tank was used for the actual sample testing, and the results are shown in Table S2, which indicates that a part of furazolidone was absorbed by the fish.

Conclusions
In conclusion, an electrochemical sensing platform based on a homemade paperbased electrode loaded with PtNP/COF TFPB−DHzDS @rGO composite was developed to detect furazolidone. The rGO−NH 2 was used to guide the growth of COF TFPB−DHzDS on its surface to prepare COF TFPB−DHzDS @rGO composites in which COF TFPB−DHzDS were covalently linked on rGO−NH 2 . Then, Pt 4+ was first coordinated with the N and S atoms of COF TFPB−DHzDS , and subsequently, PtNP/COF TFPB−DHzDS @rGO was obtained by reducing Pt 4+ . COF TFPB−DHzDS was uniformly distributed on the layered rGO, and the ultra-small PtNPs were formed on COF TFPB−DHzDS @rGO as nanozymes. The rGO and PtNPs increased the electrical conductivity of COF TFPB−DHzDS , and the catalytic activity of PtNP/COF TFPB−DHzDS @rGO was enhanced. Therefore, the proposed furazolidone electrochemical sensor based on PtNP/COF TFPB−DHzDS @rGO nanocomposites had a low detection limit (5 nM), a wide determination range (15 nM−110 µM) based on GCE, and good repeatability and stability. In contrast, the limit of detection for the paper-based electrode was 0.23 µM and the linear range was 0.69-100 µM. In addition, this work provides a reference for the covalent attachment of COFS materials on the surface of rGO and the synthesis of ultra-small-sized nanoparticles.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/bios12100904/s1. Figure S1. SEM image (a), FTIR spectrum (b), XRD pattern (c) and N2 adsorption/desorption iso-therm (d) of COFTFPB− DHzDS. Figure  S2. the particle size distribution of PtNPs in PtNPs/COFTFPB−DHzDS@rGO. Figure S3. Picture of fish with furazolidone. Table S1. The determination of furazolidone in human serum sample by PtNPs/COFTFPB−DHzDS@ rGO/PBE. Table S2. The determination of furazolidone in fish sample by PtNPs/COFTFPB−DHzDS@rGO/PBE. Author Contributions: R.C. and Y.D. conceived and planned the study. R.C. and X.P. carried out the experiments. R.C., Y.S. and X.P. reviewed and edited the manuscript. Y.D. was responsible for the supervision, project administration, and funding acquisition. All authors provided critical feedback and helped shape the research and analysis. All authors have read and agreed to the published version of the manuscript.