Mn3O4/NiO Nanoparticles Decorated on Carbon Nanofibers as an Enzyme-Free Electrochemical Sensor for Glucose Detection

Transition metal oxides have garnered a lot of attention in the field of electrocatalysis along with their unique crystal structure and excellent catalytic properties. In this study, carbon nanofibers (CNFs) decorated with Mn3O4/NiO nanoparticles were made using electrospinning and calcination. The conductive network constructed by CNFs not only facilitates electron transport, but also provides landing sites for nanoparticles, thus reducing nanoparticle aggregation and exposing more active sites. Additionally, the synergistic interaction between Mn3O4 and NiO improved electrocatalytic capacity for glucose oxidation. The Mn3O4/NiO/CNFs modified glassy carbon electrode shows satisfactory results in terms of linear range and anti-interference capability for glucose detection, suggesting that the constructed enzyme-free sensor has a promising application in clinical diagnosis.


Introduction
Electrochemical glucose sensors perform an indispensable role in monitoring glucose concentrations, which not only can be employed to control the intake of glucose in diet, but also to predict diseases associated with high or low glucose levels [1,2]. Enzyme-based sensors show excellent sensitivity and accuracy for glucose detection, but they are sensitive to temperature and pH which limits their storage and transportation [3]. In contrast, enzyme-free sensors have the advantage of being cheap, stable, and sensitive [4], thus attracting considerable attention in recent years.
Noble metals and their alloys have shown excellent catalytic properties and stability [5,6] and are often used as enzyme-free sensors for glucose detection [7,8]. However, the high cost and scarcity of resources limit wide applications of noble metals. In this regard, transition metal can overcome the drawbacks mentioned above for noble metals and be applied as catalysts for many reactions [9], such as CO 2 utilization [10], oxygen evolution reaction [11], and oxygen reduction reaction [12,13]. Generally speaking, composite transition metal oxides exhibit higher catalytic capacity and lower catalytic potential than single transition metal oxide due to the synergistic effect between different metal oxides. Wang et al. synthesized CuCo 2 O 4 /NiO nanoneedle arrays, modified on carbon cloth by hydrothermal method, which were used as self-supporting electrodes for the glucose detection [14]. Xu et al. fabricated CuO-NiO nanocomposites for the detection of glucose which exhibited excellent selectivity, reproducibility, and long-time stability [15].

Instrumentations
The scanning electron microscope (SEM, SU5000, HITACHI, Tokyo, Japan) was utilized to observe the morphology and size of materials. The high-resolution transmission electron microscope (HRTEM; JEM 2100F, JEOL, Tokyo, Japan) was further used to investigate the microscopic crystal structure of samples. X-ray diffraction (XRD, D/MAX2200, Rigaku, Tokyo, Japan) was employed to characterize the crystal structure of the materials. Fourier transform infrared spectroscopy (FTIR, AVATAR370, Thermo Scientific, Waltham, MA, USA) was performed to study chemical structure. The degree of graphitization of CNFs was studied by Raman spectrum (INVIA, Renishaw, London, UK). X-ray photoelectron spectrometer (XPS, ESCALAB-250XI, Thermo Scientific, Waltham, MA, USA) was applied to measure the surface components of materials in qualitative detail.

Synthesis of Mn 3 O 4 /NiO/CNFs
The synthesis process of Mn 3 O 4 /NiO/CNFs is illustrated in Scheme 1. Firstly, 0.5 g of PAN was mixed with 10 mL DMF and agitated to form a transparent solution. Then, 0.068 g Mn(Ac) 2 ·4H 2 O and 0.017 g Ni(Ac) 2 ·4H 2 O were added to the above solution and stirred continuously. Afterwards, the obtained brown precursor solution was moved to a 10 mL plastic injection syringe, and delivered to the metal needle at a constant stream rate of 0.6 mL h −1 under control of an injection pump. The grounded aluminum plate was used as the collector and kept away from the needle at a distance of 12 cm. After electrospinning, the collected precursor fibers were dried overnight at 60 • C under vacuum. The dried precursor fibers were pre-oxidized in air at 250 • C for 2 h, and then carbonized at 800 • C for another 2 h in an Ar stream. For comparison, Mn 3 O 4 /CNFs and NiO/CNFs were synthesized by the same method. The Raman spectrum was applied to evaluate the degree of graphitization of CNFs. In Figure 1b, the Raman spectrum shows a couple of significant peaks at 1350 and 1590

Electrochemical Measurements
The electrochemical measurements were carried out in a three-electrode system consisting of a platinum wire as counter electrode, Ag/AgCl (saturated KCl) as reference electrode, and Mn 3 O 4 /NiO/CNFs/GCE as working electrode. The three-electrode system was connected to an electrochemical workstation (CHI 842D, CH Instruments, Shanghai, China). The Mn 3 O 4 /NiO/CNFs/GCE was prepared in following steps: the bare GCE was polished with aluminum powder (0.5 µm and 0.01 µm), followed by cleaning with deionized water. Then, 5 µL of Mn 3 O 4 /NiO/CNFs uniform dispersion was dropped on GCE and then dried under infrared light to obtain Mn 3 O 4 /NiO/CNFs/GCE for glucose detection. The supporting electrolyte was 0.1 M NaOH.

XRD, Raman, FT-IR, and XPS Characterization of Mn 3 O 4 /NiO/CNFs
The crystal structure and purity of Mn 3 O 4 /NiO/CNFs were characterized with XRD. Figure

Electrochemical Measurements
The electrochemical measurements were carried out in a three-electrode system consisting of a platinum wire as counter electrode, Ag/AgCl (saturated KCl) as reference electrode, and Mn3O4/NiO/CNFs/GCE as working electrode. The three-electrode system was connected to an electrochemical workstation (CHI 842D, CH Instruments, Shanghai, China). The Mn3O4/NiO/CNFs/GCE was prepared in following steps: the bare GCE was polished with aluminum powder (0.5 μm and 0.01 μm), followed by cleaning with deionized water. Then, 5 μL of Mn3O4/NiO/CNFs uniform dispersion was dropped on GCE and then dried under infrared light to obtain Mn3O4/NiO/CNFs/GCE for glucose detection. The supporting electrolyte was 0.1 M NaOH.

XRD, Raman, FT-IR, and XPS Characterization of Mn3O4/NiO/CNFs
The crystal structure and purity of Mn3O4/NiO/CNFs were characterized with XRD. Figure 1a shows the XRD patterns of Mn3O4/CNF, NiO/CNF, and Mn3O4/NiO/CNF. The obtained material contained a set of diffraction peaks that matched with Mn3O4, NiO, and CNF. In addition, there were no other impure phases. The peak at 24.2° could be ascribed to the (200) plane of CNF. The diffraction peaks at 37.2°, 43.2°, 62.8°, and 75.4° displayed the crystal planes of (111), (200), (220), and (311) plane for NiO [22]. The diffraction peaks at 17.  The Raman spectrum was applied to evaluate the degree of graphitization of CNFs. In Figure 1b, the Raman spectrum shows a couple of significant peaks at 1350 and 1590 The Raman spectrum was applied to evaluate the degree of graphitization of CNFs. In Figure 1b, the Raman spectrum shows a couple of significant peaks at 1350 and 1590 cm −1 which were directly related to the D band and G band, respectively [24]. The intensity of the D band reflected the carbon defects, while the intensity of the G band gave information about the graphitic structure. The intensity ratio between the D and G band, R = I D /I G , can be used to reflect the degree of graphitization of CNFs [25]. The R value is 1.07, which means that there are still defects in the material under calcination at 800 • C. The explanation for the defect is that the N element in PAN has not been completely removed. As shown in Figure S1, an FT-IR spectrum of Mn 3 O 4 /NiO/CNFs was performed. The band at 3444 cm −1 is identified as the stretching vibration of O-H, which is related to the adsorption of water. The peak at 1620 cm −1 can correspond to the C=C bond. The weak peak observed at 1383 cm −1 can be attributed to the C-H bond. The band at about 1169 cm −1 is caused by C-O [26]. The peaks which appear at 618 cm −1 and 523 cm −1 can be traced back to Mn 3 O 4 and NiO, respectively [27].
The elemental composition and valence state of the Mn 3 O 4 /NiO/CNFs were reflected using XPS technique. The fully scanned spectrum ( Figure 2a) demonstrates that the synthesized materials contained elements of C, N, O, Ni, and Mn. The peaks located at 854.7 eV and 872.1 eV were associated with Ni 2p 3/2 and Ni 2p 1/2 , which were derived from Ni 2+ (Figure 2b). In addition, two satellite peaks accompanying the main peaks were observed at 861.5 eV and 879.8 eV, proving the presence of Ni 2+ from NiO [28]. The peaks located at 856.5 eV and 874.0 eV in Ni 2p spectrum corresponded to Ni 2p 3/2 and Ni 2p 1/2 from Ni 3+ , respectively. Here, more energy was required to form Ni 3+ than Ni 2+ , since the compensation of large amounts of charge leads to the introduction of Ni vacancies and oxygen gaps, which in turn cause the formation of Ni 3+ [29,30]. As shown in Figure 2c (Figure 2f), three peaks located at 398.8 eV, 400.1 eV, and 402.0 eV can be considered as pyridinic N, pyrrolic N, and graphitic N, respectively. Among them, pyridinic N and pyrrolic N are the two main doping forms in CNF [33].
Biosensors 2023, 13, x FOR PEER REVIEW 4 of 12 cm -1 which were directly related to the D band and G band, respectively [24]. The intensity of the D band reflected the carbon defects, while the intensity of the G band gave information about the graphitic structure. The intensity ratio between the D and G band, R = ID/IG, can be used to reflect the degree of graphitization of CNFs [25]. The R value is 1.07, which means that there are still defects in the material under calcination at 800 °C. The explanation for the defect is that the N element in PAN has not been completely removed. As shown in Figure S1, an FT-IR spectrum of Mn3O4/NiO/CNFs was performed. The band at 3444 cm -1 is identified as the stretching vibration of O-H, which is related to the adsorption of water. The peak at 1620 cm -1 can correspond to the C=C bond. The weak peak observed at 1383 cm -1 can be attributed to the C-H bond. The band at about 1169 cm -1 is caused by C-O [26]. The peaks which appear at 618 cm -1 and 523 cm -1 can be traced back to Mn3O4 and NiO, respectively [27]. The elemental composition and valence state of the Mn3O4/NiO/CNFs were reflected using XPS technique. The fully scanned spectrum (Figure 2a) demonstrates that the synthesized materials contained elements of C, N, O, Ni, and Mn. The peaks located at 854.7 eV and 872.1 eV were associated with Ni 2p3/2 and Ni 2p1/2, which were derived from Ni 2+ (Figure 2b). In addition, two satellite peaks accompanying the main peaks were observed at 861.5 eV and 879.8 eV, proving the presence of Ni 2+ from NiO [28]. The peaks located at 856.5 eV and 874.0 eV in Ni 2p spectrum corresponded to Ni 2p3/2 and Ni 2p1/2 from Ni 3+ , respectively. Here, more energy was required to form Ni 3+ than Ni 2+ , since the compensation of large amounts of charge leads to the introduction of Ni vacancies and oxygen gaps, which in turn cause the formation of Ni 3+ [29,30]. As shown in Figure 2c

SEM and TEM of Mn 3 O 4 /NiO/CNFs
The surface morphologies of Mn(Ac) 2 /Ni(Ac) 2 /PAN precursor nanofibers and Mn 3 O 4 /NiO/CNFs were observed through SEM. In Figure 3a, the surface of uncalcined precursor nanofibers was smooth due to the amorphous structure of Ni(Ac) 2 and Mn(Ac) 2 . After calcination at 800 • C, the nanoparticles were uniformly dispersed on the nanofiber surface (Figure 3b)

SEM and TEM of Mn3O4/NiO/CNFs
The surface morphologies of Mn(Ac)2/Ni(Ac)2/PAN precursor nanofibers an Mn3O4/NiO/CNFs were observed through SEM. In Figure 3a, the surface of uncalcine precursor nanofibers was smooth due to the amorphous structure of Ni(Ac)2 and Mn(Ac) After calcination at 800 °C, the nanoparticles were uniformly dispersed on the nanofibe surface (Figure 3b). The previous XRD result confirmed that the nanoparticle consisted o Mn3O4 and NiO. The diameter statistics and Gaussian distribution of the precursor nano fibers and Mn3O4/NiO/CNFs are shown in the insets of Figure 3a,b. Obviously, the averag diameter of the nanofibers decreased from 168 to 115 nm after calcination, a phenomeno which was ascribed to the reduction of huge amount of H, O, and N elements during th carbonization process. TEM and HRTEM were carried out to further reveal the micro scopic crystal structure of Mn3O4/NiO/CNF. As shown in Figure 3c

Electrochemical Characterization
The cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS were used to study the electrochemical performance of Mn3O4/NiO/CNFs. EIS tests of di ferent modified electrodes were conducted in the solution with 5 mM [Fe(CN)6] 3-/4-an 0.1 M KCl. The outcome of the EIS test is reflected by the Nyquist plot, where the semicir cular part of the high frequency region relates to the electron transfer restriction proces and the linear part of the low frequency region reflects the diffusion process. The decreas in the radius of the semicircle represents a decrease in the electron transfer resistance a the electrode interface (Ret). In Figure S2, compared to Mn3O4/CNFs/GCE an NiO/CNFs/GCE, Mn3O4/NiO/CNFs/GCE shows the smallest diameter, indicating the fastes charge transfer capability. To investigate the electrochemical performance o Mn3O4/NiO/CNFs, CV was employed. Figure 4a shows the CV curves of Mn3O4/CNFs/GCE NiO/CNFs/GCE, and Mn3O4/NiO/CNFs/GCE in 0.1 M NaOH solution with and without mM glucose. The oxidation current that changed for Mn3O4/CNFs/GCE in the presence o

Electrochemical Characterization
The cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were used to study the electrochemical performance of Mn 3 O 4 /NiO/CNFs. EIS tests of different modified electrodes were conducted in the solution with 5 mM [Fe(CN) 6 ] 3−/4− and 0.1 M KCl. The outcome of the EIS test is reflected by the Nyquist plot, where the semicircular part of the high frequency region relates to the electron transfer restriction process and the linear part of the low frequency region reflects the diffusion process. The decrease in the radius of the semicircle represents a decrease in the electron transfer resistance at the electrode interface (R et ). In Figure S2,  tween Ni 2+ and Ni 3+ . Notably, Mn3O4/NiO/CNFs/GCE exhibited the highest oxidation current. The mechanism of glucose oxidation on Mn3O4/NiO/CNFs can be explained by the following equations [36]: NiOOH + C6H12O6 → Ni(OH)2 + C6H10O6 (Gluconolactone) (3) Originally, in the alkaline medium, NiO strongly adsorbed water molecules to form Ni(OH)2 (Equation (1)). Next, Ni(OH)2 was oxidized to NiOOH as electron mediator (Equation (2)). Then, NiOOH obtained electrons from glucose to form Ni(OH)2, and glucose was oxidized to gluconolactone (Equation (3)). Previous studies have indicated that Mn-based materials have a wide range for enzyme-free detection of glucose [37,38]. In this work, by virtue of the synergistic interaction between Mn and Ni, the composite exhibits more significant electrocatalytic activity for the oxidation of glucose [39]. The results of the EIS also confirm that the introduction of Mn3O4 can increase the electrical conductivity of the composites.  Figure 4b describes the CV curves of Mn3O4/NiO/CNFs/GCE in different glucose concentrations (from 0 mM to 5 mM). The oxidation current signal increased with increasing glucose concentration, indicating that Mn3O4/NiO/CNFs are suitable as electrocatalytic media for the construction of glucose sensors. To further investigate the kinetic process of glucose oxidation on the electrode surface, the CV curves of Mn3O4/NiO/CNFs/GCE were tested in the range of scan rate from 20 mV s -1 to 320 mV s -1 (Figure 4c). As illustrated in Figure 4d, the currents of the anodic peak (Ipa) and cathodic peak (Ipc) maintained a good Originally, in the alkaline medium, NiO strongly adsorbed water molecules to form Ni(OH) 2 (Equation (1)). Next, Ni(OH) 2 was oxidized to NiOOH as electron mediator (Equation (2)). Then, NiOOH obtained electrons from glucose to form Ni(OH) 2 , and glucose was oxidized to gluconolactone (Equation (3)). Previous studies have indicated that Mn-based materials have a wide range for enzyme-free detection of glucose [37,38]. In this work, by virtue of the synergistic interaction between Mn and Ni, the composite exhibits more significant electrocatalytic activity for the oxidation of glucose [39]. The results of the EIS also confirm that the introduction of Mn 3 O 4 can increase the electrical conductivity of the composites. Figure 4b describes the CV curves of Mn 3 O 4 /NiO/CNFs/GCE in different glucose concentrations (from 0 mM to 5 mM). The oxidation current signal increased with increasing glucose concentration, indicating that Mn 3 O 4 /NiO/CNFs are suitable as electrocatalytic media for the construction of glucose sensors. To further investigate the kinetic process of glucose oxidation on the electrode surface, the CV curves of Mn 3 O 4 /NiO/CNFs/GCE were tested in the range of scan rate from 20 mV s −1 to 320 mV s −1 (Figure 4c). As illustrated in Figure 4d, the currents of the anodic peak (I pa ) and cathodic peak (I pc ) maintained a good linear connection with the square root of the scanning speed, indicating that the oxidation of glucose on Mn 3 O 4 /NiO/CNFs/GCE is a diffusion-controlled process.
To improve the electrocatalytic effect of Mn 3 O 4 /NiO/CNFs on glucose detection, the time-current method was used to optimize the parameters during the experiments, including the applied voltage and the concentrations of NaOH and Mn 3 O 4 /NiO/CNFs. Figure 5a shows that the response current increases with increasing applied voltage at the beginning, and then decreases when the applied voltage is higher than 0.50 V. Thus, 0.50 V is the optimum applied potential. In Figure 5b, the response current initially increases with increasing concentration of NaOH, and then starts to decrease when the concentration of NaOH exceeds 0.10 M. Therefore, 0.1 M NaOH was chosen as the optimum concentration of electrolyte. Figure 5c shows that the response current reaches its maximum when the concentration of Mn 3 O 4 /NiO/CNFs is 9 mg mL −1 ; therefore, 9 mg mL −1 was selected as the optimum modification material concentration. Figure 6a shows the current response of Mn 3 O 4 /NiO/CNFs/GCE for continuous addition of glucose under known optimal experimental conditions. The relationship between response current and glucose concentration shows linearly in the range of 5-3000 µM (R 2 = 0.997) and 3000-12,000 µM (R 2 = 0.993), and could be well-fitted to the equation: I (µA) = 27.33 C (mM) + 1.343 and I (µA) = 17.22 C (mM) + 35.015. The detection limit of Mn 3 O 4 /NiO/CNFs/GCE was calculated to be 0.73 µM (S/N = 3). The sensitivity was calculated to be 386.84 µA mM −1 cm −2 (5-3000 µM) and 243.74 µA mM −1 cm −2 (3000-12,000 µM). Additionally, the response time is a key factor in weighing up the practical application of the sensor. Figure S3 shows that the Mn 3 O 4 /NiO/CNFs/GCE at different glucose concentrations reaches its steady-state current within 2-4 s after the addition of glucose. Table 1 lists the electrocatalytic effects of some reported enzyme-free glucose sensors and enzymatic glucose sensors. Compared with them, our proposed Mn 3 O 4 /NiO/CNFs/GCE has a wider linear range as well as lower detection limit as glucose sensor. This work 3000 μM) and 243.74 μA mM -1 cm -2 (3000-12,000 μM). Additionally, the response time is a key factor in weighing up the practical application of the sensor. Figure S3 shows that the Mn3O4/NiO/CNFs/GCE at different glucose concentrations reaches its steady-state current within 2-4 s after the addition of glucose. Table 1 lists the electrocatalytic effects of some reported enzyme-free glucose sensors and enzymatic glucose sensors. Compared with them, our proposed Mn3O4/NiO/CNFs/GCE has a wider linear range as well as lower detection limit as glucose sensor.

Selectivity, Stability, Repeatability, and Reproducibility of Mn 3 O 4 /NiO/CNFs/GCE
The presence of other easily oxidized molecules in the actual sample may affect the accuracy of glucose detection. In Figure 6c, the addition of 5 µM of some interferences to 100 µM glucose solution at the ratio of 1:20 was investigated by a time-current test. The outcomes showed that the current responses caused by the addition of interferences were virtually negligible, indicating that Mn 3 O 4 /NiO/CNFs/GCE had good selectivity. Figure 6d shows the amperometric response of Mn 3 O 4 /NiO/CNFs/GCE to 100 µM glucose. The current kept stable during 2000 s, suggesting that the electrode had an excellent stability as glucose sensor. Besides, repeatability and reproducibility are also key to determine whether the synthesized materials can be used for actual sample detection. The five identical modified electrodes used to detect 0.1 mM glucose were found to have a relative standard deviation (RSD) of 6.38% ( Figure S4), indicating that the electrodes have good reproducibility. The modified electrode was repeated five times to detect 0.1 mM glucose, and the RSD was calculated to be 1.32% (Figure S5), meaning that the electrode has a good repeatability. To evaluate the long-term stability of the sensor, the Mn 3 O 4 /NiO/CNFs/GCE was stored at room temperature for 30 days. The results show that the response current remains at 95.3% of the initial current after 30 days ( Figure S6), indicating that our sensors have satisfactory long-term stability.

Actual Sample Analysis
To evaluate the practicality of Mn 3 O 4 /NiO/CNFs/GCE, serum samples with different concentrations of glucose were detected using standard addition method. The serum samples used in the experiment were from the School of Medicine, Shanghai University. The recoveries of the three samples were 100.09%, 100.20%, and 99.91%, respectively ( Table 2), indicating that the built sensor can be accepted for glucose detection in serum samples.

Conclusions
In this work, Mn 3 O 4 /NiO/CNFs were synthesized by electrospinning and following calcination, and were used as an enzyme-free sensor for glucose detection. The intricate network structure of CNFs provides the backbone for the nanoparticles and facilitates electron transport. In addition, the synergistic effect between Mn 3 O 4 and NiO is beneficial to improve the catalytic performance for glucose detection. The constructed Mn 3 O 4 /NiO/CNFs/GCE sensor has good linear range, low detection limit, and antiinterference ability, indicating that Mn 3 O 4 /NiO/CNFs have some application prospects as electrochemical sensors in clinical diagnosis.  Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Date will be made available on request.

Conflicts of Interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.