A Co3O4 Nanoparticle-Modified Screen-Printed Electrode Sensor for the Detection of Nitrate Ions in Aquaponic Systems

In this study, a screen-printed electrode (SPE) modified with cobalt oxide nanoparticles (Co3O4 NPs) was used to create an all-solid-state ion-selective electrode used as a potentiometric ion sensor for determining nitrate ion (NO3−) concentrations in aquaculture water. The effects of the Co3O4 NPs on the characterization parameters of the solid-contact nitrate ion-selective electrodes (SC-NO3−-ISEs) were investigated. The morphology, physical properties and analytical performance of the proposed NO3−-ion selective membrane (ISM)/Co3O4 NPs/SPEs were studied by X-ray diffraction (XRD), energy-dispersive spectroscopy (EDS), transmission electron microscopy (TEM), scanning electron microscopy (SEM), electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), potentiometric measurements, and potentiometric water layer tests. Once all conditions were optimized, it was confirmed that the screen-printed electrochemical sensor had high potential stability, anti-interference performance, good reproducibility, and no water layer formation between the selective membrane and the working electrode. The developed NO3−-ISM/Co3O4 NPs/SPE showed a Nernstian slope of −56.78 mV/decade for NO3− detection with a wide range of 10−7–10−2 M and a quick response time of 5.7 s. The sensors were successfully used to measure NO3− concentrations in aquaculture water. Therefore, the electrodes have potential for use in aquaponic nutrient solution applications with precise detection of NO3− in a complicated matrix and can easily be used to monitor other ions in aquaculture water.


Introduction
Aquaponics is a new type of circulating aquaculture technology that was created by combining two agricultural engineering technologies, i.e., aquaculture and hydroponics [1]. The ammonia nitrogen concentration in traditional aquaculture water, which is delivered to the hydroponic agriculture system of the aquaponic system, increases when fish excreta accumulate and gradually increases the toxicity of the aquaculture water. The ammonia nitrogen in water is decomposed by bacteria into nitrite, which is then decomposed into nitrate by nitrifying bacteria. Nitrate can be directly absorbed by plants and utilized as a metabolite [2], but widespread nitrate accumulation in various water resources is now considered one of the biggest threats to aquatic life worldwide. An increasing body of information has been curated to reveal the effects of nitrates on fish growth and health [3], which have shown that fish suffer negative effects in health, behavior, and physiology from nitrate exposure [4][5][6]. Therefore, detecting nitrate in aquaculture solutions is critical for improving the quality of fish health. In recent years, there have been numerous reports and reviews of sophisticated methods for nitrate determination, and researchers have created a number of analytical techniques for detecting nitrate from various sources. These methods include electrochemical detection [7,8], spectrophotometry [9,10], chromatography [11], and capillary electrophoresis [12]. Until now, spectrophotometry has remained the most

Preparation of Nitrate Ion-Selective Electrodes
The process for preparing the nitrate ISEs can be described as follows: the SPEs were activated in 0.1 M H 2 SO 4 by cyclic voltammetry before detection (a) and then left to dry for a few minutes (b). Three milligrams of Co 3 O 4 nanoparticles were ultrasonically dispersed in 1 mL of deionized water for 30 min to create a homogeneous dispersion. In total, three 3 µL aliquots of the Co 3 O 4 NP dispersion were drop-cast onto the SPE (c). Finally, two 5 µL aliquots of a solution containing 6% MTDDA-NO 3 − , 65% o-NOPE, and 29% PVC in 1.5 mL tetrahydrofuran (THF) [36] were drop-cast on top of the Co 3 O 4 NPs and left to dry overnight and form the two NO 3 − -ISM layers (d). All of the prepared NO 3 − -ISM/Co 3 O 4 NPs/SPEs were conditioned in 10 −2 M NaNO 3 for 24 h before use (e). The NO 3 − -ISM/Co 3 O 4 NPs/SPEs were fabricated as depicted in Figure 1.

Parameter Optimization for NO3 − -ISM/Co3O4 NPs/SPE Fabrication
In this study, the Co3O4 concentration was optimized to improve the detection performance of the proposed NO3 − -ISM/Co3O4 NPs/SPE, as shown in Figure 2. The optimization range of the Co3O4 concentration was from 1 mg/mL to 5 mg/mL. For extremely low concentrations of Co3O4, the amount of Co3O4 remaining after drying was very low. Additionally, Co3O4 with a similar dispersion grade could not be completely concentrated on the electrode surface. However, when the Co3O4 concentration was too high, the Co3O4 formed a thick layer on the electrode surface and reduced the electrode sensitivity. As such, the electrochemical properties resulting from Co3O4 modification in these two scenarios were not good enough and high levels of noise and low sensitivity resulted. Finally, we selected 3 mg/mL as the optimal Co3O4 concentration.

Parameter Optimization for NO 3 − -ISM/Co 3 O 4 NPs/SPE Fabrication
In this study, the Co 3 O 4 concentration was optimized to improve the detection performance of the proposed NO 3 − -ISM/Co 3 O 4 NPs/SPE, as shown in Figure 2. The optimization range of the Co 3 O 4 concentration was from 1 mg/mL to 5 mg/mL. For extremely low concentrations of Co 3 O 4 , the amount of Co 3 O 4 remaining after drying was very low. Additionally, Co 3 O 4 with a similar dispersion grade could not be completely concentrated on the electrode surface. However, when the Co 3 O 4 concentration was too high, the Co 3 O 4 formed a thick layer on the electrode surface and reduced the electrode sensitivity. As such, the electrochemical properties resulting from Co 3 O 4 modification in these two scenarios were not good enough and high levels of noise and low sensitivity resulted. Finally, we selected 3 mg/mL as the optimal Co 3 O 4 concentration.

Parameter Optimization for NO3 − -ISM/Co3O4 NPs/SPE Fabrication
In this study, the Co3O4 concentration was optimized to improve the detection performance of the proposed NO3 − -ISM/Co3O4 NPs/SPE, as shown in Figure 2. The optimization range of the Co3O4 concentration was from 1 mg/mL to 5 mg/mL. For extremely low concentrations of Co3O4, the amount of Co3O4 remaining after drying was very low. Additionally, Co3O4 with a similar dispersion grade could not be completely concentrated on the electrode surface. However, when the Co3O4 concentration was too high, the Co3O4 formed a thick layer on the electrode surface and reduced the electrode sensitivity. As such, the electrochemical properties resulting from Co3O4 modification in these two scenarios were not good enough and high levels of noise and low sensitivity resulted. Finally, we selected 3 mg/mL as the optimal Co3O4 concentration.   Figure 3 shows that the XRD pattern showed peaks at 2θ = 18 [24]. There were no peaks visible for additional impurity phases, demonstrating that the product was extremely pure. These results showed that Co 3 O 4 was effectively synthesized and possessed a pure spinel structure.  [24]. There were no peaks visible for additional impurity phases, demonstrating that the product was extremely pure. These results showed that Co3O4 was effectively synthesized and possessed a pure spinel structure. SEM images of the Co3O4 NPs are displayed in Figure 4a, b. The SEM results reveal the Co3O4 NP surface morphology, which was mostly homogenous and aggregative in nature after dispersion. This showed that the bulk state of the as-prepared Co3O4 NPs was separate [37]. This morphology is advantageous for the fabrication of electrodes for supercapacitor applications; in general, a supercapacitor requires a material with a large surface area [38]. However, SEM cannot be used to determine whether the Co3O4 existed as nanoparticles. The physical properties of the Co3O4 were determined using TEM analysis, as shown in Figure 4c, d. According to the TEM analysis, the synthesized nanoparticles were uniform, with an average particle size of approximately 12 nm. Figure 4d depicts the plane spacing of the nanoparticles and the presence of crystals with various orientations. Additionally, the distance between the fringes of two nearby crystals was measured. The measured planar distance of 0.28 nm was assigned to plane 220. These results were consistent with those previously reported [39,40]. The TEM results showed that this preparation method successfully produced small Co3O4 NPs.  Figure 4a,b. The SEM results reveal the Co 3 O 4 NP surface morphology, which was mostly homogenous and aggregative in nature after dispersion. This showed that the bulk state of the as-prepared Co 3 O 4 NPs was separate [37]. This morphology is advantageous for the fabrication of electrodes for supercapacitor applications; in general, a supercapacitor requires a material with a large surface area [38]. However, SEM cannot be used to determine whether the Co 3 O 4 existed as nanoparticles. The physical properties of the Co 3 O 4 were determined using TEM analysis, as shown in Figure 4c,d. According to the TEM analysis, the synthesized nanoparticles were uniform, with an average particle size of approximately 12 nm. Figure 4d depicts the plane spacing of the nanoparticles and the presence of crystals with various orientations. Additionally, the distance between the fringes of two nearby crystals was measured. The measured planar distance of 0.28 nm was assigned to plane 220. These results were consistent with those previously reported [39,40]. The TEM results showed that this preparation method successfully produced small Co 3 O 4 NPs. Figure 5 shows the EDS results for Co 3 O 4 . Figure 5b,c, and d show images for a selected area of the Co 3 O 4 NPs (Figure 5a). In Figure 5e, the elemental composition of the Co 3 O 4 NPs is clearly shown. This demonstrated that only cobalt and oxygen components were present in the manufactured specimen with no other differentiated components, indicating that the synthesized Co 3 O 4 NPs were extremely clean [25].    Figure 5b, c, and d show images for a selected area of the Co3O4 NPs ( Figure 5a). In Figure 5e, the elemental composition of the Co3O4 NPs is clearly shown. This demonstrated that only cobalt and oxygen components were present in the manufactured specimen with no other differentiated components, indicating that the synthesized Co3O4 NPs were extremely clean [25].    Figure 5b, c, and d show images for a selected area of the Co3O4 NPs ( Figure 5a). In Figure 5e, the elemental composition of the Co3O4 NPs is clearly shown. This demonstrated that only cobalt and oxygen components were present in the manufactured specimen with no other differentiated components, indicating that the synthesized Co3O4 NPs were extremely clean [25].

Electrochemical Analysis of the Proposed Sensors
Cyclic voltammetry (CV) was performed in a solution of 5 mM K 3 [Fe(CN)6] 3−/4− containing 0.1 M KCl. In each case, three cycles were performed over the potential range −0.1-0.6 V at a scan rate of 0.05 V s −1 . Figure 6a shows the different CVs of the SPE and Co 3 O 4 NPs/SPE. After modification, the redox current of the Co 3 O 4 NPs/SPE was 288.8 µA (blue line), higher than that of the bare SPE at 285.2 µA (red line). The good dispersion of Co 3 O 4 ensured adhesion to the electrode surface with good stability. Therefore, Co 3 O 4 NPs also have the advantage of stabilizing the electron transfer rate on the SPE surface. After modifying the nitrate ISM, we carried out two experiments in which we performed CV measurements before and after electrode conditioning. The inset in Figure 6a shows the CV curve obtained before conditioning. The redox current of the NO 3 − -ISM/Co 3 O 4 NPs/SPE was found to be 0.417 µA (before conditioning), which was the same as that of the electrode after conditioning at 0.625 µA. The results for both cases showed that the CV current density was very low when the membrane was modified. This was due to the ion-selective membrane's nonconductive properties.

Chronopotentiometric Test
To estimate the effect of the solid-contact layer of the Co3O4 NPs on the capacitance of the NO3 − -ISE and investigate the stability of the prepared NO3 − -ISM/Co3O4 NPs/SPE, chronopotentiometry with a constant current was used to investigate and record the potential change of the prepared NO3 − -ISM/Co3O4 NPs/SPE. In the chronopotentiometric spectrum, the potential change rate can be described as ΔE/Δt = I/C (where ΔE is the potential change, Δt is the time variation, I is the applied current, and C is the low-frequency capacitance), and it can be used to determine electrode stability and is a significant index [42]. Chronopotentiometry experiments were performed by recording the potentials of the ISE in the current range of ±1 nA in a 0.01 M NaNO3 solution-the experimental results are shown in Figure 7. In the chronopotentiometric spectrum, the difference between the two types of electrodes occurred mostly due to their low-frequency capacitance differences. The SPE potential change rate reached 3 × 10 −6 V/s when there was no Co3O4 transducing layer, while a smaller potential change rate of 1.6 × 10 −6 V/s was recorded for the Co3O4 NPs/SPE. Finally, the capacitance of the NO3 − -ISM/Co3O4 NPs/SPE was calculated to be 62.5 µF, which was higher than that of the NO3 − -ISM/SPE, which was 33 µF. It can be seen from the above results that the nitrate ISE modified with Co3O4 NPs showed a significantly higher capacitance and lower potential change rate than the SPE under the same developed conditions. These results confirmed that the Co3O4 NPs were successfully used as a transducer material. EIS measurements were performed over frequencies ranging from 10 Hz to 10 6 Hz with an open-circuit voltage of 0.358 V and an excitation amplitude of 0.005 V. Figure 6b depicts the equivalent circuit used to fit the resulting Nyquist plots, in which R b , R ct , R s , C dl , CPE, and W represent the bulk membrane resistance together with the contact resistance between the underlying conductor and the ISM, the charge-transfer resistance, the solution resistance, the double-layer capacitance, the constant phase angle element, and the Warburg impedance, respectively [41]. The R ct values of the SPE (red points) were approximately 0.72 kΩ. For the Co 3 O 4 NPs/SPE (blue points), the semicircle diameter was increased at high frequencies compared to that for the SPE, which corresponded to an R ct of 0.4 kΩ. Because the ISM was nonconductive, the impedance increased after the addition of the nitrate ISM.

Chronopotentiometric Test
To estimate the effect of the solid-contact layer of the Co 3 O 4 NPs on the capacitance of the NO 3 − -ISE and investigate the stability of the prepared NO 3 − -ISM/Co 3 O 4 NPs/SPE, chronopotentiometry with a constant current was used to investigate and record the potential change of the prepared NO 3 − -ISM/Co 3 O 4 NPs/SPE. In the chronopotentiometric spectrum, the potential change rate can be described as ∆E/∆t = I/C (where ∆E is the potential change, ∆t is the time variation, I is the applied current, and C is the low-frequency capacitance), and it can be used to determine electrode stability and is a significant index [42]. Chronopotentiometry experiments were performed by recording the potentials of the ISE in the current range of ±1 nA in a 0.01 M NaNO 3 solution-the experimental results are shown in Figure 7. In the chronopotentiometric spectrum, the difference between the two types of electrodes occurred mostly due to their low-frequency capacitance differences. The SPE potential change rate reached 3 × 10 −6 V/s when there was no Co 3 O 4 transducing layer, while a smaller potential change rate of 1.6 × 10 −6 V/s was recorded for the Co 3 O 4 NPs/SPE. Finally, the capacitance of the NO 3 − -ISM/Co 3 O 4 NPs/SPE was calculated to be 62.5 µF, which was higher than that of the NO 3 − -ISM/SPE, which was 33 µF. It can be seen from the above results that the nitrate ISE modified with Co 3 O 4 NPs showed a significantly higher capacitance and lower potential change rate than the SPE under the same developed conditions. These results confirmed that the Co 3 O 4 NPs were successfully used as a transducer material.

Potentiometric Water Layer Tests
The water layer that forms at the membrane-inner electrode interface during the working process of the solid-state ISE will obviously affect the potential stability of the SC-ISE. Therefore, potentiometric water layer test experiments are an important component of investigating the performance of a solid-state ISE. For the potentiometric water layer test experiment, we carried out the test according to the literature method [43]. If there is a water layer, a positive potential drift will occur when the electrode is moved from the main ion solution to the interference ion solution. When the electrode is removed from the interference ion solution and placed back in the main ion solution, a negative potential drift will occur. In this experiment, NO3 − -ISM/Co3O4 NPs/SPE and NO3 − -ISM/SPE were activated in 0.1 M NaNO3 solution for 12 h before the water layer tests.
The specific experimental process was as follows, using NO3 − -ISM/Co3O4 NPs/SPE as an example: First, the activated NO3 − -ISM/Co3O4 NPs/SPE was immersed in a 0.1 M

Potentiometric Water Layer Tests
The water layer that forms at the membrane-inner electrode interface during the working process of the solid-state ISE will obviously affect the potential stability of the SC-ISE. Therefore, potentiometric water layer test experiments are an important component of investigating the performance of a solid-state ISE. For the potentiometric water layer test experiment, we carried out the test according to the literature method [43]. If there is a water layer, a positive potential drift will occur when the electrode is moved from the main ion solution to the interference ion solution. When the electrode is removed from the interference ion solution and placed back in the main ion solution, a negative potential drift will occur. In this experiment, NO 3 − -ISM/Co 3 O 4 NPs/SPE and NO 3 − -ISM/SPE were activated in 0.1 M NaNO 3 solution for 12 h before the water layer tests.
The specific experimental process was as follows, using NO 3 − -ISM/Co 3

Selectivity of the NO3 − -ISM/SPE and NO3 − -ISM/Co3O4 NPs/SPE
Selectivity, as the name indicates, is another important index that characterizes these sensors. The selectivities of the NO3 − -ISM/SPE and the NO3 − -ISM/Co3O4 NPs/SPE were estimated with the selectivity coefficient values obtained in relation to ion interference. The separable solution method was used to determine the selectivity coefficients of the tested electrodes [44], and their potentiometric selectivity coefficients were calculated using the Nicolskii-Eisenman equation: where I is the primary ion, J is the interfering ion, E is the potential, T is the temperature, F is Faraday's constant, R is the ideal gas constant, z is the valency of the ion, and a is the ion activity. Table 1 displays the results of the experiment, which confirmed that the studied ISEs showed high selectivity for their primary ion, and even the addition of the Co3O4 NPs did not impact their selectivity.

Selectivity of the NO 3 − -ISM/SPE and NO 3 − -ISM/Co 3 O 4 NPs/SPE
Selectivity, as the name indicates, is another important index that characterizes these sensors. The selectivities of the NO 3 − -ISM/SPE and the NO 3 − -ISM/Co 3 O 4 NPs/SPE were estimated with the selectivity coefficient values obtained in relation to ion interference. The separable solution method was used to determine the selectivity coefficients of the tested electrodes [44], and their potentiometric selectivity coefficients were calculated using the Nicolskii-Eisenman equation: where I is the primary ion, J is the interfering ion, E is the potential, T is the temperature, F is Faraday's constant, R is the ideal gas constant, z is the valency of the ion, and a is the ion activity. Table 1 displays the results of the experiment, which confirmed that the studied ISEs showed high selectivity for their primary ion, and even the addition of the Co 3 O 4 NPs did not impact their selectivity.

Effect of pH on the Response of the NO 3 − -ISM/Co 3 O 4 NPs/SPE for NO 3 − Detection
The most crucial element influencing electrode detection is the pH value. In this work, nitrate concentrations varying between 10 −4 and 10 −3 M were used to examine the impact of the pH on the electrode potential. The pH was changed from 1.0 to 10 by adding small amounts of HNO 3 and NaOH. The effect of pH on the electrode potential response is shown in Figure 9. The findings showed that the electrode potential was relatively stable between pH 3.0 and pH 8.0. Potential drift occurred when the pH value was too high or too low, which can be caused by hydrogen-ion interference. impact of the pH on the electrode potential. The pH was changed from 1.0 to 10 by adding small amounts of HNO3 and NaOH. The effect of pH on the electrode potential response is shown in Figure 9. The findings showed that the electrode potential was relatively stable between pH 3.0 and pH 8.0. Potential drift occurred when the pH value was too high or too low, which can be caused by hydrogen-ion interference.  Figure 10 shows a comparison of the potential responses of the SPE, Co3O4 NPs/SPE, NO3 − -ISM/SPE, and NO3 − -ISM/Co3O4 NPs/SPE. The average potential response data for six different concentrations of NO3 − , 10 −7 , 10 −6 , 10 −5 , 10 −4 , 10 −3 , and 10 −2 M, were investigated. The electrodes without the nitrate ISM modification, i.e., the SPE and Co3O4 NPs/SPE, showed almost no response to nitrate ions, which gave an extremely low potential response ( Figure 10A,B), while the NO3 − -ISM modification significantly increased the sensitivities of the electrodes (Figure 10C,D). Therefore, the NO3 − -ISM used in this study showed good sensitivity for detecting nitrate ions. Additionally, the NO3 − -ISM/Co3O4 NPs/SPE exhibited a greater potential response than the NO3 − -ISM/SPE, which indicated that Co3O4 was an excellent ion and electron transport material because it has a cubic spinel crystal structure and is a magnetic p-type semiconductor oxide [45]. Furthermore, Co3O4 has intriguing electrochemical, electronic, optical, catalytic, and electrocatalytic properties that could be used to increase electrolyte diffusion and provide more ion NPs/SPE, showed almost no response to nitrate ions, which gave an extremely low potential response ( Figure 10A,B), while the NO 3 − -ISM modification significantly increased the sensitivities of the electrodes (Figure 10C,D). Therefore, the NO 3 − -ISM used in this study showed good sensitivity for detecting nitrate ions. Additionally, the NO 3 − -ISM/Co 3 O 4 NPs/SPE exhibited a greater potential response than the NO 3 − -ISM/SPE, which indicated that Co 3 O 4 was an excellent ion and electron transport material because it has a cubic spinel crystal structure and is a magnetic p-type semiconductor oxide [45]. Furthermore, Co 3 O 4 has intriguing electrochemical, electronic, optical, catalytic, and electrocatalytic properties that could be used to increase electrolyte diffusion and provide more ion transport paths in supercapacitors.

Potentiometric Characteristics
In this study, the responses of the NO3 − -ISM/SPE and the new SC-NO3 − -ISE to nitrate ions were confirmed by using a series of NaNO3 solutions with concentrations ranging from 10 −2 to 10 −7 M. The potential responses of the NO3 − -ISM/SPE and SC-NO3 − -ISE transduced by the Co3O4 NPs are shown in Figure 11. Figure 11a, c shows the dynamic potential responses of the NO3 − -ISM/SPE and NO3 − -ISM/Co3O4 NPs/SPE. The results revealed that the NO3 − -ISM/SPE response was very small at low concentrations. Compared with the NO3 − -ISM/SPE, the NO3 − -ISM/Co3O4 NPs/SPE had a shorter response time and achieved potential balance in various nitrate ion solutions even at low concentrations. Moreover, the NO3 − -ISM/Co3O4 NPs/SPE showed a steady potential response to nitrate ions as the nitrate ion concentration was changed. The inset shows the potential response over time when the concentration was changed by one order of magnitude (10 −7 to 10 −6 M), which is considered to be the electrode's slowest response in the linear range. After changing the concentration by an order of magnitude, the time required to reach a steady-state signal was typically less than 60 s. These results showed that the NO3 − -ISM/Co3O4 NPs/SPE had a very short response time, and a stable signal was obtained in approximately 5.7 s, which is sufficient for many applications. The calibration curves for the two electrodes are shown in Figure 11b, d. The NO3 − -ISM/SPE showed a Nernstian response of −28.14 mV/decade (R 2 = 0.97) with a detection limit of 2.69 × 10 −5.6 M and a quantification limit of 8.17 × 10 −5.6 M. The NO3 − -ISM/Co3O4 NPs/SPE showed a Nernstian response slope of −56.78 mV/decade (R 2 = 0.99) with a detection limit of 1.04 × 10 −8 M and a quantification limit of 3.18 × 10 −8 M, which are better than those of the NO3 − -ISM/SPE. These findings demonstrate that the electrode response was significantly improved by using CO3O4 as a solid-contact layer.

Potentiometric Characteristics
In this study, the responses of the NO 3 − -ISM/SPE and the new SC-NO 3 − -ISE to nitrate ions were confirmed by using a series of NaNO 3 solutions with concentrations ranging from 10 −2 to 10 −7 M. The potential responses of the NO 3 − -ISM/SPE and SC-NO 3 − -ISE transduced by the Co 3 O 4 NPs are shown in Figure 11. Figure 11a,c shows the dynamic potential responses of the NO 3 − -ISM/SPE and NO 3 − -ISM/Co 3 O 4 NPs/SPE. The results revealed that the NO 3 − -ISM/SPE response was very small at low concentrations. Compared with the NO 3 − -ISM/SPE, the NO 3 − -ISM/Co 3 O 4 NPs/SPE had a shorter response time and achieved potential balance in various nitrate ion solutions even at low concentrations. Moreover, the NO 3 − -ISM/Co 3 O 4 NPs/SPE showed a steady potential response to nitrate ions as the nitrate ion concentration was changed. The inset shows the potential response over time when the concentration was changed by one order of magnitude (10 −7 to 10 −6 M), which is considered to be the electrode's slowest response in the linear range. After changing the concentration by an order of magnitude, the time required to reach a steady-state signal was typically less than 60 s. These results showed that the NO 3 − -ISM/Co 3 O 4 NPs/SPE had a very short response time, and a stable signal was obtained in approximately 5.7 s, which is sufficient for many applications. The calibration curves for the two electrodes are shown in Figure 11b,d. The NO 3 − -ISM/SPE showed a Nernstian response of −28.14 mV/decade (R 2 = 0.97) with a detection limit of 2.69 × 10 −5.6 M and a quantification limit of 8.17 × 10 −5.6 M. The NO 3 − -ISM/Co 3 O 4 NPs/SPE showed a Nernstian response slope of −56.78 mV/decade (R 2 = 0.99) with a detection limit of 1.04 × 10 −8 M and a quantification limit of 3.18 × 10 −8 M, which are better than those of the NO 3 − -ISM/SPE. These findings demonstrate that the electrode response was significantly improved by using CO 3 O 4 as a solid-contact layer. Figure 12 shows a comparison of different solid-state nitrate ISEs that have recently been reported in the literature for NO 3 − detection [36,[46][47][48]. The results show that the proposed NO 3 − -ISM/Co 3 O 4 NPs/SPE has superior sensitivity, response time, and high capacitance compared to the other electrodes, which demonstrates that the proposed sensors have great sensitivity, rapid response, marginally improved capacitance, and improved selectivity for nitrate ions.  Figure 12 shows a comparison of different solid-state nitrate ISEs that have recently been reported in the literature for NO3 − detection [36,[46][47][48]. The results show that the proposed NO3 − -ISM/Co3O4 NPs/SPE has superior sensitivity, response time, and high capacitance compared to the other electrodes, which demonstrates that the proposed sensors have great sensitivity, rapid response, marginally improved capacitance, and improved selectivity for nitrate ions.    [36,[46][47][48] are compared with this research.

Lifetime Study for the NO3 − -ISM/Co3O4 NPs/SPE
An investigation was conducted to evaluate the lifetime of the proposed NO3 − -ISM/Co3O4 NPs/SPE. The electrodes were calibrated in NaNO3 solutions with concentrations ranging from 10 −2 to 10 −7 M after conditioning in 10 −2 M NaNO3 for 24 h, and the results are shown in Figure 13a. All six electrodes showed similar responses, with a standard deviation of 1.39. Moreover, the measurements were carried out every two days while using the same electrode to investigate the stability of the proposed NO3 − -ISM/Co3O4 NPs/SPE, as shown in Figure 13b. Based on an analysis of the recorded data, the NO3 − -ISM/Co3O4 NPs/SPE had good stability over the time range 0-8 d. However, there was an obvious decrease in the sensor response, which reached approximately 50% after 8 d. These results indicated that the proposed NO3 − -ISM/Co3O4 NPs/SPE has an active life of approximately 8 d.

Analyses of Real Water Samples from an Aquaponic System
To investigate the utility of the proposed NO3 − -ISM/Co3O4 NPs/SPE for monitoring NO3 − levels in aquaponic water, two real water samples obtained from an aquaponic system were used. Simultaneously, we examined other water samples to confirm the

Analyses of Real Water Samples from an Aquaponic System
To investigate the utility of the proposed NO 3 − -ISM/Co 3 O 4 NPs/SPE for monitoring NO 3 − levels in aquaponic water, two real water samples obtained from an aquaponic system were used. Simultaneously, we examined other water samples to confirm the electrode's functionality, such as domestic wastewater, tap water, and Yangtze River water samples. In addition, the detection results obtained with the NO 3 − -ISM/Co 3 O 4 NPs/SPE were compared with those from a traditional detection method, i.e., spectrophotometry. Table 2 displays the composition of the aquaculture water used in this work. As shown in Table 3, the results for determinations of nitrate ions in the aquaponic system with the solid-state NO 3 − -ISM/Co 3 O 4 NPs/SPE were essentially consistent with the data obtained using spectrophotometry, demonstrating that our sensor can be utilized to efficiently detect NO 3 − in aquaculture water.

Conclusions
In this work, SPEs were used to develop new NO 3 − -ISM/Co 3 O 4 NPs/SPEs for the detection of NO 3 − in aquaponics water, and Co 3 O 4 NPs were used to modify the electrode surfaces to inhibit formation of an aqueous layer in the membrane-inner electrode interface. The electrode capacitance was increased by orders of magnitude through the introduction of Co 3 O 4 NPs, which was required for stable output. Our sensors also exhibited an excellent sensitivity of −56.78 mV/decade and a quick response time of 5.7 s, and XRD, EDS, SEM, and TEM were used to confirm the structural integrity. The key parameters for fabrication of the NO 3 − -ISM/Co 3 O 4 NPs/SPE were optimized to improve the NO 3 − detection performance, which combined high potential stability, strong reproducibility, and low interference. The outcomes for the analyses of real water samples from an aquaponic system demonstrated that the NO 3 − -ISM/Co 3 O 4 NPs/SPE developed in this paper has good commercial potential for the detection of nitrate ions in aquaponic water.