Development of All-Solid-State Potentiometric Sensors for Monitoring Carbendazim Residues in Oranges: A Degradation Kinetics Investigation

Abstract

Monitoring fungicide residues in orange fruits is vital, as fungicides for orange cultivation are increasingly used to prevent yield loss. At the same time, increasing restrictions are added by regulatory organizations. For facile on-site monitoring of the fungicide carbendazim (MBC), five ion-selective potentiometric sensors are proposed and compared. The first two sensors were prepared with a precipitation-based technique using molybdate (sensor 1) and tetraphenylborate (TPB) (sensor 2), respectively. Furthermore, two ionophore-based sensors were prepared using β-cyclodextrin as ionophore together with TPB (sensor 3) and tetrakis(4-chlorophenyl)borate (TpClPB) (sensor 4) as ion-exchanger. Further incorporation of multi-walled carbon nanotubes (MWCNTs) between the graphite rod and the sensing membrane of sensor 4 (sensor 5) further improved the stability and significantly lowered the limit of detection (LOD). Their performance was evaluated according to IUPAC recommendations, revealing linear response in the concentration range 1 × 10⁻⁴–1 × 10⁻² M, 1 × 10⁻⁵–1 × 10⁻² M, 1 × 10⁻⁵–1 × 10⁻³ M, 1 × 10⁻⁶–1 × 10⁻³ M, and 1 × 10⁻⁷–1 × 10⁻³ M with a Nernstian slope of 54.56, 55.48, 56.00, 56.85, and 57.34 mV/decade, respectively. The LOD values for the five sensors were found to be 7.92 × 10⁻⁵, 9.98 × 10⁻⁶, 9.72 × 10⁻⁶, 9.61 × 10⁻⁷, and 9.57 × 10⁻⁸ M, respectively. The developed potentiometric sensors were successfully applied to determine the residue and degradation rate of MBC in orange samples. After the researched fungicide was applied to the orange trees, the preharvest interval (PHI) could be calculated based on the MBC degradation kinetics determined in the tested orange samples.

1. Introduction

Carbendazim (MBC), a systemic benzimidazole fungicide widely employed to combat a broad range of plant fungal infections [1], has garnered attention from the World Health Organization (WHO), which categorizes it as a hazardous substance [2]. The use of MBC has been banned in Australia, most of the European Union, and the United States due to its severe toxicity and persistence [3]. It has been identified as a potential human carcinogen. It harms developing mammals and causes teratogenesis, developmental toxicity, embryotoxicity, and germ cell death in several mammalian species. It also has endocrine-disrupting effects and induces hepatocellular dysfunction. Residual traces of MBC in water, grains, soils, vegetables, and fruits pose significant health risks to consumers [4].

The widespread use of MBC is notable in citrus cultivation, considering the global importance of citrus fruits and their susceptibility to fungal infections [5,6]. Egypt, a leading exporter of citrus fruits, must adhere to established maximum residue limits (MRL) set by governmental organizations such as Codex Alimentarius and the European Commission (EU) to ensure food safety and minimize environmental contamination [7]. The EU has set the MRL for carbendazim in orange samples to 0.2 mg/kg, necessitating the development of highly sensitive analytical methods capable of detecting and quantifying this fungicide at exceedingly low concentrations.

In the literature, various methods have been reported for determining MBC in diverse environmental matrices. These methods include spectroscopic methods [8,9,10,11,12,13], voltammetric methods [14,15,16,17], and high-performance liquid chromatography (HPLC) coupled with ultraviolet [18,19,20,21,22], fluorescence [23,24,25,26], and mass spectrometry detection systems [7,27,28,29,30]. While many published methods are sensitive and selective, they often entail prolonged analysis times, expensive instrumentation, and intricate sample pretreatment procedures. Given these limitations, the development and application of potentiometric sensors as an alternative for determining carbendazim concentrations in environmental samples with minimum sample preparation is of great interest.

Despite the flexibility of all-solid-state potentiometric sensors, allowing for convenient on-site analyses, no potentiometric methods for carbendazim have been reported in the literature to date to the best of our knowledge. Thus, we aimed to design solid-contact ion-selective electrodes for the first time to determine MBC in orange samples, ensuring no interference from the matrix. Various graphite electrodes were constructed using ionophore and precipitation techniques. Furthermore, MWCNTs were used as an interface between the graphite and the sensing membrane to further improve the limit of detection (LOD), potential stability, and reproducibility of the proposed solid-contact electrodes to enable the reliable determination of carbendazim in orange samples at low concentrations.

The developed potentiometric sensors were subsequently employed in a real-life application to assess the degradation kinetics of carbendazim on oranges previously treated with this fungicide. The degradation kinetics served as a basis for estimating the preharvest interval (PHI), a critical parameter for determining the safe time to harvest oranges treated with MBC, showing the ability of the sensors to monitor the MBC concentration and their reliable determination of the MBC residual concentration in real samples.

2. Materials and Methods

2.1. Apparatus

A Jenway digital ion analyzer (model 3330, Essex, UK) was used for potentiometric measurements coupled with an Ag/AgCl double junction external reference electrode (Thermo Scientific Orion 900200, (Waltham, MA, USA); 3 M KCl saturated with AgCl as the inner filling solution and 1% KNO₃ as a bridge electrolyte).

A Jenway pH glass electrode no. 924005-BO3-Q11C (Essex, UK) was used for pH adjustments. A Bandelin Sonorex magnetic stirrer and heater, model Rx510S (Budapest, Hungary), was used for temperature adjustments, and a vortex mixer (F20230176 ZX3, VELP China Co. Ltd., Shanghai, China) for homogenization of citrus fruit.

The fruits were sprayed with the fungicide using a knapsack sprayer equipped with a nozzle connected to a 5 L capacity tank (Egypt Power, Giza, Egypt).

2.2. Standards and Samples

2.2.1. Pure Standards

Carbendazim was obtained from Sigma Aldrich as a certified fungicide standard. Its purity was confirmed to be 98.37 ± 0.31% using the reported method (LC-MS/MS method) [31].

2.2.2. Commercial Samples

The Ministry of Agriculture (Giza, Egypt) kindly provided Sendo^® (Shoura, Starchem, Cairo, Egypt), a 50% wettable powder (WP) of carbendazim.

2.3. Chemicals and Reagents

• Sodium Hydroxide (NaOH), hydrochloric acid (HCl), calcium chloride, fructose, and potassium chloride (El Nasr Company, Cairo, Egypt);
• High molecular weight polyvinyl chloride (PVC), tetrahydrofuran (THF) of HPLC grade, dioctyl phthalate (DOP), multi-walled carbon nanotubes (MWCNTs), β-cyclodextrin hydrate, and potassium tetrakis(4-chlorophenyl)borate (KTpClPB) (Sigma Aldrich, Darmstadt, Germany);
• Sodium tetraphenylborate (NaTPB), Fluka (Seelze, Germany);
• Ammonium molybdate, BDH Chemicals Ltd. (Poole, UK);
• Deionized water (DW) (MilliQ Plus, Millipore Iberica, Navalafuente, Spain);
• Thiabendazole (TBZ), (Sigma Aldrich, Germany);
• Sodium citrate, Prolabo (West Chester, PA, USA);
• A Britton-Robinson buffer was prepared by mixing equal volumes of 0.1 M acetic acid, 0.1 M boric acid, and 0.1 M phosphoric acid. The required pH was then adjusted using 1 M NaOH solution [32];
• Lactose monohydrate (PIOCHEM, Giza, Egypt).

2.4. Standard Solutions

2.4.1. Stock Standard Solutions

A 1 × 10⁻² M stock solution was prepared by dissolving 0.1912 g of MBC in the smallest amount of 0.1 M HCl, which was then completed to 100 mL with Britton-Robinson buffer pH 3 followed by sonication for 5 min.

2.4.2. Working Standard Solutions

Working standard solutions (1 × 10⁻⁹–1 × 10⁻³ M) were prepared by diluting the previously mentioned stock solution (1 × 10⁻² M) with Britton-Robinson buffer pH 3.

2.5. Procedure

2.5.1. Preparation of the Ion-Association Complexes

The sensing elements were prepared by mixing 100 mL of a saturated MBC solution with either 100 mL of a saturated molybdate aqueous solution or 100 mL of a saturated TPB aqueous solution. The resulting yellowish-white precipitate of molybdate–MBC ion-association complex or TPB-MBC ion-association complex was filtered, washed with distilled water, and air-dried at room temperature. The resulting powder was then further ground to obtain a fine powder.

2.5.2. Ion Selective Membrane Composition and Sensor Fabrication

Precipitation Technique

For sensor 1: 0.01 g molybdate-MBC ion-association complex, 0.35 mL DOP plasticizer, and 0.19 g PVC were mixed in a 5 cm glass Petri dish and dissolved in 5 mL THF.

For sensor 2: 0.01 g TPB-MBC ion-association complex, 0.35 mL DOP plasticizer, and 0.19 g PVC were mixed in a 5 cm glass Petri dish and dissolved in 5 mL THF.

Ionophore Technique

For sensor 3: 0.005 g TPB, 0.01 g β-cyclodextrin hydrate, 0.35 mL DOP plasticizer, and 0.19 g PVC were mixed in a 5 cm glass Petri dish and dissolved in 5 mL THF.

For sensor 4: 0.005 g potassium tetrakis(4-chlorophenyl)borate (KTpClPB), 0.01 g β-cyclodextrin hydrate, 0.35 mL DOP plasticizer, and 0.19 g PVC were mixed in a 5 cm glass Petri dish and then dissolved in 5 mL THF.

For sensors from 1 to 4, a graphite rod was immersed in the previously mentioned dispersion cocktail twice for about 30 s to achieve a uniform coating of the coating material on the graphite rod. Afterward, the coating was visually inspected to ensure a uniform and complete coverage of the graphite rod.

For sensor 5: 40 mg MWCNTs were weighed and dispersed in 10 mL THF. The mixture was then sonicated for 5 min to obtain a homogeneous dispersion. The sensor was prepared by dipping the graphite electrode into the MWCNTs suspension for about 30 s, followed by a visual inspection to ensure homogenous coverage of the graphite electrode. The electrode was then left to dry for 1 h at room temperature. Then, 20 µL of the previous dispersion cocktail for sensor 4 was drop-casted twice over the previously deposited MWCNTs. Finally, the prepared electrode was left in the air for 1 h to evaporate the solvent.

2.5.3. Potential Measurement

The coated graphite electrode was activated before its use for potentiometric measurements by conditioning it in a 1 × 10⁻² M MBC solution for 24 h.

The proposed sensors and the Ag/AgCl reference electrode were used to perform the potentiometric measurements. The conditioned electrodes were then separately immersed in 20 mL of each working standard solution of MBC, and the potential readings for the corresponding fungicide concentrations were recorded. Between each measurement, the electrodes were cleaned with deionized water. Regression equations were calculated and plotted as the measured electromotive force (emf) against the logarithmic MBC concentrations.

2.5.4. Experimental Conditions

Identification of Electrochemical Properties of the Proposed Electrodes

The electrochemical properties of the proposed electrodes were evaluated according to IUPAC recommendations [33].

To check the dynamic response time of the studied electrodes, different concentrations of MBC solutions (ranging from 1 × 10⁻⁷ to 1 × 10⁻² M) were used. The measurements were taken in a sequence from low to high concentration and back. The reproducibility and stability of the readings through time intervals from 10 s to 2 min were examined to determine the response time of the electrodes.

The lifetime of the proposed electrodes was assessed by periodically testing the slope of the working standard solutions.

pH Effect

The response of the electrodes under study was examined at two concentration levels (1 × 10⁻³ and 1 × 10⁻⁴ M) while varying the pH from 1 to 10. Adjusting pH involved adding small volumes of 1 M HCl and 1 M NaOH.

The potential measured for both concentration levels was plotted against the respective pH values.

Effect of Foreign Compounds

The ability of an electrode to respond to its primary ion in the presence of other interfering ions is known as electrode selectivity [34]. Potentiometric selectivity coefficients can be used to evaluate selectivity using the separate solution method (SSM) or the matched potential method (MPM) [35]. The potential selectivity coefficients (K^pot_A,B) were calculated using MPM according to IUPAC [33,36] using the following equation:

K^pot_A,B = (a′_A − a_A)/a_B

where a′_A: the activity of the primary ion;

a_A: the activity of the reference solution of the primary ion;

a_B: the activity of the interfering ion;

Calcium chloride, sodium citrate, potassium chloride, fructose, lactose, and the chemically related TBZ are used for this study.

Potentiometric Water Layer Test

The development of an aqueous layer between the membrane and solid-contact electrode has a significant impact on the electrode’s long-term durability. Therefore, the water layer test became a standard protocol for the assessment of the performance characteristics of solid-contact ion-selective electrodes (SC-ISE) [37].

The test relies on sensing any potential drift while changing concentrations from MBC solution (1 × 10⁻⁴ M) to a solution of higher concentration of an interfering ion, TBZ (1 × 10⁻² M), and then returning to the MBC solution. Any potential drift detected indicates the formation of a water layer due to the change in the ionic composition of the membrane caused by ion fluxes [38]. Each electrode was soaked separately in each solution for one hour.

2.5.5. Applications

Application to Spiked Orange Samples

Ten grams of well-ground blank orange samples were spiked with 10 mL of MBC solutions in a range of (10⁻⁴ to 10⁻² M) for sensors 1 and 2 and in the range of (10⁻⁵ to 10⁻³ M) for sensors 3, 4, and 5. The ground orange samples were mixed and sonicated for 5 min with the respective MBC solutions and were finally filtered through a qualitative filter paper. The electrodes were individually immersed in the filtered samples, and the emf was measured. Between each measurement, the electrode was cleaned with deionized water. The emf values obtained were recorded, and the concentration was determined using the previously established regression equations. Recovery% ± SD is used to describe the results.

Applications to the Treated Orange Samples

a.

Field Experiment

The experiment was conducted at the Horticulture Research Institute (HRI), located at 9 Cairo University St., Orman, Giza. The field was planted with orange trees and separated into two sections, one for carbendazim fungicide and the other for non-fungicide treatment (control). Each section had two trees.

The experiment started on 1 January 2019. The MBC area was treated with the Ministry of Agriculture’s prescribed dose of the fungicide, which was applied to orange trees as follows: 7.5 gm of Sendo^®, 50% WP, was dissolved in 15 L of water, which was enough for two trees.

A knapsack sprayer fitted with a nozzle was used to apply the diluted fungicide to the designated area.

b.

Sampling and Storage

In the treatment and the control areas, random samples of 2 kg orange fruits were taken to represent all the plants in that region.

To assess the degradation kinetics of MBC and establish its preharvest interval (PHI), sampling was performed on several dates: 0 (2 h after treatment), 1, 2, 5, 7, 14, 21, and 28 days after treatment.

The collected samples were immediately transported in iceboxes to the lab. The oranges were ground to produce a homogenized sample, and the homogenized sample was placed in labeled bags and stored at −20 °C until analysis.

After homogenization, 10 g of each treated orange sample was mixed with 10 mL of Britton-Robinson buffer pH 3 and sonicated for 5 min to extract the MBC. The extract was then filtered using filter paper. The measurement process was the same as for the spiked orange sample described in Section 3.6.

3. Results and Discussion

In this study, we present solid-contact potentiometric sensors for the first time for the determination of MBC fungicide, both unmodified and modified with MWCNTs. A graphite rod served as the solid contact, while the sensing component of the polymeric membrane was either precipitation- or ionophore-based, with β-CD acting as the ionophore. These sensors offer high sensitivity and selectivity for the potentiometric detection of MBC in orange samples, permitting them to follow its degradation kinetics while benefiting from straightforward sample preparation procedures.

3.1. Sensor Fabrication

To develop sensitive and selective potentiometric sensors for the determination of MBC in orange samples, the components of the sensing polymeric membrane have to be carefully selected and optimized. Therefore, precipitation- and ionophore-based techniques for MBC sensing are tested and compared in this work to determine the best sensing cocktail for MBC.

For the preparation of precipitation-based potentiometric sensing membranes, appropriate precipitating agents that can form ion-association complexes with MBC of low solubility and known stoichiometry have to be selected. For that, the ionization of MBC has to be taken into account. MBC, with a pKa of 4.48, forms an HCl salt in an acidic medium and possesses a protonated imidazole ring that serves as an anionic exchanger [39]. The cationic site within the ionized MBC structure interacts with the anionic site in cationic exchangers such as molybdate or TPB, resulting in the formation of a 1:1 water-insoluble ion pair (as shown in Figure 1) with a low solubility product—this forms the basis for the preparation of potentiometric sensors (sensors 1 and 2). The Nernstian response of the proposed sensors confirmed the 1:1 ratio, characteristic for monovalent substances, yielding a slope of approximately 60 mV (

Table 1. Potentiometric characteristics of MBC membrane sensors.

Parameter	Sensor 1	Sensor 2	Sensor 3	Sensor 4	Sensor 5
Slope (mV/decade) *	54.56	55.48	56.00	56.85	57.34
Intercept (mV)	526.5	438.33	448.53	467.37	506.18
Response time (s)	25	20	20	15	10
Working pH range	2–4.5	2–5	2–4	2–4.5	2.5–4.5
Linearity range (M)	1 × 10−4–1 × 10−2	1 × 10−5–1 × 10−2	1 × 10−5–1 × 10−3	1 × 10−6–1 × 10−3	1 × 10−7–1 × 10−3
Stability (days)	40	45	50	60	65
Accuracy (Mean ± SD) **	99.42 ± 1.13	98.94 ± 0.73	99.43 ± 0.95	99.31 ± 1.03	99.36 ± 0.91
Correlation coefficient	0.9997	0.9999	0.9998	0.9999	0.9998
Intra-day precision (RSD%) ***	1.49	0.51	1.17	1.14	0.65
Inter-day precision (RSD%) ***	1.74	1.07	1.59	1.19	0.69
LOD (M) ****	7.92 × 10−5	9.98 × 10−6	9.72 × 10−6	9.61 × 10−7	9.57 × 10−8
MRL (M)	1.05 × 10−6

* Average of three determinations. ** Mean of five concentrations (1 × 10⁻², 5.5 × 10⁻³, 1 × 10⁻³, 5.5 × 10⁻⁴, 1 × 10⁻⁴ M) for sensor 1, (5.5 × 10⁻³, 1 × 10⁻³, 5.5 × 10⁻⁴, 1 × 10⁻⁴, 1 × 10⁻⁵ M) for sensors 2 and 3, (1 × 10⁻³, 5.5 × 10⁻⁴, 5.5 × 10⁻⁵, 5.5 × 10⁻⁶, 1 × 10⁻⁶ M) for sensors 4 and 5. *** Mean of three concentrations (1 × 10⁻², 5.5 × 10⁻⁴, 1 × 10⁻⁴ M) for sensors 1 and 2, (5.5 × 10⁻⁴, 1 × 10⁻⁴, 1 × 10⁻⁵ M) for sensor 3, (5.5 × 10⁻⁴, 1 × 10⁻⁴, 5.5 × 10⁻⁶ M) for sensors 4 and 5. **** Limit of detection (measured by intersection of the extrapolated arms of the corresponding calibration curve).

).

Additionally, the ionophore-based method was employed for sensors 3, 4, and 5, relying on the molecular recognition of MBC through inclusion-complex formation between MBC and β-cyclodextrin hydrate. Various interactions, such as Van der Waals forces and hydrogen bonds, occur between the host and guest molecules [40], with binding strength primarily determined by the guest’s size and geometry [41]. The inclusion complex confines the hydrophobic part of MBC within the hydrophobic pocket of β-cyclodextrin hydrate, where it was found that the benzene moiety of MBC is inserted into the host cavity, resulting in a non-polar/non-polar association complex [42].

After selecting and preparing the different recognition elements for MBC in the sensing membrane, further components of the sensing membrane need to be carefully selected. In the membranes with the ionophore-based method, sodium tetraphenylborate is used as a lipophilic ion additive to increase ionic mobility in the sensor matrix while stabilizing the charged inclusion complex and reducing anion interference [43,44].

For the polymer matrix harboring the sensing components, a suitable polymer enabling optimum diffusion of membrane components and giving the membrane suitable mechanical properties is needed. PVC is a relatively cheap polymer and enables good entrapment of the membrane components, so it is considered one of the most used polymers [45,46]. Furthermore, it enables good partitioning of hydrophobic compounds from aqueous solutions [47], which is highly desirable for sensitive sensing.

The plasticizer is an essential component in PVC membranes as it adjusts its physical and mechanical properties, increasing the membrane’s flexibility [48]. Plasticizers are water-immiscible, high-boiling organic solvents that act as the membrane solvent and control the dielectric constant of the membrane phase, thus impacting the target ion extraction efficiency of the membrane electrode and enhancing the electrochemical characteristics of PVC membranes [49]. Among plasticizers, phthalates are popular for their ability to increase membrane flexibility and durability. They are esters of polycarboxylic acids that function by embedding themselves between the polymer chains, thereby increasing the free volume and spacing the chains apart. The addition of more plasticizer results in greater flexibility and durability of the membrane [50]. Notably, the DOP-PVC electrode exhibited superior performance when compared to other plasticizers [49].

The use of nanomaterials, such as MWCNTs, in sensor fabrication can enhance sensor performance due to their extraordinary properties, including enhanced conductivity and high surface area. MWCNTs, in particular, have been found to be efficient ion-to-electron transducers in solid-contact (SC) potentiometric sensors, contributing to improved transduction of chemical signals to electrical signals [44]. Additionally, the high hydrophobicity of MWCNTs prevents water-layer formation, thereby increasing the stability and reproducibility of the SC ion-selective electrode [51,52]. Thus, in this study, a sensor was developed where MWCNTs were integrated as a film between the graphite rod and sensing ion-selective membrane to enhance overall sensor performance and lower the LOD of the potentiometric sensor.

3.2. Sensor Calibration and Response Time

According to IUPAC recommendations, the electrochemical performance of the suggested sensors was assessed [33], as shown in Table 1.

In order to determine the most appropriate ion-exchanger, sensors 1 to 4 were prepared with molybdate, TPB, TPB, and KTpClPB, respectively, with sensors 1 and 2 employing a precipitation-based approach and sensors 3 and 4 utilizing an ionophore-based method. The response of the four sensors to the change in MBC concentration was recorded. The linear response was in the range of 1 × 10⁻⁴–1 × 10⁻² M, 1 × 10⁻⁵–1 × 10⁻² M, 1 × 10⁻⁵–1 × 10⁻³ M, and 1 × 10⁻⁶–1 × 10⁻³ M, respectively, with a near-ideal monovalent cationic Nernstian slope of 54.56, 55.48, 56.00, and 56.85 mV/decade, respectively. The correlation coefficients for all sensors were notably high at 0.9997, 0.9999, 0.9998, and 0.9999, respectively, indicating strong linearity in their responses. The corresponding calibration curves are shown in Figure 2.

Determining the limit of detection (LOD) involved identifying the intersection point between the two linearly extrapolated segments of the calibration curve, resulting in LOD values of 7.92 × 10⁻⁵, 9.98 × 10⁻⁶, 9.72 × 10⁻⁶, and 9.61 × 10⁻⁷ M for sensors 1 to 4, respectively. Additionally, these sensors exhibited rapid response times, stabilizing within ±1 mV of the final equilibrium emf value within 15 to 25 s after a tenfold increase in fungicide concentration. Moreover, the electrodes demonstrated good stability with a lifetime ranging from 40 to 60 days. The developed sensors showed good accuracy and precision results, with mean recoveries in the range of 98.94% to 99.43% with RSD% in the range of 0.51% to 1.74% (Table 1).

Comparing the two sensors with the same ion-exchanger TPB (sensor 2 and 3) but based on different preparation techniques, the results show that sensor 3 exhibited slightly superior performance than sensor 2, which is most probably attributed to the incorporation of β-CD in the membrane. The presence of an ionophore slightly improved detection limit, selectivity, electrochemical performance, and stability, as the β-CD ionophore has a lipophilic core, thus forming a stable complex with the studied fungicide [53,54].

Sensor 4 is a further development of sensor 3, where the ionophore-based technique has been shown to be superior, with the ion exchanger changed to KTpClPB. Sensor 4 approached the monovalent Nernstian slope with a further improvement of the detection limit, stability, and response time, indicating the favorable interaction between MBC and TpClPB⁻ (sterically favored) owing to its hydrophobic nature. The TpClPB⁻ is more hydrophobic than molybdate or TPB, which improves the electrochemical performance of the investigated sensor, as there was a clear correlation between lipophilicity of the cation-exchanger and an enhancement of the detection limit. Furthermore, the increased lipophilicity of the ion exchanger minimizes its leaching and, hence, improves the sensor lifetime [53,54,55]. Therefore, KTpClPB was selected as the ion exchanger for the design of MWCNTs modified sensor.

For sensor 5, the implementation of MWCNTs as a thin film between the graphite electrode and the sensing membrane further improved the electrochemical performance of the proposed sensor. MWCNTs can improve the conductivity of the sensor by increasing the transduction of the chemical signal to an electrical signal [44]. By increasing the transduction, the potential response of the sensor reached Nernstian values of 57.34 with a correlation coefficient of 0.9998. The calibration curve is shown in Figure 2.

By increasing the conductivity and improving the ion-to-electron transduction of the sensor, the dynamic working range and response time of the sensor can be improved [44,56]. Thus, the modification of the sensor with MWCNTs with its high surface area resulted in a wider dynamic working range (1 × 10⁻⁷–1 × 10⁻³ M), shorter response time (10 s), and a LOD of 9.57 × 10⁻⁸ M, surpassing regulatory limits (MRL of 0.2 mg/kg which is equivalent to 1.05 × 10⁻⁶ M) indicating the high sensitivity of the sensor. The electrode demonstrated a prolonged lifetime of 60 days and exhibited excellent accuracy and precision, with a recovery rate of 99.36% and RSD% of 0.69%.

Overall, the developed sensors offer promising capabilities for sensitive and selective detection of MBC in orange samples, laying the foundation for practical applications in agricultural and environmental monitoring.

3.3. Effect of pH

Figure 3 presents the outcomes of a distinct investigation conducted across two concentration levels (1 × 10⁻³–1 × 10⁻⁴ M) across the pH range of 1–10, exploring the impact of pH on the potential response of the five proposed sensors. The potential remained essentially constant for the respective sensors for the five proposed sensors across the pH ranges of 2–4.5, 2–5, 2–4, 2–4.5, and 2.5–4.5. However, at higher and lower pH values, the presence of hydroxyl and hydronium ions, occurring at comparatively higher concentrations than the primary ion, could interfere. Moreover, the salting out of the MBC base occurred at elevated pH levels. Furthermore, MBC was ensured to be fully protonated at a pH lower than one unit of its pKa (pKa = 4.48), which is important for the reliable and accurate performance of the ion-selective electrode. Hence, the pH was regulated using Britton-Robinson buffer at pH 3 throughout the investigation to maintain consistency.

3.4. Sensors’ Selectivity

One of the most crucial aspects of ISEs is the selectivity coefficient. It shows that the sensor can sense the target ion even in the presence of complex matrices and other interfering ions. The potential selectivity coefficients (K^pot_A,B) were computed using MPM, the IUPAC recommended method [33,36]. Utilization of MPM adds strength to our study, as it operates independently of the Nikolsky–Eisenman equation. Unlike the Nikolsky–Eisenman equation, which assumes a Nernstian response for both the primary and interfering ions and may not accurately represent responses in situations where ions of different charges significantly contribute to the potential [57,58], MPM offers a robust alternative.

According to this method, the potentiometric selectivity coefficient is defined as the quotient of the activity of the primary and interfering ions that yield an equal potential change under identical conditions. Practically, this is performed by adding an MBC solution of known activity (a′_A) to a reference MBC solution having a fixed activity (a_A). The change in potential, labeled as ∆E, is recorded. This is followed by the addition of a solution containing an interfering ion (sodium citrate, KCl, CaCl2, lactose, fructose, or TBZ) to the reference MBC solution until it achieves the same potential difference ∆E, and noting the used interfering ion activity. It is crucial that the potential change generated under a constant primary ion background in both cases remains identical [57,58].

The results of the potentiometric selectivity coefficients of the proposed sensors are shown in

Table 2. Potentiometric selectivity coefficients K^pot_{MBC, interference} for the investigated MBC sensors.

Interferant		Sensor 1	Sensor 2	Sensor 3	Sensor 4	Sensor 5
Calcium Chloride	KpotMBC, int.	9.81 × 10−1	7.44 × 10−1	2.98 × 10−1	5.69 × 10−3	2.84 × 10−3
	±SE	1.79 × 10−1	7.30 × 10−2	4.39 × 10−2	1.40 × 10−3	1.20 × 10−3
Sodium Citrate	KpotMBC, int.	7.78 × 10−1	8.89 × 10−1	3.53 × 10−2	9.80 × 10−2	8.0 × 10−3
	±SE	9.45 × 10−2	8.85 × 10−2	7.90 × 10−3	1.45 × 10−3	1.50 × 10−3
Potassium Chloride	KpotMBC, int.	1.48 × 10−1	6.48 × 10−2	3.89 × 10−1	1.45 × 10−3	3.53 × 10−3
	±SE	8.55 × 10−2	7.35 × 10−2	6.60 × 10−2	1.20 × 10−3	1.30 × 10−3
Lactose	KpotMBC, int.	1.98 × 10−1	4.69 × 10−2	3.75 × 10−2	5.94 × 10−3	1.13 × 10−3
	±SE	6.65 × 10−2	1.07 × 10−2	1.14 × 10−2	1.10 × 10−3	1.10 × 10−3
Fructose	KpotMBC, int.	1.15 × 10−1	3.73 × 10−2	2.36 × 10−2	9.1 × 10−3	1.79 × 10−3
	±SE	1.22 × 10−1	1.24 × 10−2	1.69 × 10−2	1.50 × 10−3	7.00 × 10−4
TBZ	KpotMBC, int.	5.35 × 10−1	8.08 × 10−2	3.53 × 10−2	9.65 × 10−2	1.78 × 10−2
	±SE	1.33 × 10−1	1.46 × 10−2	1.36 × 10−2	2.88 × 10−2	2.10 × 10−3

SE = Standard Error.

. K^pot_A,B values for all five sensors were below 1, indicating their high selectivity toward MBC even in the presence of other interferences and similar fungicides. Remarkably, the sensor modified with MWCNTs (sensor 5) exhibited the most favorable selectivity outcomes among the majority of tested interferences.

3.5. Potentiometric Water Layer Test

Assessment of the formation of a water layer between the sensing membrane and the SC electrode can be investigated using the water layer test developed by Fibbioli et al. [59]. This test holds significance as the formation of a water layer can substantially impact the sensor’s performance, prolonging the time required to achieve stable readings and reducing the potentiometric sensor’s lifespan due to possible contact loss and membrane delamination from the SC electrode over time [60].

As shown in Figure 4, no noticeable drift in the potential response (33 µV/h) of the proposed sensor 5 was observed, indicating that the inclusion of MWCNTs effectively prevents the formation of an aqueous layer due to their highly lipophilic nature [61].

Sensors 1 to 4 likely experienced varying degrees of water film formation between the SC electrode and the sensing membrane, as evidenced by the varying degrees of potential drift observed for each sensor. The calculated potential drifts were determined to be 5.13 mV/h, 1.96 mV/h, −1.93 mV/h, and −1.26 mV/h for sensors 1 to 4, respectively. In addition, it was also noticed that the potential drift decreased with the ascending order of the lipophilicity of the investigated ion pair [62,63].

3.6. Application to Spiked Orange Samples and Study of the Degradation Rate of MBC in Orange Samples

The performance of the proposed sensors was evaluated using spiked orange samples to assess their suitability for real-world applications. Following the preparation and analysis of the spiked orange samples utilizing our validated method, MBC recoveries were computed, yielding acceptable results, as presented in

Table 3. Potentiometric determination of MBC in spiked orange samples using the proposed sensors.

Spiked Concentrations (M)	Recovery * %
	Sensor 1	Sensor 2	Sensor 3	Sensor 4	Sensor 5
10−2	95.89	105.65	---------	-----------	----------
10−3	113.69	101.20	99.11	97.30	100.26
10−4	103.26	100.41	102.61	103.63	102.45
10−5	---------	---------	111.99	98.19	94.41
Mean ± SD	104.28 ± 8.94	102.42 ± 2.83	104.57 ± 6.66	99.71 ± 3.43	99.04 ± 4.16

* Average of three determinations.

.

To investigate the degradation of MBC in oranges, treated orange samples were collected at various time intervals.

Table 4. Measured concentrations of declining MBC residues in orange samples using the proposed sensors.

Time (Days)	Sensor 1		Sensor 2		Sensor 3		Sensor 4		Sensor 5
	Conc (M) *	% Loss	Conc (M) *	% Loss	Conc (M) *	% Loss	Conc (M) *	% Loss	Conc (M) *	% Loss
0	N.D **	------	3.16 × 10−5	0	4.17 × 10−5	0	3.66 × 10−5	0	4.31 × 10−5	0
1	N.D	------	2.82 × 10−5	10.76	3.53 × 10−5	15.35	2.43 × 10−5	33.61	3.89 × 10−5	9.74
2	N.D	------	2.45 × 10−5	22.47	2.83 × 10−5	32.13	1.99 × 10−5	45.63	3.02 × 10−5	29.93
5	N.D	------	1.76 × 10−5	44.30	2.18 × 10−5	47.72	1.79 × 10−5	51.09	2.63 × 10−5	38.98
7	N.D	------	1.02 × 10−5	67.72	1.42 × 10−5	65.95	6.91 × 10−6	81.12	1.29 × 10−5	70.07
14	N.D	------	N.D	------	N.D	------	3.99 × 10−6	89.10	5.01 × 10−6	88.38
21	N.D	------	N.D	------	N.D	------	1.43 × 10−6	96.09	1.71 × 10−6	96.03
28	N.D	------	N.D	------	N.D	------	N.D	------	4.79 × 10−7	98.89

* Average of three determinations. ** N.D: not detected.

shows the concentrations of residual MBC in the treated orange samples determined at different time points using the different sensors. Due to the wider dynamic working range and much lower detection limit of sensor 5, it was possible to track the residual MBC concentration in orange samples for a longer time (until it dropped below the MRL), while this was not possible with the other sensors. However, by plotting the residual MBC concentration versus time and determining the kinetics of MBC degradation, it is possible with the other sensors to estimate the time when the residual MBC concentration drops below the MRL and thus the PHI, as shown in

Table 5. Results of the kinetic study of MBC residues in orange samples using the proposed sensors.

	Sensor 2	Sensor 3	Sensor 4	Sensor 5
Order	First order (At = Aoe−kt)
R2	0.9954	0.9982	0.9940	0.9989
Degradation constant (k) * (days−1) ± SE	0.1569 ± 0.0176	0.1482 ± 0.0113	0.1495 ± 0.0136	0.1582 ± 0.0069
Half-life (t1/2) ** (days)	4.42	4.68	4.64	4.38
PHI (days)	22	25	23	24

SE = Standard Error. * k = −Slope. ** t_1/2 = 0.693/k.

.

Plotting the MBC residue concentration against time, as shown in Figure 5, revealed an exponential relationship, indicating that the degradation of MBC in orange samples followed first-order kinetics as determined by the following equation [64]:

A_t = A_oe^−kt

where A_t is the concentration of the fungicide residue at time interval t (days) after fungicide application, A_o denotes the initial residue concentration at zero time (t = zero, after 2 h from fungicide application), and k is the degradation rate constant.

Table 5 also provides a summary of the degradation parameters of MBC obtained from each sensor, where k is the degradation rate constant, and t_1/2 represents the half-life (equal to 0.693/k). It is noteworthy that even with the other sensors (sensors 2–4), the determined PHI (through extrapolation of the degradation kinetics) is very similar to the PHI determined with sensor 5, where no extrapolation was needed due to its ability to measure concentrations well below the MRL. This shows the accuracy of the developed sensors and the subsequent exponential fitting of the first-order degradation kinetics of MBC, allowing reliable estimation of the PHI.

As MBC is not approved by the EU Commission and USDA (the United States Department of Agriculture), our proposed method allows the determination of MBC in minute concentrations in a very short time with a simple purification step of the collected orange samples, thus preventing their rejection for export to other countries.

3.7. Statistical Analysis

The Student’s t-test and variance ratio F-test were employed to assess the validity of the proposed method. A statistical comparison between the method proposed in this study and the previously referenced method for MBC [31] is presented in

Table 6. Statistical comparison between the proposed potentiometric method using five sensors and the reported method.

Parameter	Proposed Method					Reported Method **
	Sensor 1	Sensor 2	Sensor 3	Sensor 4	Sensor 5
Mean	99.42	98.95	99.43	99.31	99.36	98.37
SD	1.13	0.73	0.95	1.03	0.91	0.31
Variance	1.28	0.53	0.90	1.06	0.83	0.09
N	5	5	5	5	5	3
Student’s t-test (2.447) *	1.535	1.284	1.824	1.507	1.790	----------
F (39.248) *	13.733	5.710	9.705	11.375	8.791	----------

* The theoretical values of Student’s t-test and F at p = 0.05. ** The reported method described in [31] utilized an LC-MS/MS approach with an Inertsil1 ODS-4 column (50 mm × 2.1 mm i.d., 3 μm particle size). The mobile phase consisted of HPLC-grade water with 5 mM ammonium formate (eluant A) and HPLC-grade methanol with 5 mM ammonium formate (eluant B). The gradient elution proceeded as follows: 0–8 min: 95% B; 8–12 min: 5% B using a flow rate of 0.5 mL/min and an injection volume of 15 μL.

.

The results indicate that there is no statistically significant difference between the reported and the proposed method, as indicated by the estimated t and F values being lower than the theoretical ones at p = 0.05.

3.8. Comparison to Other Electrochemical Methods

A literature survey for other electrochemical methods for the determination of MBC in orange reveals that no potentiometric determination of MBC has been conducted to date. The other electrochemical methods for the determination of MBC found in the literature are mainly voltammetric methods. A collection of these electrochemical methods can be found in

Table 7. Comparison of the different electrochemical methods for the determination of carbendazim in orange.

Technique	Sensor Used	Linearity Range	Limit of Detection (LOD)	Application and Matrix	Sample Pretreatment	Degradation Kinetics Study	Reference
Potentiometric Methods
Potentiometry with Ion-Selective Electrodes (ISE)	Detection of carbendazim using modified electrodes with precipitation-based and ionophore-based techniques, including MWCNTs for enhanced sensitivity	1 × 10−7−1 × 10−3 M (0.019–190 µg/mL)	9.57 × 10−8 M (0.0183 µg/mL)	Orange samples	Grinding and homogenization of orange fruit, mixing with Britton-Robinson buffer, sonication, and finally, filtration using filter paper	Yes Sensors used to track degradation kinetics of MBC in oranges	This work
Voltammetric Methods
Square-wave anodic stripping voltammetry	Carbon paste electrode modified with cobalt phthalocyanine	49.7–384.6 × 10−9 M	5.7 × 10−10 M	Orange and pineapple juice samples	None	No	Severo et al. (2024) [65]
Differential-pulse anodic stripping voltammetry	Glassy carbon electrode modified with molecularly imprinted polymer on CdMoO4/g-C3N4 nanocomposite	0.1–10 × 10−10 M	2.5 × 10−12 M	Orange and apple juice	Sonication for 20 min, centrifugation, filtration of the supernatant with a 0.50 µm filter, and dilution with 0.1 M PBS *	No	Yola et al. (2022) [66]
Differential pulse voltammetry	Carbon paste electrode modified with silver nanoparticles on fumed silica (FS@Ag)	5.0 × 10−8–3.0 × 10−6 M	9.4 × 10−10 M	River water, tomato juice, and commercial orange and apple juices	Filtration through 0.45 µm membrane filter and dilution with PBS *	No	Özcan et al. (2021) [67]
Differential pulse voltammetry	Carbon paste electrode modified with biochar and reduced graphene oxide nanocomposite	3.0 × 10−8–9.0 × 10−7 M (30–900 nmole/L)	2.3 × 10−9 M (2.3 nmole/L)	Spiked whole orange juice, lettuce leaves, drinking water, and wastewater samples	Extraction in methanol in an ultrasonic bath, then filtration of the extract by membrane syringe filter and concentration under a flow of N2 gas	No	Sant’Anna et al. (2020) [68]
Differential pulse voltammetry	Carbon paste electrodes modified with the nanocomposite based on ZnCdTe semiconductor nanocrystals synthesized in situ on reduced graphene oxide	9.98 × 10−8 to 1.18 × 10−5 M	9.16 × 10−8 M	Spiked orange juice samples.	Extraction with a manual extractor, the collected juice was centrifuged. The samples were stored in an amber glass bottle in the refrigerator	No	Santana et al. (2019) [69]
Differential pulse voltammetry	Carbon paste electrode	1.49 × 10−8–2.38 × 10−7 M (2.84 to 45.44 μg/L)	5.02 × 10−9 M (0.96 μg/L)	Spiked ultrapurified water, river water, and orange juice	Filtered through cotton	No	Arruda et al. (2016) [70]
Square-wave voltammetry	Glassy carbon electrode modified with a thin film of mesoporous silica/multi-walled carbon nanotubes	2.0 × 10−7 –4.0 × 10−6 M (0.2 to 4.0 μM)	5.6 × 10−8 M (0.056 μM)	Spiked orange juice sample	None	No	Razzino et al. (2015) [71]

* PBS: Phosphate buffer solution.

.

Although the developed MWCNT-based potentiometric sensor (sensor 5) does not show the lowest LOD among the developed electrochemical sensors found in the literature, it is still similar or even more sensitive to some voltammetric sensors [69,70,71]. Furthermore, the LOD achieved through the developed potentiometric sensors (especially sensor 5) was well below the MRL of MBC, enabling the monitoring of MBC residues in orange fruits. Comparing the linear range of the potentiometric sensors with the different voltammetric sensors shows a much wider linearity range of the developed potentiometric sensors.

Comparing the application of the different voltammetric sensors with the application of the potentiometric sensors developed in this work shows that all voltammetric sensors were applied to orange juice, while in this work, the whole orange fruit was used for the determination of MBC, enabling the detection of MBC residues present in the fruit before being further processed to juice. Furthermore, the degradation kinetics of MBC were determined for the first time using an electrochemical sensor, which is important to determine the PHI of the orange fruit sprayed with MBC fungicide.

4. Conclusions

The present work presents the development of an optimal potentiometric sensor for the sensitive determination of MBC in orange samples. An in-depth comparative study was carried out where the performance of the potentiometric sensors with different ion-exchangers was performed using molybdate, TPB, and KTpClPB. The use of the ionophore β-CD together with the ion-exchanger KTpClPB in the sensing membrane showed good sensing properties. The performance, sensitivity, and stability of the sensor were further enhanced by inserting a thin film of MWCNTs between the graphite electrode and the sensing membrane, resulting in a sensor with the highest sensitivity and lowest detection limit in the tested sensor series.

In general, the proposed sensors have short response time, high accuracy, precision, simple construction, low detection limit, and good selectivity. The prepared sensors enabled the selective detection of MBC in an orange matrix in the presence of potential interferences, demonstrating their potential for practical applications in MBC residue monitoring. This study also demonstrated the successful application of these sensors for tracking the degradation kinetics of MBC in treated orange samples, providing valuable insights into the safety and preharvest interval considerations for citrus cultivation.

Overall, the development of these potentiometric sensors represents a significant advancement in analytical methodologies for monitoring pesticide residues in agricultural products. These sensors offer a rapid, cost-effective, and reliable alternative to conventional analytical techniques, with potential implications for food safety and regulatory compliance. They allow the measurement of MBC degradation kinetics in oranges without the need for sample extraction procedures and allow regular testing of orange shipments prior to export to the EU or USA to detect the presence of trace levels of MBC. This work represents the first potentiometric measurement of MBC and the first potentiometric analysis of MBC degradation kinetics.