Glutathione and Magnetic Nanoparticle-Modified Nanochannels for the Detection of Cadmium (II) in Cereal Grains

Abstract

We developed a novel and portable magnetic nanochannel electrochemical sensor for the sensitive detection of cadmium ions (Cd²⁺), which pose serious risks to food safety and human health. The sensor was fabricated by co-modifying an anodic aluminum oxide (AAO) nanochannel membrane with a composite of glutathione (GSH) and ferric oxide nanoparticles (Fe₃O₄), denoted as GSH@Fe₃O₄. This modified membrane was then integrated with a screen-printed carbon electrode (SPCE) to construct the GSH@Fe₃O₄/GSH@AAO/SPCE sensing platform. The performance of the sensor was evaluated using differential pulse voltammetry (DPV), which demonstrated a strong linear correlation between the peak current response and the concentration of Cd²⁺ in the range of 5–120 μg/L. The calibration equation was I_DPV(μA) = −0.31 + 0.98·C_{Cd²⁺(μg/L)}, with an excellent correlation coefficient (R² = 0.999, n = 3). The calculated limit of detection (LOD) was as low as 0.1 μg/L, indicating the high sensitivity of the system. These results confirm the successful construction of the GSH@Fe₃O₄/GSH@AAO/SPCE portable nanochannel sensor. This innovative sensing platform provides a rapid, sensitive, and user-friendly approach for the on-site monitoring of heavy metal contamination in agricultural products, especially grains.

1. Introduction

Heavy metal ions (such as cadmium, lead, and mercury) are major global environmental pollutants owing to their high toxicity and persistence. These ions enter the environment through various pathways such as industrial emissions, agricultural activities, transportation, and mining activities. Their presence in ecosystems causes severe ecological damage and poses serious risks to human health [1]. Cd²⁺ is a common heavy metal pollutant that exhibits bioaccumulation and long-term latency, progressively accumulating throughout the food chain and posing major threats to human health [2]. Cd²⁺ promotes apoptosis, oxidative stress, DNA methylation, and DNA damage [3]. The primary organs affected by Cd²⁺ toxicity are the kidneys, lungs, and bones [4]. This metal is also considered a potent carcinogen that affects the kidneys, lungs, pancreas, and prostate [5]. Therefore, the detection and monitoring of Cd²⁺ in food and the environment are crucial for mitigating its harmful effects.

Traditional analytical techniques, such as atomic absorption spectroscopy (AAS) [6] and inductively coupled plasma mass spectrometry (ICP-MS) [7], offer high sensitivity and accuracy, yet their complexity, high cost, and limited portability make them less suitable for field-based or real-time detection. In contrast, electrochemical sensors offer significant advantages such as portability, rapid analysis, low cost, and high sensitivity [8,9]. However, conventional electrochemical sensors often suffer from limited sensitivity and selectivity owing to the low surface area and insufficient number of active sites of traditional electrode materials.

To overcome these limitations, nanochannel-based electrochemical sensors have emerged as a promising strategy [10]. Anodic aluminum oxide (AAO) nanochannel membranes formed via electrochemical anodization offer uniform, vertically aligned, and size-tunable pores with high chemical stability and biocompatibility [11]. Their large specific surface area and confined mass transport properties enable signal amplification by enriching target analytes and reducing background interference [12,13]. Recent studies have demonstrated that nanochannel membranes can be directly integrated with screen-printed carbon electrodes (SPCE) to fabricate compact and high-performance electrochemical platforms [14].

In this study, the nanochannel membrane was chemically functionalized with glutathione (GSH), a tripeptide with carboxyl, amine, and thiol groups capable of forming strong coordination bonds with heavy metal ions, thus enhancing the selectivity and capture efficiency [15,16,17,18]. Moreover, magnetic Fe₃O₄ nanoparticles were introduced into the nanochannel membrane surface. These nanoparticles not only offer excellent conductivity and electron transfer properties but also allow external magnetic control, contributing to signal enhancement and membrane immobilization. Previous research has shown that Fe₃O₄ nanoparticles are effective modifiers in electrochemical sensors for detecting toxic metals such as Cu²⁺, Pb²⁺, and Cr²⁺ [19,20,21,22], typically through surface adsorption, redox facilitation, or peroxidase-like catalytic effects.

Compared with traditional bismuth-film-based electrodes, which are widely used for Cd²⁺ and Pb²⁺ detection owing to their low toxicity and favorable stripping voltammetry behavior, several recently reported sensors, such as RGO/Au-Bi-based systems, have achieved ultralow detection limits in the sub-ppb or nanomolar range [23]. While these systems offer excellent sensitivity, some still involve material synthesis steps or require pretreatment protocols that may limit their direct on-site application. In contrast, the GSH@Fe₃O₄/GSH@AAO-based system proposed in this study offers distinct advantages: (i) enhanced surface area and target selectivity via nanochannel confinement; (ii) dual functionality through glutathione-mediated recognition and conductive magnetic Fe₃O₄ support; and (iii) simplified modular integration with SPCE, facilitating cost-effective, reproducible, and field-deployable sensor fabrication without the need for additional post-processing.

Therefore, this study aimed to develop a nanochannel-based electrochemical sensing platform by fabricating a GSH@Fe₃O₄/GSH@AAO composite membrane and integrating it with the SPCE. The sensor was evaluated using differential pulse voltammetry (DPV) for Cd²⁺ detection in real samples, offering insights into the role of nanochannels and nanomaterials in advancing portable environmental sensing technologies.

2. Experimental Procedures

2.1. Materials and Apparatus

GSH was purchased from Aladdin Reagent Co., Ltd., Shanghai, China. Fe₃O₄ nanoparticles (average diameter: ~20 nm) were obtained from Macklin Biochemical Co., Ltd., Shanghai, China. The nanoparticles were well-dispersed in an aqueous solution and used without further surface modification. 3-Aminopropyltriethoxysilane (APTES) was obtained from Macklin Biochemical Co., Ltd., Shanghai, China, and N-hydroxysuccinimide (NHS) was obtained from Aladdin Reagent Co., Ltd., Shanghai, China. The MES buffer solution (pH 5.5) and cadmium standard solution (1000 μg/mL) were purchased from the National Nonferrous Metals and Electronic Materials Analysis and Testing Center, Beijing, China. Conductive adhesive K-181 was obtained from Zhuhai Jinsn Technology Co., Ltd., Zhuhai, China. The AAO membrane (5 mm diameter, 5000 nm thickness, pore diameter 390 nm, and pore spacing 450 nm) was purchased from Shenzhen Topology Precision Membrane Technology Co., Ltd., Shenzhen, China. The cost per membrane was approximately RMB 30 (USD 4–5), depending on the order quantity and specifications. The “flexible magnetic disk” refers to a 0.5 mm-thick magnetized PVC sheet (10 mm diameter) purchased from Shanghai Haoqi Magnetics Co., Ltd., Shanghai, China. It serves as a mechanical support and allows magnetic fixation during sensing without participating in the electrochemical process. The screen-printed electrode system, including a platinum wire auxiliary electrode, Ag/AgCl reference electrode, and carbon-based working electrode, was purchased from Qingdao Boron Carbon Technology Co., Ltd., Qingdao, China. The barley, wheat, and corn samples were purchased from a local agricultural market in Nanjing, China.

The following instruments were used: a PalmSens4 electrochemical workstation, PalmSens, Utrecht, The Netherlands. SB25-12DT ultrasonic cleaner, Ningbo Xinzhi Biotechnology Co., Ltd., Ningbo, Zhejiang, China. DHG-9013A drying oven, Shanghai Yiheng Scientific Instrument Co., Ltd., Shanghai, China. TM3000 SEM, Hitachi Ltd., Tokyo, Japan, and contact angle goniometer, Ningbo New Boundary Scientific Instrument Co., Ltd., Ningbo, Zhejiang, China. All the electrochemical experiments were conducted at room temperature (25 °C).

2.2. Modification of Porous AAO with GSH

The AAO membrane was ultrasonically treated with a 0.25% NaOH (wt.) solution for 3 min to enhance the hydrophilicity. The membrane was then soaked in a 5% (v/v) APTES anhydrous ethanol solution for 2 h to form siloxane bonds, washed with anhydrous ethanol, and incubated in 0.1 M N-hydroxysuccinimide MES buffer solution for 2 h to activate the surface hydroxyl groups. The activated membrane was further incubated in a 5 mM GSH solution for 1 h, which allowed the GSH to firmly bind to the membrane surface, thus forming a GSH@AAO membrane.

2.3. Surface Modification of Fe₃O₄ Nanoparticles with GSH

First, 1.0 g of Fe₃O₄ nanoparticles was mixed with 10 mL of anhydrous ethanol in a flask. The nanoparticles were uniformly dispersed via stirring in ethanol. Subsequently, 30 mL of water was added to the nanoparticles, and the mixture was ultrasonically treated at 25 °C for 2 h. Later, 0.4 g of reduced GSH was added to the mixture, and ultrasonic treatment was continued for 1 h. The GSH@Fe₃O₄ nanoparticles were collected using a magnet and washed repeatedly with deionized water. Finally, the nanoparticles were dried in a vacuum at 60 °C for 8 h to obtain a GSH@Fe₃O₄ nanoparticle suspension.

2.4. Preparation of the GSH@Fe₃O₄/GSH@AAO Membrane

First, 10 μL of the GSH@Fe₃O₄ nanoparticle suspension was applied onto one side of the GSH@AAO membrane. The membrane was then placed in an oven at 50 °C and dried for 30 min to prepare a functionalized GSH@Fe₃O₄/GSH@AAO composite membrane.

2.5. Preparation of GSH@Fe₃O₄/GSH@AAO/SPCE

A 0.5 mm-thick flexible magnetic sheet was cut into 3 mm-diameter disks by using a hole puncher. A flexible magnetic disk was attached to the working electrode using a silver conductive adhesive. The prepared GSH@Fe₃O₄/GSH@AAO composite membrane was placed onto the working electrode, ensuring that the GSH@Fe₃O₄-modified side faced the silver conductive adhesive. To ensure structural integrity and prevent leakage in practical applications, the edges of the porous AAO membrane were sealed with a silicone sealant and allowed to dry at room temperature for 2 h. This assembly formed a nanochannel-based screen-printed electrochemical sensor (GSH@Fe₃O₄/GSH@AAO/SPCE) in which the nanochannels in the AAO membrane effectively captured the target molecules. The magnetic materials used ensured that the Fe₃O₄ nanoparticles were securely attached to the membrane, thereby enhancing the stability and durability of the composite material. Scheme 1 outlines the steps followed for fabricating the GSH@Fe₃O₄/GSH@AAO/SPCE.

The sensor was tested using a DPV system with a portable EmStat4R electrochemical workstation (PalmSens, The Netherlands). During the test, the sensor was exposed to a solution containing Cd²⁺ ions, with the ions reduced to a specific potential. The current response was recorded using EmStat4R, thus allowing the quantitative detection of Cd²⁺. The performance and sensitivity of the sensor were evaluated by comparing the electrochemical behavior of buffer solutions with different Cd²⁺ concentrations.

2.6. Preparation and Pretreatment of Real Samples

Barley, wheat, and corn samples were obtained from a local market and milled into fine powders using a high-speed grinder. Precisely 2.0 g of each sample was mixed with 10 mL of 0.1 M acetate buffer (pH 5.0) and subjected to ultrasonic extraction for 30 min at room temperature. After centrifugation at 8000 rpm for 10 min, the supernatant was collected and filtered through a 0.22 μm membrane to remove particulates. The resulting extract was diluted threefold with the same buffer prior to the electrochemical analysis. For recovery assessment, defined concentrations of the Cd²⁺ standard solution were spiked into the treated samples under identical conditions.

3. Results and Discussion

3.1. Characterization of the GSH@Fe₃O₄/GSH@AAO Membrane

Figure 1A shows an SEM image of the bare AAO membrane at 10,000× magnification, displaying a well-ordered and uniform nanochannel array. After GSH modification, the pore structure remained intact (Figure 1B) while a slight change in the surface contrast was observed, which may indicate the presence of an organic molecular layer. Owing to the resolution limitations at this magnification, the GSH modification layer is not directly visible, but is inferred from surface texture differences.

Figure 1C presents the SEM image of the original Fe₃O₄ nanoparticles at 80,000× magnification, showing predominantly spherical particles with a uniform size distribution and minimal aggregation. The observed morphology and particle size (~20 nm in diameter) were consistent with the manufacturer’s specifications.

Figure 1D shows the surface morphology of the GSH@Fe₃O₄ after surface modification. Compared to the bare and GSH-modified AAO, the surface becomes rougher and shows aggregated particulate structures, indicating the successful immobilization of GSH@Fe₃O₄ nanoparticles. Although individual nanoparticles were not fully resolved due to overlapping and partial aggregation, morphological changes were evident.

To confirm surface composition, energy-dispersive X-ray spectroscopy (EDS) was performed at selected regions (Figure 1D), and the spectra (Figure 1E) revealed that iron is the dominant element (>94 wt%) with minor contributions from aluminum and silicon, likely from the AAO membrane and the substrate, respectively. These findings validated the effective immobilization of Fe₃O₄ on the membrane surface.

Figure 1F presents the FT-IR spectra (32 scans, 4 cm⁻¹ resolution) of AAO, GSH@AAO, and GSH@Fe₃O₄/GSH@AAO membranes. The GSH@AAO membrane (curve a) exhibited peaks at 1650 and 1569 cm⁻¹, indicating the presence of amide bonds (C=O and N–H) from GSH. The GSH@Fe₃O₄/GSH@AAO membrane (curve b) showed peaks at 1162 and 1038 cm⁻¹, corresponding to C–O or C–N bonds, and Fe–O bonds between 1000 and 500 cm⁻¹, thereby confirming that Fe₃O₄ was successfully attached [24]. The AAO membrane (curve c) displayed a flat spectrum with no significant peaks, exhibiting only the Al–O bond characteristics.

Figure 1G shows the XRD patterns of AAO, GSH@AAO, and GSH@Fe₃O₄/GSH@AAO membranes. The bare AAO membrane exhibited characteristic peaks at 2θ ≈ 37.0° and 45.0°, corresponding to γ-Al₂O₃ (JCPDS No. 10-0425). In the GSH@AAO membrane, the γ-Al₂O₃ peaks remained but showed decreased intensity, possibly due to lattice distortions induced by GSH modification. The GSH@Fe₃O₄/GSH@AAO membrane displayed additional peaks at 2θ = 30.2°, 35.6°, 43.3°, 57.3°, and 62.9°, which correspond to the (220), (311), (400), (511), and (440) planes of Fe₃O₄ (JCPDS No. 19-0629), confirming the successful deposition of Fe₃O₄ [25]. The simultaneous presence of both Fe₃O₄ and γ-Al₂O₃ peaks indicates a mixed-phase structure, with weakened alumina signals resulting from partial coverage and lattice interactions.

Together, these findings suggest the effective surface modification of the AAO membrane with both GSH and Fe₃O₄ nanoparticles, as indicated by the morphology, elemental composition, and spectral features.

3.2. Electrochemical Characterization of the GSH@Fe₃O₄/GSH@AAO/SPCE Sensor

The AAO membrane was attached to the SPCE using a silver conductive adhesive and sealed with silicone to obtain the AAO/SPCE configuration. CV was conducted in a 3 mM potassium ferricyanide solution at a scan rate of 100 mV/s. As shown in Figure 2, the redox current gradually increases over successive scans and stabilizes after five cycles. This behavior suggests progressive wetting and electrochemical equilibration within the AAO nanochannels.

The CV curves exhibited distinct oxidation and reduction peaks, indicating that redox reactions of potassium ferricyanide occurred at the electrode interface. The increase in the peak current over the initial cycles may be attributed to the gradual infiltration and adsorption of electroactive species into the confined nanochannel structure. After stabilization, the peak positions showed minor shifts, likely due to interfacial polarization or changes in the local ion concentration. The data presented in the subsequent figures (e.g., Figure 3 and Figure 4) correspond to the fifth scan after the system reaches an electrochemical steady state.

The electrochemical behavior of the GSH@Fe₃O₄/GSH@AAO/SPCE sensor was further investigated by recording cyclic voltammograms at different scan rates (20–140 mV/s) in a 3 mM potassium ferricyanide solution. Figure 3 presents the CV responses obtained from the fifth scan after signal stabilization. Clear redox peaks were observed at all scan rates, with the peak currents increasing proportionally with the scan rate. A linear relationship was found between the peak current (Ip) and scan rate (v), with R² = 0.967 for oxidation and 0.983 for reduction. According to electrochemical theory, such Ip–v linearity is indicative of a surface-confined (adsorption-controlled) redox process rather than a diffusion-controlled process, which typically exhibits a linear Ip–$\surd v$ relationship. The confined geometry of the nanochannels and presence of thiol-rich glutathione may contribute to this adsorption-driven electrochemical behavior. Additionally, the peak potential shifts with an increasing scan rate suggest a kinetic contribution to the electron transfer process.

Figure 4A compares the CV responses of the bare SPCE, GSH@AAO/SPCE, GSH@Fe₃O₄/SPCE, and GSH@Fe₃O₄/GSH@AAO/SPCE electrodes in a potassium ferricyanide solution. It is important to note that the bare SPCE was tested without a magnetic strip, whereas a small magnet was placed beneath each modified electrode during the measurement to stabilize the magnetic components.

The bare SPCE exhibited distinct, yet relatively low redox peak currents, attributable to its limited electroactive surface area. Upon modification, particularly with both GSH-functionalized AAO and Fe₃O₄ nanoparticles, the peak currents increased significantly. This enhancement can be ascribed to improved electron transfer, increased surface area, and recognition sites introduced by the AAO nanochannels. Among all the electrodes, the dual-modified electrode exhibited the highest current response, indicating a synergistic effect between the nanochannel confinement and the electrical conductivity of the magnetic nanoparticles.

Figure 4B presents the Nyquist plots of SPCE, GSH@AAO/SPCE, GSH@Fe₃O₄/SPCE, and GSH@Fe₃O₄/GSH@AAO/SPCE electrodes recorded in a 3.0 mM [Fe(CN)₆]³⁻/⁴⁻ solution containing 0.1 M KCl. The electrochemical impedance spectroscopy (EIS) measurements were performed by applying a sinusoidal AC voltage with an amplitude of 5 mV over a frequency range from 100 kHz to 0.1 Hz, with a DC bias of +0.2 V (vs. Ag/AgCl). The impedance spectra were fitted using a modified Randles equivalent circuit model (as shown in the inset of Figure 4B), which includes solution resistance (R1), double-layer capacitance (C1), charge transfer resistance (R2), and Warburg impedance (W1), representing the diffusion process. The bare SPCE displayed a large semicircular region in the high-frequency range, indicating a high charge transfer resistance (Rct), likely due to the limited electroactive surface and poor intrinsic conductivity of the unmodified carbon surface. In contrast, both GSH@AAO/SPCE and GSH@Fe₃O₄/SPCE exhibited markedly reduced semicircle diameters, suggesting improved electron transfer kinetics. The enhancement can be attributed to the electron-donating groups provided by glutathione on the AAO membrane and the conductive and redox-active nature of the Fe₃O₄ nanoparticles. Notably, the semicircle of GSH@AAO/SPCE is smaller than that of GSH@Fe₃O₄/SPCE, which may be attributed to the additional effect of the silver conductive adhesive in the AAO composite assembly.

Among all configurations, the GSH@Fe₃O₄/GSH@AAO/SPCE exhibited the smallest semicircle, corresponding to the lowest Rct. This observation supports a synergistic improvement in the charge transfer properties arising from the combination of nanochannel confinement, surface functionalization, and nanoparticle conductivity enhancement.

The effective electroactive surface area (ECSA) of the modified electrode is determined using the Randles–Sevcik formula, as shown in Equation (1):

$$I_p= 0.4463nFAC\sqrt{{\displaystyle \frac{nFDv}{RT}}}$$

(1)

where n denotes the number of electrons in the electrochemical reaction, I_p represents the peak voltammetric current of the electrochemical process, F stands for the Faraday constant (96,485 C/mol), v represents the applied voltammetric scanning rate (V/s), T indicates the Kelvin temperature (273.15), R signifies the universal gas constant (8.314 J/mol·K), A denotes the electrically active area of the electrodes (cm²), and D represents the diffusion coefficient (cm²/s).

Based on this calculation, the ECSA of the GSH@Fe₃O₄/GSH@AAO/SPCE was estimated to be 4.1 × 10⁻⁴ cm², while that of the unmodified SPCE was only 9.7 × 10⁻⁵ cm². Although the geometric area of the SPCE working electrode was approximately 0.071 cm² (3 mm diameter), the effective electroactive area was significantly lower, likely owing to surface passivation and the limited accessibility of conductive pathways in the untreated screen-printed carbon surface. In contrast, the increased ECSA observed in the modified electrode suggests that the introduction of AAO nanochannels and Fe₃O₄ nanoparticles provides additional active sites and enhances electron transfer efficiency. These findings highlight the potential of nanochannel-based modifications to improve the interfacial properties of electrochemical sensors.

3.3. Electroanalytical Characteristics of Cd²⁺ on Different Electrodes

Figure 5 presents the DPV response curves of the different modified electrodes to Cd²⁺ in an acetate–acetic acid buffer solution. The electrochemical performance of these electrodes was assessed by analyzing their curves. The SPCE (bare electrode) exhibits a relatively small current response with an indistinct peak. This indicates the basic performance of the unmodified electrode with low sensitivity for Cd²⁺ detection. It should be noted that the bare SPCE was tested without a magnetic disk, whereas all other electrodes were tested under magnetic fixation to maintain the Fe₃O₄ components. The GSH@AAO/SPCE exhibited a significantly increased current response with a more distinct peak, which suggested that the GSH-modified AAO membrane improved the sensitivity of the electrode to Cd²⁺ detection; the nanochannel structure of the AAO membrane likely provided more active sites, thus augmenting the electrochemical performance. The GSH@Fe₃O₄/SPCE displayed a slightly increased current response, indicating that the Fe₃O₄ nanoparticles alone exerted a limited effect on Cd²⁺ detection. While the Fe₃O₄ nanoparticles may exhibit some electrochemical activity, their distribution and effective surface area might be constrained. GSH@Fe₃O₄/GSH@AAO/SPCE displayed the highest current response with a considerably more pronounced peak than the other electrodes, indicating that dual modification (GSH@Fe₃O₄ and GSH@AAO) significantly boosted the sensitivity of the electrode to Cd²⁺ detection.

3.4. Optimization of Experimental Parameters for the GSH@Fe₃O₄/GSH@AAO/SPCE Electrochemical Sensor

3.4.1. Optimization of Hydrophilicity of the AAO Membrane

To evaluate the wetting behavior and potential electrolyte infiltration capability of AAO nanochannels, contact angle measurements were performed under different surface treatment conditions. The improved hydrophilicity is expected to facilitate ion transport and enhance the electrochemical response of nanochannel-based sensors.

The porous AAO membrane (pore diameter: 390 nm) was treated with 0.25% NaOH under ultrasonic agitation and the contact angles were measured at different treatment times (Figure 6A). The untreated membrane had a contact angle of 57.6°, suggesting partial hydrophobicity. After 1 min of ultrasonic treatment, the angle decreased to 51.0° and further to 49.9° after 2 min, indicating increased hydrophilicity. At 3 min, the angle significantly decreased to 16.7°, likely owing to pore wall exposure and surface hydroxylation. These results demonstrate that a mild alkaline ultrasound treatment substantially improves the AAO surface wettability.

Although the contact angle of the fully modified sensor (GSH@Fe₃O₄/GSH@AAO) was not directly measured, it was expected to remain low or even decrease slightly, given that both GSH and Fe₃O₄ nanoparticles possess hydrophilic functional groups (e.g., thiol, carboxyl, and hydroxyl), which may further promote electrolyte penetration. The overall improved surface wettability supports efficient ionic diffusion into the nanochannels and contributes to the enhanced electrochemical performance observed in the later sections.

These findings emphasize the importance of optimizing surface wettability in nanochannel-based electrochemical sensors, where efficient electrolyte infiltration is crucial for maximizing sensitivity and signal stability.

3.4.2. Optimization of pH for DPV

The DPV response curves for 20 μg/L Cd²⁺ were recorded in acetate–acetic acid buffer solutions of pH (3.5~6.5). The DPV parameters were set as follows: calibration time of 1 s, testing voltage from −0.5~0 V, step potential of 0.01 V, and effective sweep rate of 0.002 V/s. In this study, the term “effective sweep rate” refers to the ratio of the step potential (0.01 V) to the pulse period applied in the DPV measurement, serving as an analog to scan rate for comparison purposes. As shown (Figure 6B), both the peak current intensity and potential varied significantly with the pH. Among the tested conditions, pH 5.0 yielded the highest peak current, indicating optimal sensitivity for Cd²⁺ detection under these conditions.

The enhanced signal at pH 5.0 is likely due to the optimal protonation state of the GSH functional groups on the electrode surface. At lower pH values (<4.5), excess H⁺ ions can protonate the carboxylate and thiol groups of GSH, thereby reducing its ability to chelate Cd²⁺ via electrostatic or covalent interactions. At a higher pH (>6), the deprotonation of functional groups may increase, but competition from hydrolyzed Cd species (e.g., Cd(OH)⁺, Cd(OH)₂) could reduce free Cd²⁺ availability and binding efficiency. Thus, pH 5 provides a balance where GSH remains sufficiently deprotonated to interact with Cd²⁺, while minimizing interference from Cd hydrolysis.

Notably, the peak potential remained relatively stable across the tested pH range, despite significant variations in the current intensity. This electrochemical invariance suggests that the redox process of Cd²⁺ at the GSH@Fe₃O₄/GSH@AAO/SPCE interface is not predominantly governed by the proton-coupled electron transfer (PCET), but rather by a surface-confined coordination mechanism [26]. In addition, the constrained geometry of the nanoporous AAO structure may facilitate a local buffering effect, wherein the microenvironment within the nanochannels maintains a quasi-stable pH, thereby minimizing the influence of the bulk solution pH on the redox thermodynamics [27]. Similar behavior has been reported in other nanostructured electrochemical systems, where ion transport and interfacial electrostatics are modulated by nanoscale confinement, resulting in stable redox potentials. These findings underscore the dual role of pH in regulating the Cd²⁺ binding efficiency while having a limited impact on electron transfer kinetics at the modified electrode surface.

3.4.3. Optimization of the Effective Sweep Rate for DPV

The effective sweep rate is a key parameter in DPV that influences signal intensity, peak sharpness, and background noise. To identify the optimal effective sweep rate for Cd²⁺ detection, DPV measurements were performed in acetate buffer (pH 5.0) containing 20 μg/L Cd²⁺, using an effective sweep rate from 0.002~0.01 V/s (Figure 6C). The other DPV parameters were fixed: step potential of 0.01 V, testing voltage from −0.5~0 V, and equilibration time of 1 s without deposition.

As shown in the figure, the peak current gradually decreases with an increasing effective sweep rate. At the lowest effective sweep rate of 0.002 V/s, the response signal was the highest and the most defined. This can be attributed to the longer interaction time between the analyte and the sensing interface, which allows for more efficient adsorption and electron transfer. In contrast, a higher effective sweep rate reduces the time available for faradaic processes, thereby lowering the current response and broadening the peaks due to kinetic limitations.

Based on these results, an effective sweep rate of 0.002 V/s was selected as optimal for subsequent measurements, providing a balance between the sensitivity and signal resolution in the absence of a deposition step.

3.5. Detection of Cd²⁺ Using the GSH@Fe₃O₄/GSH@AAO/SPCE Electrochemical Sensor

Acetate–acetic acid buffer solutions of Cd²⁺ were prepared at concentrations of 5, 10, 20, 40, 60, 80, 100, and 120 μg/L. Then, 100 μL of each solution was applied to the GSH@Fe₃O₄/GSH@AAO/SPCE membrane electrode prepared as described in Section 2. The electrode was connected to a portable electrochemical workstation and Cd²⁺ was detected using the DPV method. The DPV parameters were set as follows: calibration time, 1 s; testing voltage, −0.5 V~0 V; step potential, 0.01 V; and scan rate, 0.002 V/s. The pH of the buffer solution was maintained at five.

The different Cd²⁺ concentrations (5–120 μg/L) correspond to the curves of different colors, thereby demonstrating a clear concentration dependency. As the Cd²⁺ concentration increased, the peak current steadily increased (Figure 7). The inset displays the relationship between the peak current and the Cd²⁺ concentration, showing a clear linear relationship. The linear regression equation was as follows: I_DPV(μA) = −0.31 + 0.98 C_{Cd²⁺(μg/L)}, with a correlation coefficient (R²) of 0.999, where I_DPV(μA) and C_{Cd²⁺(μg/L)} represent the DPV response peak current and Cd²⁺ concentration, respectively. Using the formula LOD = 3 s/b (where s is the standard deviation of the blank sample and b is the slope of the calibration curve), the detection limit was calculated as 0.1 μg/L. Similarly, the limit of quantification (LOQ), calculated as 10 s/b, was determined to be 0.33 μg/L. These results demonstrate the viability of using the constructed electrochemical sensor to detect Cd²⁺.

Table 1. Comparison of different methods and materials for Cd²⁺ detection.

Electrodes	Technique	Linear Range (μg L−1)	LOD (μgL−1)	Reference
Ru-GO/Nafion screen-printed gold electrode	SWASV	50~350	4.2	[28]
Ru-Au/Nafion screen-printed gold electrode	SWASV	10~300	12.01	[28]
UiO-66-NH2/GA	DPSV	18.5~925	6.2	[29]
CeMOF@MWCNTs/CC	DPASV	10~1200	2.2	[8]
RGO/Au-Bi	DPASV	0.1~300	0.02	[23]
N, S-PCNFs	DPASV	2~500	0.7	[30]
Mo-WO3/CC	DPV	0.1~100	0.017	[31]
SPGE-Nafion/Bi	SWASV	20~300	4	[32]
BiSP/SPE	DPASV	0.9~25.0	0.3	[33]
GSH@Fe3O4/GSH@AAO/SPCE	DPV	5~120	0.1	This work

compares the key sensing characteristics, such as the linear range and detection limit, of several platforms developed based on various modified electrode materials [8,28,29,30,31,32,33]. These findings confirm the practical value of the proposed modifiers for the detection of Cd²⁺.

Detecting Cd²⁺ in real samples with high selectivity and without significant interference is challenging. To address this, the influence of common coexisting ions was evaluated using 20 μg/L of Cd²⁺. Even in the presence of a 20-fold excess of Na⁺, Mg²⁺, Cu²⁺, Pb²⁺, and Hg²⁺, no significant variation in the Cd²⁺ signal was observed (Figure 8A), indicating the strong selectivity of the electrode. The lack of interference from Cu²⁺, despite its known affinity for thiol groups, may be due to its distinct redox potential, which prevents overlap with Cd²⁺ in the scanning window. Similarly, although Pb²⁺ shares a comparable ionic radius and often coexists with Cd²⁺ in environmental matrices, the sensor response remains stable, suggesting selective recognition at the sensing interface.

The high selectivity of the GSH@Fe₃O₄/GSH@AAO/SPCE sensor toward Cd²⁺ over other potentially interfering metal ions such as Cu²⁺, Pb²⁺, and Zn²⁺ can be attributed to a combination of thermodynamic and kinetic factors under specific sensing conditions. First, although GSH contains thiol, amino, and carboxyl groups capable of coordinating multiple heavy metal ions, the relative stability of the resulting complexes varies with the pH and ionic radius. Studies have shown that under mildly acidic conditions (pH 4.5~5.5), the stability constant (log K) of the Cd–GSH complex is higher than that of GSH complexes with Zn²⁺ or Pb²⁺, and the reduction potential of Cd²⁺ is more negative, which aligns with the DPV window applied in this work (−0.5~0 V) [25].

Secondly, Cu²⁺, which has a high affinity for thiol ligands, often undergoes redox cycling near 0 V, which is kinetically fast and overlaps with background processes. However, in our system, the electrochemical window avoids the Cu²⁺/Cu⁺ redox pair (typically around 0.15~0.2 V), preventing its interference. This is supported by our experimental data (Figure 8A), in which even a 20-fold excess of Cu²⁺ did not significantly alter the Cd²⁺ signal.

Moreover, the confined geometry of the AAO nanochannels plays a critical role in enhancing selectivity. The spatial constraint and electrostatic environment favor smaller ions with a suitable hydration radius and charge density. Cd²⁺, which has a moderate ionic radius (0.97 Å) and stable coordination geometry, diffuses more efficiently into the nanochannels than the larger Pb²⁺ (1.19 Å) or smaller but more tightly hydrated Zn²⁺ (0.74 Å), leading to preferential enrichment. These effects are further magnified by GSH modification on both Fe₃O₄ and AAO interfaces, which promotes multisite binding and preconcentration [34,35,36]. This synergy between ligand chemistry and nanostructure-induced confinement provided a robust explanation for the experimentally observed high selectivity toward Cd²⁺.

In addition, the reproducibility of the sensor was evaluated by testing five independently prepared electrodes using a 20 μg/L Cd²⁺ solution. The relative standard deviation (RSD) of the peak current was 3.37% (Figure 8B), indicating good fabrication reproducibility. Furthermore, five consecutive measurements using a single electrode yielded an RSD of 4.4% (Figure 8C), demonstrating acceptable operational stability. After storage at 4 °C for 33 d, the sensor retained 76.6% of its initial response, indicating reasonable long-term stability (Figure 8D). Although only five consecutive measurements were conducted, the stable current responses indicated that the sensor could be reused for at least five cycles without any notable performance loss. Based on the current results, the sensor is expected to maintain an acceptable stability for short-term repeated use. However, further investigation is required to determine the maximum number of operational cycles before replacement or recalibration is required. For field applications, it may also be designed as a single-use disposable device to avoid fouling and to simplify operation.

Although the sensor was not specifically designed for regeneration, the relatively stable responses across consecutive uses suggest a degree of resistance to fouling. In future work, regeneration strategies, such as gentle rinsing or electrochemical cleaning, will be explored to further extend sensor reusability in complex food matrices.

3.6. Analysis of Real Samples

To evaluate the practical applicability of the developed sensor, it was used to determine Cd²⁺ in barley, wheat, and corn samples. As shown in

Table 2. Recovery of Cd²⁺ in spiked cereal samples (n = 3). Measured values are presented as mean ± standard deviation. AAS was used as a reference method for result validation.

Sample	Quantity Added (μg/kg)	Measured Value (μg/kg)	Measured by AAS (μg/kg)	Recovery Rate (%)
Barley	50	48.5 ± 1.5	47.5	97
	100	97.5 ± 2	96.5	97.5
Wheat	50	54 ± 1	53	108
	100	106 ± 1.5	103.5	106
Corn	50	52.5 ± 1.5	51	105
	100	102 ± 1	99	102

, the measured values are presented as the mean ± standard deviation (n = 3), and the recovery rates for Cd²⁺ ranged from 97% to 108% after spiking known concentrations into the pretreated extracts. Furthermore, the results obtained with the proposed Fe₃O₄/GSH@AAO/SPCE sensor were consistent with those measured by atomic absorption spectroscopy (AAS), confirming the accuracy and reliability of the sensor in detecting trace levels of Cd²⁺ in complex food matrices. Although electrochemical analyses were conducted using filtered and diluted sample extracts, the high recovery rates and consistency with AAS measurements suggest that matrix effects were effectively minimized under the present experimental conditions. The sample preparation steps, including centrifugation and membrane filtration, were designed to remove particulate matter and reduce interference from organic compounds, such as proteins and polyphenols, commonly found in food matrices. According to EU regulations, the maximum permissible concentration of Cd²⁺ in cereals, such as wheat and barley, is 100 μg/kg (dry weight basis). The Cd²⁺ levels detected in this study, including those in spiked samples, remained below this limit after conversion to μg/kg, thereby demonstrating that the proposed electrochemical sensor is suitable for monitoring Cd contamination in food matrices in accordance with regulatory standards.”

In future studies, we plan to investigate the performance of the sensor in raw or minimally processed grain matrices to further evaluate its antifouling ability and suitability for direct on-site detection without extensive pretreatment.

3.7. Scalability and Reproducibility Assessment

The practical deployment of electrochemical sensors requires not only high analytical performance but also scalable and reproducible fabrication. In this study, the construction of the GSH@Fe₃O₄/GSH@AAO/SPCE sensor involved a series of wet-chemical modifications—including glutathione functionalization of the AAO membrane and ferric oxide nanoparticle integration—followed by physical assembly onto the screen-printed carbon electrode. Although the process includes multiple steps, each step is based on well-established protocols that can potentially be parallelized for batch production.

Vacuum drying and manual attachment were used in this proof-of-concept study to ensure structural integrity and proper membrane electrode alignment. However, these steps can be further streamlined using mold-guided membrane placement, robotic dispensing, or modular electrode platforms, which could support scale-up manufacturing.

Furthermore, the sensor exhibited good fabrication reproducibility, as demonstrated by the low relative standard deviation (RSD) of 3.37% across five independently prepared electrodes (Figure 8B). The consistency in the electrode performance and modular nature of the assembly support the feasibility of standardizing the process for semi-automated or commercial-scale sensor production, reinforcing the platform′s applicability in real-world settings.

4. Conclusions

In this study, a nanochannel-based electrochemical sensor was successfully developed by the dual modification of an AAO membrane with GSH and Fe₃O₄ nanoparticles, forming a GSH@Fe₃O₄/GSH@AAO composite membrane. This composite significantly enhanced the electrochemical reactivity and sensitivity of the electrode. After optimization using DPV, the sensor exhibited a strong linear response toward Cd²⁺ in the range of 5~120 μg/L with a detection limit as low as 0.1 μg/L under pH 5 conditions. The dual modification strategy improved electron transfer and increased the number of active sites, producing a pronounced synergistic effect. The sensor demonstrated a high analytical performance for trace Cd²⁺ detection, offering a promising platform for environmental and food safety monitoring.

Despite these advantages, the current fabrication process involves multiple modifications and assembly steps, which may present challenges for large-scale production. However, the reliance on established wet chemistry and membrane integration techniques indicates the potential for future adaptation to automated or modular fabrication systems. Future research will focus on simplifying the electrode preparation process, enhancing the antifouling capacity, and extending the applicability of this sensor to other heavy metal ions in complex food and environmental matrices. In particular, the performance of the sensor in unprocessed or high-organic-content food matrices will be further explored to assess its practical utility under real field conditions.