Methylene-Blue-Encapsulated Metal-Organic-Framework-Based Electrochemical POCT Platform for Multiple Detection of Heavy Metal Ions in Milk

Considering the high risk of heavy metal ions (HMIs) transferring through the food chain and accumulating in milk, a flexible and facile point-of-care testing (POCT) platform is urgently needed for the accurate, sensitive, and highly selective on-site quantification of multiple HMIs in milk. In this work, a cost-effective disk with six screen-printed electrodes (SPEs) was designed for hand-held electrochemical detection. Metal organic frameworks (MOFs) were adopted to amplify and enhance the electrochemical signals of methylene blue (MB). Using differential pulse voltammetry (DPV) methods, low limits of detection for four HMIs (Cd2+, 0.039 ppb; Hg2+, 0.039 ppb; Pb2+, 0.073 ppb; and As3+, 0.022 ppb) were achieved within four minutes. Moreover, the quantitative POCT system was applied to milk samples. The advantages of low cost, ease of on-site implementation, fast response, and accuracy allow for the POCT platform to be used in practical monitoring applications for the quantitation of multiple HMIs in milk samples.


Introduction
With the improvements in quality of life and the development of industrialization, the risk of contaminating nutritious milk products has gradually gained attention. In particular, heavy metal ions (HMIs) in milk have become a significant problem for food safety [1]. HMIs often exist in raw milk and milk powder, which may enter the human body through ingestion and cause considerable harm owing to recalcitrance, high toxicity, and facile enrichment [2,3]. The maximum permitted concentrations of Hg 2+ , Pb 2+ , and As 3+ in raw milk are limited to 0.01, 0.05, and 0.1 mg/kg by the National Standard of the People's Republic of China (GB 2762-2012), respectively. Conventional detection methods for HMIs mainly include atomic absorption spectrometry (AAS) [4], inductively coupled plasma mass spectrometry (ICP-MS) [5], atomic fluorescence spectrometry (AFS) [6], surface-enhanced Raman spectroscopy (SERS) [7,8], immunoassays [9], and fluorescence (FL) methods [10]. However, these techniques rely on expensive instruments and are too sophisticated to support rapid point-of-care testing (POCT). Hence, HMI sensing devices that are portable, simple to use, and cost-effective must be developed.
POCT allows non-specialists to quickly perform on-site analysis, using low and disposable materials compared with tests in laboratories [11]. POCT devices, such as microfluidic platforms [12,13] and paper-based sensors [2], have the advantages of facile operation, low material consumption, rapid output, and versatile application with existing infrastructure [14]. Therefore, POCT devices have been adopted in food safety analysis [15], clinical diagnosis [16], and environmental monitoring [17].
Integrated electrochemical POCT sensing devices have received attention because of their small size and high sensitivity [18]. Furthermore, the electrochemical POCT devices produce direct electrical signals based on the oxidation-reduction reaction. Therefore, it The disks containing six SPEs, a PET substrate, and a polydimethylsiloxane (PDMS) insulating loop were obtained from Weihai Poten Technology Co., Ltd. (Weihai, China). The six SPEs included four carbon electrodes as the working electrodes, Ag/AgCl as the reference electrode, and a carbon electrode as the counter electrode. DPV was performed using a BIOSYS p15e max biosensor system from Shenzhen Refresh Biosensing Technology Co., Ltd. (Shenzhen, China). This hand-held electrochemical station was driven by a smartphone installed with an application that uses Blue tooth. CV and EIS were conducted using a CHI660E electrochemical station (Chenhua, Shanghai, China). The morphologies and energy-dispersive X-ray (EDX) patterns of MOFs and MB@MOFs were obtained through high-resolution transmission electron microscopy (HRTEM) using an HT7700 Exalens microscope (Hitachi, Tokyo, Japan) at an accelerating voltage of 120 kV. X-ray diffraction (XRD) patterns were obtained using a D2 PHASER diffractometer (BRUKER, Mannheim, Germany). Adsorption-desorption isotherms were assessed using an automated surface area and porosity analyzer (Quantachrome, Boynton Beach, FL, USA).

Preparation of AuNPs
The synthesis of AuNPs was conducted using a citrate reduction method, in accordance with previous reports [49]. Briefly, 500 µL of 2% HAuCl 4 aqueous solution and 100 mL of distilled water were mixed in a flask and boiled under constant stirring. Next, 500 µL of 1% sodium citrate was added and the mixture was stirred for an additional 20 min. The wine-red AuNP solution was cooled to room temperature and stored at 4 • C.

Preparation of Zr-MOFs
UiO-66-NH 2 was synthesized using the procedure reported in a previous report with a slight modification [50]. First, 120 mg of ZrCl 4 , 1900 mg of benzoic acid, and 110 mg of 2-aminoterephthalic acid (NH 2 -BDC) were dissolved in 10 mL of DMF under ultrasonication for 3 min. The dispersion was then reacted at 120 • C for 24 h. Next, the precipitate was obtained after washing twice with DMF and methanol. Finally, the Zr-MOF precipitates were dried at 65 • C for 12 h and kept in a desiccator.

Preparation of MB@Zr-MOFs
First, 1 mg of MB was mixed with 10 mg of Zr-MOFs in 1 mL of distilled water using a rotary mixer at room temperature for 24 h [42]. Then, the precipitate was centrifuged and washed three times with ultrapure water. The resulting blue powder was dried at 65 • C overnight and stored in the dry environment for further use.

Preparation of MB@Zr-MOFs-apt and MB-apt
Here, 300 µL of 10 µM aptamers targeting different HMIs was activated with 5 µL of 1 M TCEP for 30 min at room temperature to convert the disulfide bonds to thiols. Then, 800 µL of 2.5 µM aptamers with carboxyl groups was mixed with 400 µL of EDC/NHS (400 mM/100 mM) in 10 mM PBS solution (pH 7.4) [51]. The mixture was stirred for 30 min to activate the aptamer carboxy groups. Next, the amide reaction was conducted with the addition of 800 µL of 1 mg/mL MB@Zr-MOFs for 2 h. The light blue precipitate was obtained after centrifugation at 12,000 rpm for 20 min. Finally, the MB@Zr-MOFs-apt complex was attained and redispersed in 800 µL of 10 mM PBS solution (pH 7.4). For preparation of the MB-apt, the steps were basically same as those for MB@Zr-MOFs-apt except for the addition of 800 µL of 1 mg/mL MB.

Fabrication of the Portable Electrochemical Platform
Prior to modification, the bare SPEs were cleaned in 0.5 M H 2 SO 4 under CV from −0.5 to +1.5 V until reproducible voltammograms were achieved [52]. Then, the four working SPEs were modified using 2.5 µL of AuNPs and dried in air [51]. Each working SPE was incubated with 2.5 µL of MB@Zr-MOFs-apt for 8 h at 4 • C to form strong Au-S bonds and promote complete coverage. Then, the aptamer-modified SPEs were treated with 2.5 µL of 5.0 mM MCH for 1 h to block the unspecific binding sites [53]. Finally, the as-prepared aptasensor was thoroughly rinsed with 10 mM PBS (pH 7.4). To perform HMI detection using the electrochemical POCT aptasensing platform, the disk was completely immersed in 200 µL of PBS (pH 7.4) containing 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ . The disk was incubated for 30 min and then thoroughly rinsed with ultrapure water and stored at 4 • C.
The electrochemical responses were measured and recorded based on the DPV method using PBS (pH 7.4) as the supporting electrolyte. The parameters of the DPV method were set as follows: amplitude potential = 0.1 V, pulse width = 0.1 s, pulse period = 1 s, quiet time = 2 s, and a potential range from −0.2 to +0.5 V. Each experiment was carried out three times, and results were recorded as the mean ± standard deviation (SD).

Detection Principle of the Electrochemical POCT Platform
A photograph of the portable electrochemical workstation and electrochemical POCT platform is provided in Figure S1. As shown in Scheme 1, the electrochemical POCT system consisted of a PET disk with a PDMS insulation loop and a portable electrochemical station. The outer diameter of the disk was 2 cm (almost a size of a coin), and the inner diameter was 1.2 cm. The diameters of the working SPEs were 1.8 mm. Furthermore, we designed the platform with six SPEs in specific positions, to simplify the operation complexity. For example, the reference electrode was placed opposite the counter electrode, with the same diameters, to avoid spilling of the PBS reaction solution and short circuiting of the conducting wires. In advance, aptamers modified with carboxyl groups were linked covalently with MB@Zr-MOFs modified with amino groups via amide reactions. AuNPs were used on all four working SPEs to provide effective electron transfer and serve as the loading substrates to support the thiolated aptamers. Then, the MB@Zr-MOFs-apts were immobilized on the surfaces of the AuNPs via the Au-S bond, and MCH was added to block the non-specific binding sites. The relatively weak DPV responses were observed owing to the resistive conformation of the electroactive MB@Zr-MOFs on the SPE surfaces. The target HMIs, were recognized by the corresponding specific aptamers upon addition and captured on the SPE surfaces, leading to a change in the conformation and the enhanced proximity of electroactive MB@Zr-MOFs with the electrode surface. Therefore, the DPV signals of the aptasensors were greatly enhanced. The DPV current increased with an increasing number of HMIs (∆ is the difference in the presence and absence of the four HMIs). The whole test lasted approximately 4 min: 2 min for DPV detection of the four HMIs and 2 min for changing PBS solutions and four conducting WEs. Thus, the proposed electrochemical POCT system enabled the convenient on-site detection of four HMIs. ensors 2023, 13, x FOR PEER REVIEW 5 o The target HMIs, were recognized by the corresponding specific aptamers upon addit and captured on the SPE surfaces, leading to a change in the conformation and the hanced proximity of electroactive MB@Zr-MOFs with the electrode surface. Therefore, DPV signals of the aptasensors were greatly enhanced. The DPV current increased w an increasing number of HMIs (Δ is the difference in the presence and absence of the f HMIs). The whole test lasted approximately 4 min: 2 min for DPV detection of the f HMIs and 2 min for changing PBS solutions and four conducting WEs. Thus, the propo electrochemical POCT system enabled the convenient on-site detection of four HMIs.

Scheme 1.
Principle of the electrochemical POCT aptasensing platform for detection of four HM

Characterization of Zr-MOFs and MB@Zr-MOFs
The morphologies of Zr-MOFs and MB@Zr-MOFs were characterized using HRTE The Zr-MOF is a light-yellow powder ( Figure S2B) with a uniform particle shape and s of 180 nm ( Figure 1A). After MB was loaded into the Zr-MOFs, the size slightly increa to 200 nm, and the shape became round, demonstrating that loading with the small m ecules has little influence on the size ( Figure 1B). The loading of MB on Zr-MOFs w confirmed through EDX mapping ( Figures S3 and S4), which revealed an increase in sulfur concentration after loading. Moreover, the scanning electron microscopy (SEM the AuNPs revealed their homogeneous dispersion with a diameter of 20 nm on the S Scheme 1. Principle of the electrochemical POCT aptasensing platform for detection of four HMIs.

Characterization of Zr-MOFs and MB@Zr-MOFs
The morphologies of Zr-MOFs and MB@Zr-MOFs were characterized using HRTEM. The Zr-MOF is a light-yellow powder ( Figure S2B) with a uniform particle shape and size of 180 nm ( Figure 1A). After MB was loaded into the Zr-MOFs, the size slightly increased to 200 nm, and the shape became round, demonstrating that loading with the small molecules has little influence on the size ( Figure 1B). The loading of MB on Zr-MOFs was confirmed through EDX mapping ( Figures S3 and S4), which revealed an increase in the sulfur concentration after loading. Moreover, the scanning electron microscopy (SEM) of the AuNPs revealed their homogeneous dispersion with a diameter of 20 nm on the SPEs ( Figure 1C). As shown in Figure 1D, the XRD patterns of simulated Zr-MOFs, Zr-MOFs, and MB@Zr-MOFs showed almost the same diffraction peak positions, indicating that MB had little impact on the crystalline structure of Zr-MOFs. Adsorption-desorption isotherms were obtained at 77.3 K to verify the synthesis of MB@Zr-MOFs [54]. As shown in Figure 1E, the Brunauer−Emmett−Teller (BET) surface area of Zr-MOFs was calculated as 814 m 2 /g, whereas the BET surface area decreased to 465 m 2 /g when the electroactive dyes occupied the pores of the Zr-MOFs, demonstrating the successful encapsulation of MB in Zr-MOFs. The loading efficiency of MB in Zr-MOFs was estimated as (814 − 465)/814×100% = 42.9%. The morphologies and adsorption characteristics were comparable with those reported in previous work [42]. In addition, the sensing stability of MB@Zr-MOFs was inspected based on the TGA curve under an oxygen atmosphere. As shown in Figure 1F, no weight loss was observed before 32 • C. Moreover, the weight loss was lower than 5 wt% before 45 • C, which clearly illustrated that MB@Zr-MOFs were kept stable under the sensing environment at room temperature. ensors 2023, 13, x FOR PEER REVIEW 6 of isotherms were obtained at 77.3 K to verify the synthesis of MB@Zr-MOFs [54]. As show in Figure 1E, the Brunauer−Emmett−Teller (BET) surface area of Zr-MOFs was calculat as 814 m 2 /g, whereas the BET surface area decreased to 465 m 2 /g when the electroact dyes occupied the pores of the Zr-MOFs, demonstrating the successful encapsulation MB in Zr-MOFs. The loading efficiency of MB in Zr-MOFs was estimated as (814 465)/814×100% = 42.9%. The morphologies and adsorption characteristics were compa ble with those reported in previous work [42]. In addition, the sensing stability of MB@Z MOFs was inspected based on the TGA curve under an oxygen atmosphere. As shown Figure 1F, no weight loss was observed before 32 °C. Moreover, the weight loss was low than 5 wt% before 45 °C, which clearly illustrated that MB@Zr-MOFs were kept sta under the sensing environment at room temperature.

Amplification Electrochemical Behavior of the POCT Sensor
To explore the electrochemical amplification effect of Zr-MOFs, DPV signals of M were tested with and without being loaded into Zr-MOFs. The bare SPEs were clean and modified using the same protocol of the established platform, except for incubati with 2.5 µL of MB-apt for 8 h at 4 °C to form tight Au-S bond. The final concentration apt in MB-apt was the same with that of 2.5 µM in MB@Zr-MOFs-apt. As shown in Figu 2, the DPV peak of MB-apt/AuNPs/SPE (red curve) was located at approximately +0.10 and the current was 0.623 µA. In contrast, the DPV peak of MB@Zr-MOFs-apt/AuNPs/S was located at approximately +0.09 V and the current was 4.46 µA, increasing by 7.2 tim compared with that of the red curve. The results suggest that loading MB into Zr-MO can greatly enhance the DPV signals of MB.

Amplification Electrochemical Behavior of the POCT Sensor
To explore the electrochemical amplification effect of Zr-MOFs, DPV signals of MB were tested with and without being loaded into Zr-MOFs. The bare SPEs were cleaned and modified using the same protocol of the established platform, except for incubation with 2.5 µL of MB-apt for 8 h at 4 • C to form tight Au-S bond. The final concentration of apt in MB-apt was the same with that of 2.5 µM in MB@Zr-MOFs-apt. As shown in Figure 2, the DPV peak of MB-apt/AuNPs/SPE (red curve) was located at approximately +0.10 V, and the current was 0.623 µA. In contrast, the DPV peak of MB@Zr-MOFs-apt/AuNPs/SPE was located at approximately +0.09 V and the current was 4.46 µA, increasing by 7.2 times compared with that of the red curve. The results suggest that loading MB into Zr-MOFs can greatly enhance the DPV signals of MB.

Electrochemical Behavior of the POCT Sensor
The electrochemical behavior of four HMIs was investigated using DPV. The SPEs on the disk with different aptamer modifications were tested in PBS (pH 7.4) solution. As shown in Figure 3, slight DPV peak currents located at approximately +0.11 V were observed for the four SPEs with the addition of MCH, suggesting that the aptasensors were slightly electroactive in the potential windows. The peak currents increased after the disk was immersed in PBS (pH 7.4) solution containing 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ ( Figure 3A for Cd 2+ , Figure 3B for Hg 2+ , Figure 3C for Pb 2+ and Figure 3D for As 3+ ). These results demonstrated that the conformational changes of the aptamers that captured the HMIs could shorten the distance between the electroactive MB@Zr-MOFs and the electrode surface [31].

Electrochemical Behavior of the POCT Sensor
The electrochemical behavior of four HMIs was investigated using DPV. The SPEs on the disk with different aptamer modifications were tested in PBS (pH 7.4) solution. As shown in Figure 3, slight DPV peak currents located at approximately +0.11 V were observed for the four SPEs with the addition of MCH, suggesting that the aptasensors were slightly electroactive in the potential windows. The peak currents increased after the disk was immersed in PBS (pH 7.4) solution containing 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ ( Figure 3A for Cd 2+ , Figure 3B for Hg 2+ , Figure 3C for Pb 2+ and Figure 3D for As 3+ ). These results demonstrated that the conformational changes of the aptamers that captured the HMIs could shorten the distance between the electroactive MB@Zr-MOFs and the electrode surface [31].

Electrochemical Behavior of the POCT Sensor
The electrochemical behavior of four HMIs was investigated using DPV. The SPEs on the disk with different aptamer modifications were tested in PBS (pH 7.4) solution. As shown in Figure 3, slight DPV peak currents located at approximately +0.11 V were observed for the four SPEs with the addition of MCH, suggesting that the aptasensors were slightly electroactive in the potential windows. The peak currents increased after the disk was immersed in PBS (pH 7.4) solution containing 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ ( Figure 3A for Cd 2+ , Figure 3B for Hg 2+ , Figure 3C for Pb 2+ and Figure 3D for As 3+ ). These results demonstrated that the conformational changes of the aptamers that captured the HMIs could shorten the distance between the electroactive MB@Zr-MOFs and the electrode surface [31].

Analytical Performance of the Electrochemical POCT Sensor
Under optimized conditions using 6 µM aptamer concentrations and 45 min of incubation time for HMIs ( Figure S10A,B), DPV was performed to determine the analytical performance of the proposed aptasensor for HMI detection. The results for different HMI concentrations are shown in Figure 4. As shown in Figure 4A-D, the DPV current increased linearly with the increase in each HMI concentration. The aptasensing platform exhibited useful analytical performance, with a dynamic range of 0.2-20 ppb for Cd 2+ , 0.1-10 ppb for Hg 2+ , 0.1-10 ppb for Pb 2+ , and 0.1-20 ppb for As 3+ .

Analytical Performance of the Electrochemical POCT Sensor
Under optimized conditions using 6 µM aptamer concentrations and 45 min of incubation time for HMIs ( Figure S10A,B), DPV was performed to determine the analytical performance of the proposed aptasensor for HMI detection. The results for different HMI concentrations are shown in Figure 4. As shown in Figure 4A-D, the DPV current increased linearly with the increase in each HMI concentration. The aptasensing platform exhibited useful analytical performance, with a dynamic range of 0.2-20 ppb for Cd 2+ , 0.1-10 ppb for Hg 2+ , 0.1-10 ppb for Pb 2+ , and 0.1-20 ppb for As 3+ . As shown in Figure 4E-H, a series of different standard concentrations of HMIs was applied to estimate the analytical performance. Utilizing a series of standard concentrations of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ , the current difference of aptasensor signals in the presence and absence of each HMI was defined as the Δ current. The DPV Δ current responses exhibited linearity as the concentration increased from a concentration of 0.2 to 20 ppb for Cd 2+ , 0.1 to 10 ppb for Hg 2+ , 0.1 to 10 ppb for Pb 2+ , and 0.1 to 20 ppb for As 3+ . Four calibration graphs were observed with the linear regression equations Δ current (µA) = 12.18 lg CCd( Ⅱ ) +21.73 (R 2 = 0.9981), Δ current (µA) = 14.00 lg CHg( Ⅱ ) +26.49 (R 2 = 0.9952), Δ current (µA) = 10.31 lg CPb( Ⅱ ) +16.25 (R 2 = 0.9943), and Δ current (µA) = 6.281 lg CAs( Ⅲ ) +19.93 (R 2 = 0.9839). The LOD was calculated using 3σ/S, where σ is the SD of blank signals (n = 3), and S represents the slope of the calibration curves [55]. The LODs were 0.039 ppb for Cd 2+ , 0.039 ppb for Hg 2+ , 0.073 ppb for Pb 2+ , and 0.022 ppb for As 3+ . Notably, the LODs obtained using the disk aptasensing platform were lower than the national standard values [56,57], indicating that the constructed electrochemical POCT system is suitable and practical for the quantitative on-site detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ . Compared with previously reported aptasensors for Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ analysis, the proposed electrochemical POCT aptasensing platform achieved lower LODs and a relatively wide linear range (Table S1), indicating that the fabricated aptasensing platform was acceptable and has potential for real sample analysis. The electrochemical POCT platform exhibited good As shown in Figure 4E-H, a series of different standard concentrations of HMIs was applied to estimate the analytical performance. Utilizing a series of standard concentrations of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ , the current difference of aptasensor signals in the presence and absence of each HMI was defined as the ∆ current. The DPV ∆ current responses exhibited linearity as the concentration increased from a concentration of 0.2 to 20 ppb for Cd 2+ , 0.1 to 10 ppb for Hg 2+ , 0.1 to 10 ppb for Pb 2+ , and 0.1 to 20 ppb for As 3+ . Four calibration graphs were observed with the linear regression equations ∆ current (µA) = 12.18 lg C Cd(II) + 21.73 (R 2 = 0.9981), ∆ current (µA) = 14.00 lg C Hg(II) + 26.49 (R 2 = 0.9952), ∆ current (µA) = 10.31 lg C Pb(II) + 16.25 (R 2 = 0.9943), and ∆ current (µA) = 6.281 lg C As(III) + 19.93 (R 2 = 0.9839). The LOD was calculated using 3σ/S, where σ is the SD of blank signals (n = 3), and S represents the slope of the calibration curves [55]. The LODs were 0.039 ppb for Cd 2+ , 0.039 ppb for Hg 2+ , 0.073 ppb for Pb 2+ , and 0.022 ppb for As 3+ . Notably, the LODs obtained using the disk aptasensing platform were lower than the national standard values [56,57], indicating that the constructed electrochemical POCT system is suitable and practical for the quantitative on-site detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ . Compared with previously reported aptasensors for Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ analysis, the proposed electrochemical POCT aptasensing platform achieved lower LODs and a relatively wide linear range (Table S1), indicating that the fabricated aptasensing platform was acceptable and has potential for real sample analysis. The electrochemical POCT platform exhibited good performance when compared to some previously reported methods for the detection of HMIs (Table S1).

Stability, Reproducibility, and Selectivity of Electrochemical POCT Platform
It is essential for a POCT aptasensing platform to have good stability. The aptasensors were prepared using the same procedure and stored at 4 • C for further use. As shown in Figure 5A, the aptasensors retained 96.5% for Cd 2+ , 98.4% for Hg 2+ , 95.3% for Pb 2+ , and 92.7% for As 3+ of its initial electrochemical response after five days. Moreover, it retained 94.8% for Cd 2+ , 97.6% for Hg 2+ , 95.3% for Pb 2+ , and 91.9% for As 3+ of its initial electrochemical response after ten days. Therefore, the aptasensors remained bioactive within 10 days. The obtained results demonstrated that the proposed aptasensors have good storage and utilization stability.
Biosensors 2023, 13, x FOR PEER REVIEW 9 of 13 performance when compared to some previously reported methods for the detection of HMIs (Table S1).

Stability, Reproducibility, and Selectivity of Electrochemical POCT Platform
It is essential for a POCT aptasensing platform to have good stability. The aptasensors were prepared using the same procedure and stored at 4 °C for further use. As shown in Figure 5A, the aptasensors retained 96.5% for Cd 2+ , 98.4% for Hg 2+ , 95.3% for Pb 2+ , and 92.7% for As 3+ of its initial electrochemical response after five days. Moreover, it retained 94.8% for Cd 2+ , 97.6% for Hg 2+ , 95.3% for Pb 2+ , and 91.9% for As 3+ of its initial electrochemical response after ten days. Therefore, the aptasensors remained bioactive within 10 days. The obtained results demonstrated that the proposed aptasensors have good storage and utilization stability. For the reproducibility test, as shown in Figure 5B, the relative standard deviations (RSDs) for three parallel tests using HMI aptasensors under the same experimental conditions were 1.7% for Cd 2+ , 1.5% for Hg 2+ , 2.1% for Pb 2+ , and 1.8% for As 3+ . The results demonstrated their acceptable reproducibility.
The selectivity of the constructed electrochemical POCT aptasensing platform toward Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ was investigated by including other metal ions, including Ag + , Cu 2+ , Zn 2+ , and Mn 2+ . Specifically, the mixture solution contained 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ and 30 ppb each of Ag + , Cu 2+ , Zn 2+ , and Mn 2+ . As indicated in Figure  5C, the aptasensor displayed a negligible current signal difference in the presence and absence of the additional metal ions, even with concentrations that were 3-fold higher than those of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ . When the mixture of target HMIs and additional metal ions was incubated on the four SPEs, a slight change in the DPV response was observed compared with that of the standard HMI solutions on the corresponding SPEs. The mixture solution exhibited almost the same DPV signal differences as those observed independently in the standard HMI solutions (relative errors: 6.12%, 2.96%, 2.98%, and 1.97% for detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ , respectively). The obtained results demonstrated that the proposed electrochemical POCT aptasensing platform is highly selective for the detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ under competitive HMI conditions.

Real Sample Analysis Using Milk
To assess the suitability of the proposed aptasensor for HMI analytical applications using real milk samples, 5 mL solutions of raw milk samples were preprocessed by adding 0.1, 1, and 10 ppb each of the four HMIs. To remove proteins from milk that are easily For the reproducibility test, as shown in Figure 5B, the relative standard deviations (RSDs) for three parallel tests using HMI aptasensors under the same experimental conditions were 1.7% for Cd 2+ , 1.5% for Hg 2+ , 2.1% for Pb 2+ , and 1.8% for As 3+ . The results demonstrated their acceptable reproducibility.
The selectivity of the constructed electrochemical POCT aptasensing platform toward Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ was investigated by including other metal ions, including Ag + , Cu 2+ , Zn 2+ , and Mn 2+ . Specifically, the mixture solution contained 10 ppb each of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ and 30 ppb each of Ag + , Cu 2+ , Zn 2+ , and Mn 2+ . As indicated in Figure 5C, the aptasensor displayed a negligible current signal difference in the presence and absence of the additional metal ions, even with concentrations that were 3-fold higher than those of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ . When the mixture of target HMIs and additional metal ions was incubated on the four SPEs, a slight change in the DPV response was observed compared with that of the standard HMI solutions on the corresponding SPEs. The mixture solution exhibited almost the same DPV signal differences as those observed independently in the standard HMI solutions (relative errors: 6.12%, 2.96%, 2.98%, and 1.97% for detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ , respectively). The obtained results demonstrated that the proposed electrochemical POCT aptasensing platform is highly selective for the detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ under competitive HMI conditions.

Real Sample Analysis Using Milk
To assess the suitability of the proposed aptasensor for HMI analytical applications using real milk samples, 5 mL solutions of raw milk samples were preprocessed by adding 0.1, 1, and 10 ppb each of the four HMIs. To remove proteins from milk that are easily adsorbed or cause the reduction of HMIs, 2 mL of trichloroacetic acid was added and stirred on a vortex shaker for 5 min. The mixed solutions were centrifuged at 10,000× g for 15 min, and the supernatant was collected and filtered through a 0.22 µm membrane filter. To prevent the interference of the residual trichloroacetic acid, the DPV curves of negative controls with milk pretreated in the same way without the HMIs were measured. No electrochemical signals were observed within the voltage window, indicating that there was no or very few HMIs detected. Then, the samples were prepared by adding different concentrations (10 and 30 µM) of HMI solutions into the commercial milk.

Conclusions
In this work, we constructed a sensitive, accurate, and flexible POCT aptasensing platform for the on-site detection of Cd 2+ , Hg 2+ , Pb 2+ , and As 3+ in milk samples. Furthermore, the proposed disk-like aptasensing platform was portable, cost-effective, and user-friendly owing to the highly efficient data collection enabled by the smartphone application and Bluetooth connection. Moreover, Zr-MOFs enabled considerable loading of the electroactive MB molecule for signal amplification. Under the optimal conditions, the platform exhibited a high sensitivity for the four target HMIs, with LODs as low as 0.039 ppb for Cd 2+ , 0.039 ppb for Hg 2+ , 0.073 ppb for Pb 2+ , and 0.022 ppb for As 3+ (S/N = 3). Additionally, the individual target HMIs can be specifically detected in a mixture of HMIs and kept active for ten days, demonstrating the useful specificity and stability of the sensing platform. Overall, the integrated electrochemical POCT system can be extended for practical on-site applications in milk sample monitoring. Further studies can focus on developing more versatile multi-channel POCT sensors for the simultaneous detection of more HMIs.
Author Contributions: Validation, H.X.; formal analysis, H.X.; resources, P.L.; writing-original draft preparation, H.X. and P.L.; writing-review and editing, F.C. and H.C.; supervision, H.C. and J.K., funding acquisition, H.C. and J.K. All authors have read and agreed to the published version of the manuscript.