Development of a Tannic Acid-Gelatin Modified Filter Paper Membrane for Trace Lead Detection in Water

Abstract

A novel membrane-based solid-phase extraction (SPE) device was developed using cellulose filter paper modified with a tannic acid–gelatin (TA/GE) composite for the preconcentration of lead (Pb­(II)) ions from aqueous samples. The TA/GE sorbent layer was fabricated via cross-linking gelatin with tannic acid at pH 9, forming a stable, partially insoluble network that enhances the membrane’s mechanical strength and ion-binding capability. Characterization by SEM confirmed the successful modification, while micro-XRF analysis demonstrated the membrane’s effective adsorption and elution of Pb­(II). After preconcentration, square wave anodic stripping voltammetry (SWASV) was used for quantitative determination. The method exhibited linear range of 10–100 μg·L^–1, with a limit of detection (LOD) of 2.7 μg·L^–1. Method validation using a certified reference material for trace metals in water produced a combined measurement uncertainty (U) of 9 μg·L^–1 at the 95% confidence level for a sample containing 61 μg·L^–1 Pb­(II), showing no significant deviation from the certified value. The developed method was further applied to drinking water and river water samples, achieving recoveries of 82–110% for spiked samples. These results confirm that the TA/GE-modified membrane provides a simple, low-cost, and environmentally friendly platform for trace Pb­(II) analysis, with performance meeting both WHO (10 μg·L^–1) and Thailand Pollution Control Department (50 μg·L^–1) regulatory standards for drinking and surface water.

1. Introduction

Lead (Pb) is an environmental contaminant with high human toxicity and persistence in ecosystems. Even at trace levels, lead can disrupt various metabolic and physiological processes in the human body, adversely affecting the hematopoietic, nervous, and circulatory systems, among others. , Consequently, the World Health Organization (WHO) has established a maximum allowable concentration of 10 μg L^–1 for lead in drinking water. Therefore, the accurate and sensitive determination of Pb­(II) in environmental and drinking water is critically important for both public health and regulatory compliance.

Although significant efforts have been made to monitor trace lead in water, many conventional analytical techniques still suffer from inadequate sensitivity for detecting Pb­(II) at ultra trace levels. Moreover, while modern instrumentation such as ICP–MS and AAS has advanced detection capabilities, the high operational costs and maintenance requirements restrict their convenient use, particularly in resource-limited settings. To address these limitations, sample pretreatment and preconcentration techniques particularly solid-phase extraction (SPE) have been widely developed. SPE enhances detection limits and selectivity by concentrating analytes onto selective sorbents before elution. , It offers several advantages, including high enrichment factors, rapid processing, cost-effectiveness, and operational simplicity. It should be noted that when SPE is employed for preconcentration rather than for sample cleanup, it is essential that the analyte can also be efficiently eluted from the sorbent. Membrane-based SPE has recently emerged as a particularly attractive approach. Compared with conventional packed-cartridge SPE, the membrane format enables high flow rates and rapid extraction due to its thin and porous structure, ensuring effective analyte–sorbent interaction with minimal backpressure. Notably, membrane-based SPE allows direct treatment of natural water samples without requiring filtration prior to preconcentration, thereby reducing sample preparation steps and saving analytical time. These features make the membrane-based strategy especially suitable for routine or large-scale environmental monitoring. ,

In parallel with conventional spectrometric techniques, electrochemical methods have attracted increasing interest for trace Pb­(II) determination owing to their low cost, portability, and rapid response. Advanced electrochemical sensing platforms employing nanostructured interfaces or affinity-based recognition elements have demonstrated exceptional analytical sensitivity, with detection limits reaching the pico- to femtomolar range. For example, DNAzyme-mediated electrochemical systems have been reported to achieve subpicomolar Pb­(II) detection through highly specific, metal-dependent catalytic reactions. More recently, Amin et al. developed a printed circuit board–based affinity electrochemical sensor integrated with microfluidic enrichment, achieving a detection limit as low as 1.85 fM without the need for an external preconcentration step. In addition, alternative portable approaches, such as smartphone-based nanocolorimetry, have also been explored for on-site Pb­(II) analysis, enabling sensitive optical quantification of lead chromate nanoparticles formed from dissolved Pb­(II) in water samples.

In this context, the integration of membrane-based SPE with portable electrochemical detection offers a balanced and practical strategy, combining effective matrix cleanup and analyte preconcentration with cost-effective, and field-deployable analysis. Tannic acid (TA), a naturally occurring polyphenolic compound, has attracted significant interest as a metal-binding material due to the presence of multiple adjacent phenolic hydroxyl groups capable of strong chelation with heavy metal ions. TA has been reported to efficiently bind metals such as Cr­(VI), Hg­(II), Cd­(II), Ni­(II), and Pb­(II). , However, the high water solubility of TA limits its direct application in aqueous systems, necessitating immobilization onto insoluble supports. Various strategies have therefore been explored to immobilize TA onto substrates such as cellulose, collagen fibers, gelatin-based nanofibers, and composite beads for heavy metal adsorption. −

Gelatin (GE), a biodegradable polypeptide derived from collagen, contains abundant amino and carboxyl functional groups that facilitate metal ion binding and chemical modification. , Nevertheless, native GE exhibits limited thermal and mechanical stability and tends to swell or dissolve under aqueous conditions. To address these drawbacks, TA has been widely employed as a natural cross-linking agent for GE, forming hydrogen-bonded and covalently cross-linked networks that enhance mechanical strength, reduce swelling, and decrease solubility through the consumption of free amino groups. −

In this study, cellulose filter paper was modified with a cross-linked TA/GE polymer to develop a membrane-based solid-phase extraction sorbent with high affinity for Pb­(II) ions. In this system, TA provides polyphenolic functional groups capable of chelating Pb­(II), while GE acts as a biopolymeric matrix that immobilizes TA onto the cellulose fibers, enhances the mechanical stability of the membrane, and provides additional Pb­(II) adsorption through hydrogen-bonding interactions. The resulting SPE device enables efficient preconcentration of Pb­(II) from aqueous samples at high flow rates with minimal backpressure. Adsorbed Pb­(II) is subsequently eluted using dilute nitric acid and quantitatively determined by square wave anodic stripping voltammetry (SWASV) employing a screen-printed gold electrode and a portable potentiostat. To the best of our knowledge, this work represents the first report of a TA/GE-modified paper membrane integrated with electrochemical detection for trace heavy metal analysis, offering a simple, cost-effective, and regulatorily relevant approach for environmental Pb­(II) monitoring.

2. Materials and Methods

2.1. Reagents and Apparatus

Stock standard solution of lead (Pb) at a concentration of 1000 mg L^–1 was purchased from Merck (Darmstadt, Germany) and diluted to the required concentrations with Millipore-grade deionized (DI) water. Gelatin (GE) from bovine skin (Type B, Sigma-Aldrich) and tannic acid (TA) powder were obtained from HiMedia (India). A 0.1 M acetate buffer solution at pH 4.5, prepared by mixing acetic acid and sodium acetate (Merck, Darmstadt, Germany), was used as the supporting electrolyte. Other reagents, including sodium hydroxide, hydrochloric acid (37%, fuming), nitric acid (65%), and acetic acid (99.8%), were also from Merck (Darmstadt, Germany). A certified reference material (CRM) of mixed solution of Cd, Cu, Pb and Zn (TRM-E-2002, 100 mL) was provided by the National Institute of Metrology Thailand (NIMT). To eliminate trace metal contamination, all glassware was precleaned by soaking in 10% HNO₃ for 3 days, followed by thorough rinsing with DI water prior to use. For the preconcentration process, a peristaltic pump (Ismatec, Germany), a reusable 13 mm polycarbonate syringe filter holder (Sartorius Stedim Biotech, Germany), and a 50 mL plastic syringe were utilized. The volumes of all solutions-including the sample, rinsewater, and 0.1 M HNO₃ eluentwere dispensed using graduated pipettes.

For quantitation of Pb­(II), electrochemical measurements were carried out using a three-electrode system consisting of a commercial screen-printed gold electrode (SPGE) obtained from Zensor (Taiwan), which included a gold working electrode, an interdigitated gold counter electrode, and an Ag/AgCl reference electrode. Square wave anodic stripping voltammetry (SWASV) was performed using a PalmSens potentiostat (PalmSens BV, Houten, The Netherlands). During the electrodeposition and cleaning steps, a magnetic stirrer with a Teflon-coated stir bar was used to ensure uniform mixing inside the 30 mL electrochemical cell.

Four commercial drinking water samples, with different brands, were obtained from local markets in Bangkok, Thailand. In addition, four surface water samples were collected from different locations along the Mae Klong River in Kanchanaburi Province, Thailand. The river water samples were collected in sterilized polyvinyl chloride (PVC) containers to prevent contamination. All water samples were stored at 4 °C prior to analysis.

2.2. Preparation of TA/GE Membrane

A 3% w/v TA solution was prepared by dissolving solid TA in DI water, followed by pH adjustment to pH 9 using 1 M NaOH. Separately, a 1% w/v GE solution was prepared by dissolving GE in DI water and heating at 45 °C for 15 min, after which the pH was also adjusted to pH 9. The cross-linking reaction between GE and TA was initiated by mixing 30 mL of the 3% TA solution with 30 mL of the 1% GE solution. The mixture was maintained at 60 °C for 1 h at constant pH, with the pH measured every 30 min and adjusted to pH 9 every 30 min with 1 M NaOH as required.

A precut 7 cm × 9 cm sheet of Whatman No. 1 filter paper (GE Healthcare, UK) was immersed in the freshly prepared TA/GE mixture contained in a flat tray for approximately 10 s. The coated paper was allowed to dry at room temperature overnight. Circular discs (13 mm diameter) were subsequently obtained using a paper punch. The discs were stored in a clean container until use.

2.3. Physical Characterization of TA/GE Paper Membrane

The morphology of the TA/GE-modified filter paper membrane was examined using a FTIR Spectrometer (Spectrum Two FT-IR Spectrometer, PerkinElmer Scientific, USA) to observe interaction between Pb and TA/GE modified filter paper and a field emission scanning electron microscope (FE-SEM, LEO 1455VP, England) to observe surface structure and coating distribution. The elemental composition of the paper discs was investigated using a Micro X-ray Fluorescence (Micro-XRF) spectrometer (M4 Tornado Plus, Bruker, Thailand), equipped with a Rhodium (Rh) anode X-ray tube operating at 50 kV and 600 μA.

2.4. Point of Zero Charge Analysis

The point of zero charge pH (PZC) of the TA/GE-modified membrane was determined using the pH drift method with 0.01 M KNO₃ as the background electrolyte. The membrane was cut into small pieces (0.5 × 0.5 cm²), and 0.10 g was added to 10.00 mL of the electrolyte solution. The initial pH was adjusted from 2 to 12 using 0.1 M HNO₃ or NaOH. After equilibration under gentle shaking at room temperature for 24 h, the final pH_f was recorded. The pHPZC was obtained from the plot of ΔpH (pH_f–pH₀) versus pH₀ at ΔpH = 0.

2.5. Preconcentration Procedure

Four pieces of dry TA/GE-modified filter paper membranes were placed inside a 13 mm reusable syringe membrane holder, and the holder lid was screwed tightly. The device was then connected to the reservoir syringe and the peristaltic pump, as shown in Figure . First, the membrane stack was washed with 20.0 mL of DI water stored in the 50 mL plastic syringe reservoir connected to the inlet of the membrane holder. This step was performed to remove any unreacted TA that may have been adsorbed onto the membrane surface. The outlet of the membrane holder was connected via tubing to a peristaltic pump and then to a waste beaker or collection vial. After the washing step, 50.0 mL of standard/sample Pb­(II) solution (adjusted to pH 5) was dispensed into the reservoir syringe. The solution was drawn through the extraction device at a flow rate of approximately 1 mL/min using the peristaltic pump. The membranes were then rinsed with 1.0 mL of DI water to remove any weakly bound ions. Lead ions retained on the TA/GE membrane stack were subsequently eluted with 3.0 mL of 0.1 M HNO₃ into a clean 10 mL plastic vial. The functionalized filter paper was designed for single-use operation; therefore, regeneration procedures and repeated-use experiments were not performed.

1.

Schematic representation of the experimental setup for Pb­(II) preconcentration using a syringe filter holder containing TA/GE-modified filter paper discs. The filter holder is connected to a syringe reservoir for sample loading and a peristaltic pump, which draws the solution through the holder at a constant flow rate of 1 mL/min to either waste or a collection vial.

A photograph of the experimental setup is provided in Supporting Information S1.

2.6. Quantitation Procedure

Square wave anodic stripping voltammetry (SWASV) were performed using a portable PalmSens4 potentiostat equipped with PSTrace 5.9 software. The electrode was a screen-printed gold electrode which was immersed in an electrochemical cell containing 30.0 mL of 0.1 mol L^–1 acetate buffer solution (pH 4.5) as the supporting electrolyte. A 2.0 mL aliquot of the eluted Pb­(II) solution was added to the electrochemical cell. Only for analysis of the CRM, which contained Cu­(II) ions, an aliquot of 0.35 mL of aqueous 0.1 M KCN was also added as a masking agent. After stirring with a Teflon coated magnetic bar for 1 min the SWASV measurement was started. The SWASV procedure consisted of four steps: conditioning at 0.2 V (vs Ag/AgCl) for 15 s, electrodeposition at −0.8 V (vs Ag/AgCl) for 180 s, stripping from −0.8 to 0.0 V (vs Ag/AgCl) after an equilibration time of 10 s, at step potential of 5 mV and amplitude of 25 mV. The frequency was 50 Hz. Finally, a cleaning step at 0.2 V (vs Ag/AgCl) for 30 s to remove any residual species from the electrode surface. Stirring was applied only during the conditioning, electrodeposition, and cleaning steps to ensure solution homogeneity and enhance the mass transport to the electrode surface. Each sample was measured in triplicates. Between consecutive measurements, the electrode was rinsed thoroughly with deionized water and gently dried with tissue paper.

Potassium cyanide (KCN) was handled in accordance with institutional safety regulations and under appropriate laboratory safety conditions. All cyanide-containing waste solutions generated during the experiments were collected separately and transferred to the University Toxic Waste Disposal Unit for proper treatment and disposal.

3. Results and Discussion

3.1. Physical Characterization of TA/GE Modified Paper

3.1.1. FTIR Analysis

Figure compares the ATR–FTIR spectra of the TA/GE-modified membrane before and after Pb­(II) loading. For the pristine TA/GE membrane, a broad absorption band at ∼3335 cm^–1 is observed, corresponding to overlapping O–H and N–H stretching vibrations arising from phenolic groups of TA, amino groups of GE, and hydroxyl groups of cellulose. The band at ∼1535 cm^–1 is assigned to the amide II vibration of GE (N–H bending coupled with C–N stretching)., confirming the successful incorporation of GE within the TA-based polymeric network.

2.

ATR–FTIR spectra of TA/GE-modified membrane before and after Pb­(II) adsorption. The decreased intensities of bands at ∼3335 cm^–1 (O–H/N–H stretching) and ∼1535 cm^–1 (amide II) after Pb­(II) loading indicate the involvement of hydroxyl and amino groups in Pb²⁺ binding, while the change at ∼1641 cm^–1 suggests perturbation of the amide environment due to metal–ligand interactions.

Following Pb­(II) adsorption, a clear reduction in the intensity of both the ∼3335 and ∼1535 cm^–1 bands is observed, indicating the involvement of hydroxyl and amino functional groups in Pb²⁺ binding, likely through metal–ligand coordination and hydrogen-bonding interactions. In addition, the appearance of a band at ∼1641 cm^–1 can be attributed to amide I (CO stretching), whose change in intensity suggests perturbation of the local chemical environment of GE upon Pb²⁺ complexation. Besides, the retention of the polysaccharide-related bands in the 1200–1000 cm^–1 region suggests that the cellulose backbone remains structurally intact after Pb²⁺ adsorption. Collectively, these spectral changes confirm that Pb²⁺ ions interact primarily with oxygen- and nitrogen-containing functional groups on the TA/GE-modified membrane, supporting the proposed adsorption mechanism.

Based on the FTIR analysis, which indicates the involvement of hydroxyl and amide functional groups in Pb²⁺ binding, a schematic representation of the cross-linking reaction and Pb²⁺ adsorption mechanism is proposed. Figure schematically illustrates the formation of the TA/GE cross-linked network and the subsequent Pb­(II) adsorption process. As shown in Figure a, TA interacts with GE under alkaline conditions through hydrogen bonding and coordination-type interactions, resulting in a stable cross-linked network immobilized within the cellulose matrix. Upon exposure to a Pb­(II) solution (Figure b), the phenolic hydroxyl groups of TA and the amide and amino functionalities of GE act as active binding sites, enabling effective chelation of Pb²⁺ ions. The proposed adsorption mechanism is consistent with the FTIR results, which indicate the participation of O–H and amide groups in metal binding. The cross-sectional illustration (Figure c) highlights the uniform distribution of the TA/GE network throughout the porous membrane, providing abundant accessible binding sites while maintaining high permeability, which is crucial for efficient Pb­(II) preconcentration.

3.

Schematic illustration of (a) the cross-linking reaction between tannic acid (TA) and gelatin (GE) on cellulose filter paper under alkaline conditions, (b) the proposed Pb²⁺ adsorption mechanism on the TA/GE-modified membrane via coordination with phenolic hydroxyl and amide functional groups, and (c) a cross-sectional view of the TA/GE-modified membrane showing the homogeneous distribution of active binding sites within the porous structure.

3.1.2. Scanning Electron Microscopy (SEM) Analysis

Scanning electron microscopy (SEM) was used to examine the surface morphologies of three sample conditions: unmodified filter paper, TA/GE- modified filter paper, and TA/GE filter paper after loading with 25 mL of a 1 mg·L^–1 Pb­(II) solution. Images were captured at magnifications of ×500 and ×5000, as shown in Figure . The unmodified filter paper (Figure a,d) exhibited a typical cellulose fibrous structure with smooth surfaces and large interfiber voids, indicating a highly porous and uncoated material. Following surface modification with the TA/GE layer (Figure b,e), distinct morphological changes were observed. The fiber surfaces appeared more compact and were coated with a thin, irregular polymeric film, resulting in a rougher texture. These changes confirm the successful immobilization of the TA/GE composite onto the cellulose substrate. After flushing with the solution of Pb²⁺ ions additional structural modifications were evident (Figure c,f). The surface became more compact and uneven, with small particulates or aggregates visible, particularly at higher magnification. These features are indicative of Pb­(II) adsorption or deposition on the functionalized membrane. The presence of wrinkled textures and localized distortions (indicated by arrows) further supports the formation of Pb²⁺- ligand interactions between the metal ions and the TA/GE matrix.

4.

SEM images of (a,d) unmodified filter paper, (b,e) TA/GE-modified filter paper, and (c,f) TA/GE-modified filter paper after Pb²⁺ adsorption (25 mL of 1 mg·L^–1 Pb­(II)), recorded at magnifications of ×500 and ×5000. Compared with the smooth surface of the unmodified filter paper, the TA/GE-modified samples exhibit a rougher and more heterogeneous morphology. After Pb­(II) adsorption, pronounced wrinkled textures and localized surface distortions (indicated by arrows) are clearly observed, confirming successful surface modification and subsequent Pb²⁺ ligand interactions.

3.1.3. X-ray Fluorescence Analysis (XRF)

Micro X-ray fluorescence (micro-XRF) imaging was employed to examine the distribution of Pb on the TA/GE membrane before and after the adsorption–desorption process. In this experiment, 25 mL of a 1 mg·L^–1 of Pb­(II) solution was loaded onto the membrane, followed by elution with 5 mL of 0.1 M HNO₃. As shown in Figure , a distinct Pb signal (represented in green) was observed on the membrane surface after loading, confirming the presence and surface accumulation of Pb­(II) on the TA/GE membrane. After elution with 0.1 M HNO₃, the signal intensity markedly decreased, approaching that of the unexposed control membrane. These results confirm the effectiveness of the TA/GE membrane for Pb­(II) preconcentration.

5.

Micro X-ray fluorescence (XRF) images of the TA/GE membrane: (a) before Pb­(II) loading, (b) after loading with 25 mL of 1 mg·L^–1 Pb­(II), and (c) after elution with 5 mL 0.1 M HNO₃.

In addition, adsorption and desorption of other metal ions commonly found in contaminated water, including Cu­(II), Cd­(II), and Zn­(II), were also examined. As presented in Supporting Information S2, both Cu­(II) and Zn­(II) generated detectable signals, indicating that the TA/GE membrane can also preconcentrate these ions under the studied conditions. In contrast, no significant Cd­(II) signal was detected, which is attributed to the limited sensitivity of Cd­(II) under the employed XRF conditions, where an Rh anode was used for sample excitation.

3.2. Optimization of TA–GE Membrane Formation

All test samples were preconcentrated as described in Section using 50 mL of 0.05 mg·L^–1 Pb­(II) solution at pH 5. The SWASV conditions were as described in Section .

3.2.1. Tannic Acid Concentration

The effect of TA concentration on the electrochemical response was investigated in the range of 1–7% w/v, with GE fixed at 1% w/v, as shown in Figure a. TA played a dual role: (i) as a cross-linking agent for GE, through interactions between its phenolic hydroxyl groups and the amino groups in the helical protein structure of GE, thereby forming a more stable, less water-soluble network − and (ii) as a source of phenolic groups for Pb­(II) binding.

6.

SWASV response at a gold electrode as a function of (a) tannic acid concentration (GE fixed at 1% w/v) and (b) GE concentration (tannic acid fixed at 3% w/v). A 25.0 mL aliquot of 1 mg·L^–1 Pb­(II) solution was used for preconcentration.

The peak current increased with TA concentration up to 3% w/v, reflecting enhanced Pb­(II) adsorption due to the increased availability of active binding sites. At higher concentrations, the current gradually decreased, which is attributed to excessive TA–GE cross-linking, which is expected to produce a denser GE network with reduced porosity, permeability, and polymer chain mobility, thereby limiting Pb²⁺ diffusion to internal binding sites. This trend is consistent with previous reports showing that increased TA content reduces GE solubility and increases network rigidity in TA–GE systems. Supporting this interpretation, at a TA concentration of 10% w/v, the formation of a visibly sticky gel was observed during material preparation, indicating extensive cross-linking and unsuitability for adsorption on the immersed paper.

3.2.2. Gelatin Concentration

The effect of GE concentration was evaluated in the range of 1–10% w/v, with TA fixed at 3% w/v, as shown in Figure b. The peak current increased as the GE concentration increased from 0 to 1% (w/v), indicating improved Pb­(II) adsorption due to the formation of a continuous GE coating that facilitated effective cross-linking with TA and provided accessible binding sites. However, a slight decrease in signal was observed at 3% (w/v), which is attributed to the formation of a thicker and more compact GE layer, reducing membrane porosity and hindering Pb­(II) diffusion. At GE concentrations above 10% (w/v), significant back pressure prevented solution flow through the modified filter, consistent with excessive GE deposition and markedly reduced permeability. Based on these findings, 1% w/v GE combined with 3% w/v TA was determined to be optimal, producing a stable coating that maximized Pb­(II) adsorption and electrochemical response.

3.2.3. Reaction Time

The effect of reaction time on cross-linking between GE and TA was investigated in the range of 0–210 min (Supporting Information S3). The current signal increased slightly during the first 30 min, indicating that initial cross-linking enhanced the sorbent’s ability to retain Pb²⁺ ions. Beyond 30 min, the current remained relatively constant, suggesting that the cross-linking reaction reached completion within this period. These results indicate that the tannin–GE network forms rapidly and is fully established within 30 min. Therefore, a cross-linking time of 30 min is sufficient to produce a stable and effective sorbent layer on the paper substrate.

3.2.4. pH

The effect of pH on the cross-linking efficiency between GE and TA was studied over the range of pH 7–11 (Supporting Information S4). The results indicate that the cross-linking reaction proceeds effectively within this range. The mechanism involves hydrogen bonding and covalent interactions between the phenolic hydroxyl groups of TA and the amino groups (−NH₂) of GE. , At near-neutral to mildly alkaline pH (7–10), the amino groups in GE remain sufficiently nucleophilic to react with electrophilic sites of TA (such as oxidized phenolic groups), leading to the formation of a stable and porous polymeric network. Additionally, under mildly basic conditions, the phenolic hydroxyls in TA are more ionized, which enhances hydrogen bonding and ionic interactions with GE and further stabilizes the film structure on the paper substrate. At pH 11, however, a significant reduction (∼35%) in current was observed, indicating diminished Pb­(II) retention. This is likely due to partial GE denaturation under high alkaline conditions, leading to disruption of the triple-helix structure and loss of reactive amino groups. Therefore, pH 9 was selected as the optimal condition for cross-linking.

3.3. Optimization of Preconcentration Process

The Pb­(II) test samples were preconcentrated as described in Section . The initial test conditions used 50 mL of 0.05 mg·L^–1 Pb­(II) solution. SWASV measurements were carried out following the steps described in Section .

3.3.1. Effect of Solution pH and Point of Zero Charge on Pb²⁺ Adsorption

The pH of the Pb­(II) solution plays a crucial role in the adsorption efficiency of the TA/GE-modified membrane. As shown in Figure a, the electrochemical response increased as the pH rose from 4 to 5 and then gradually decreased at higher pH values. At pH < 4, the phenolic hydroxyl groups of TA and the amino groups of GE are predominantly protonated, resulting in a limited availability of active sites for Pb­(II) coordination and, consequently, low adsorption efficiency.

7.

Effect of (a) pH of Pb­(II) solution, (b) volume of 0.05 mg·L^–1 Pb­(II) solution, (c) number of TA/GE discs installed in holder, and (d) volume of 0.1 M HNO₃ in the elution step.

To further elucidate the pH-dependent adsorption behavior, the point of zero charge pH (pHPZC) of the TA/GE-modified membrane was determined (results shown in Figure ) and found to be 4.74. Below this pH, the membrane surface carries a net positive charge, leading to electrostatic repulsion toward Pb²⁺ ions. As the pH increases above the pHPZC, deprotonation of phenolic and amino functional groups occurs, rendering the membrane surface negatively charged. This change enhances Pb²⁺ adsorption through electrostatic attraction and coordination interactions. The marked increase in signal from pH 4 to 5 is attributed to increased ionization of phenolic −OH groups of TA and the enhanced affinity of the membrane for Pb²⁺ ions.

8.

Determination of the point of zero charge (pHPZC) of the TA/GE-modified membrane using the pH drift method in 0.01 M KNO₃ electrolyte.

At pH above 5, the electrochemical signal gradually decreased, likely due to hydrolysis of Pb²⁺ and the formation of insoluble Pb­(OH)₂ species. This effect is most pronounced in the pH range of 6.5–7.0. The process reduces the concentration of free Pb²⁺ ions and limits their availability for adsorption onto the membrane. Based on these results, pH 5 was selected as optimal for Pb²⁺ retention, as it lies slightly above the pHPZC, maximizes electrostatic and coordination interactions, and avoids Pb­(II) precipitation. Consequently, this condition provides the highest and most reproducible electrochemical response.

3.3.2. Effect of Volume of Pb­(II) Solution

The volume of Pb­(II) solution used in the preconcentration step also affects the extent of interaction between lead ions and the TA/GE membrane. To investigate this, 0.05 mg·L^–1 Pb­(II) solutions were loaded in volumes ranging from 25 to 100 mL through the four-layer modified paper membrane. As shown in Figure b, the peak current increased with increasing sample volume, consistent with the increasing amount of Pb­(II) loaded on the membrane. A plateau value appears to be reached at volumes greater than 100 mL, indicating that the maximum sorbent capacity is approximately 5 μg Pb. Thus, a sample volume of 50 mL is suitable for Pb­(II) concentrations not exceeding 0.05 mg·L^–1.

3.3.3. Number of Paper Discs in Holder

The effect of the number of TA/GE membrane layers in the holder on lead extraction efficiency was investigated. The number of layers was varied from 1 to 4, which is the maximum thickness that allows the holder lid to be screw closed. As shown in Figure c, the electrochemical signal increased proportionally with the number of layers, indicating enhanced lead adsorption capacity. This trend suggests that increasing the number of modified layers provides more available binding sites for Pb²⁺ ions, improving preconcentration efficiency. Based on these results, four layers were used in subsequent experiments, as this number provided the highest sorption capacity within the physical constraints of the device.

3.3.4. Elution Volume of 0.1 M HNO₃

The effect of eluent volume on the desorption efficiency of Pb­(II) from the TA/GE membrane was investigated using 0.1 M HNO₃ volumes ranging from 2 to 5 mL. As shown in Figure d, the electrochemical signal increased when the eluent volume was raised from 2 to 3 mL but decreased with further increases in volume. This decrease can be attributed to a dilution effect: although more Pb²⁺ ions are desorbed, the larger eluent volume lowers the final concentration of Pb­(II) for detection. Therefore, an eluent volume of 3 mL was selected for subsequent experiments, as it provides the highest analytical signal.

3.4. Comparison of Pb²⁺ Adsorption Using Different Paper Modifications

Previous studies have shown that filter paper, GE, and TA each possess ability to adsorb Pb­(II) ions. To evaluate the enhanced performance of the newly developed TA/GE membrane (1% w/v GE and 3% w/v TA), a comparative study was conducted using four types of sorbents: (i) unmodified filter paper (control), (ii) filter paper modified with 1% w/v GE only, (iii) filter paper modified with 3% w/v TA only, and (iv) filter paper modified with both 1% GE and 3% TA. All samples were tested under identical preconcentration conditions using 50 mL of 0.05 mg·L^–1 Pb­(II) solution.

As illustrated in Figure , unmodified filter paper exhibited the lowest electrochemical signal, with a peak current of 3.2 μA, reflecting minimal adsorption capability. The GE-modified paper produced a moderate signal of 3.4 μA, likely due to improved surface structure and increased hydrophilicity. The TA-modified paper showed a signal of 3.8 μA, attributed to the presence of multiple phenolic hydroxyl groups capable of chelating Pb²⁺ ions. Most notably, the comodified TA/GE membrane yielded the highest peak current of 9.1 μA, highlighting a synergistic effect: GE provides a flexible structural matrix supporting film formation, while TA contributes abundant metal-binding sites. Together, these components enhance the adsorption capacity far beyond that of either component alone. These findings confirm that dual-component modification significantly improves the sorptive performance of the paper-based membrane, making it a promising platform for trace-level heavy metal analysis.

9.

Effect of different paper modifications on SWASV signals: (i) unmodified filter paper (control), (ii) filter paper modified with 1% w/v GE, (iii) filter paper modified with 3% w/v TA, and (iv) filter paper modified with both 3% TA and 1% GE.

3.5. Interference Study of Coexisting Metal Ions in Pb­(II) Analysis

To evaluate the selectivity of the developed analytical method, an interference study was conducted by introducing potential coexisting metal ions into a Pb­(II) solution. Standard solutions of Cu­(II), Cd­(II), and Zn­(II) were individually spiked into a 50 μg·L^–1 Pb­(II) solution to yield final concentrations of 500 μg·L^–1, 20 μg·L^–1, and 600 μg·L^–1, respectively. These concentrations were selected to represent typical or higher levels of these metals present in the certified reference material (CRM) obtained from the National Institute of Metrology Thailand (NIMT). Aliquot of 50 mL of the spiked samples were preconcentrated and analyzed by SWASV as described in Sections and .

The results showed that Cd­(II) and Zn­(II) did not significantly affect the Pb­(II) signal, as the measured current responses remained within three standard deviations (±3SD) of the signal obtained from the Pb­(II) solution alone. This indicates that the method is tolerant to these metal ions under the tested conditions. In contrast, Cu­(II) at 100 μg·L^–1 caused noticeable signal suppression, with the Pb­(II) peak current decreasing by approximately 40%. This effect is due to the overlap of the reduction potentials of Pb­(II) and Cu­(II) on gold electrodes in SWASV. To mitigate Cu­(II) interference, 0.01 M potassium cyanide (KCN) was added to the electrolyte solution to selectively mask Cu­(II) through the formation of stable [Cu­(CN)₄]^3– complexes, preventing copper deposition on the electrode during SWASV measurement. The addition of KCN effectively restored the Pb­(II) signal to within the acceptable range (±3SD) of the original response. (Data in Supporting Information S5).

These results demonstrate that the developed method is highly selective for Pb­(II) in the presence of common coexisting metal ions in environmental water samples, and that Cu­(II) interference can be effectively suppressed by complexation with cyanide ions.

3.6. Analytical Performance of the Developed Method

3.6.1. Linear Calibration and LOD

To evaluate the sensitivity enhancement provided by the developed TA/GE modified paper-based solid-phase extraction, calibration curves were constructed for Pb­(II) standard solutions over the concentration range of 0.01–0.1 mg·L^–1, both with and without the preconcentration step. Without preconcentration, the linear calibration was found to be y = (8.39 ± 0.36) x + (0.02 ± 0.25), with r ² = 0.9964. With preconcentration, the calibration curve was y = (157.6 ± 2.1) x −(0.22 ± 0.12), with r ² = 0.9990, corresponding to a 19-fold increase in sensitivity. Both calibration curves exhibited excellent linearity (r ² > 0.996).

The limit of detection (LOD) was calculated as three times the standard deviation of the regression (S _y/x ) divided by the slope of the regression line (S). Based on the calibration equation obtained in Figure , the LOD was determined to be 2.7 μg·L^–1, while the limit of quantification (LOQ), defined as ten times S _y/x divided by S, was 9.10 μg·L^–1. These values demonstrate the high sensitivity of the proposed method, enabling reliable detection of Pb­(II) at low concentrations suitable for environmental and drinking water monitoring. ,

10.

SWASV voltammograms recorded for Pb­(II) at concentrations ranging from 0.01 to 0.10 mg·L^–1 obtained from the eluted solutions after membrane-based preconcentration, measured versus an Ag/AgCl reference electrode. The inset shows the corresponding calibration curve.

Comparative evaluation of the present method with previously reported Pb­(II) preconcentration and determination strategies is summarized in Table . Various solid-phase extraction–based approaches have been developed using a wide range of sorbent materials, including carbon xerogels, silver nanoparticle–modified cellulose, biopolymer composites, ion-imprinted polymers, magnetic metal–organic frameworks, and MOF-based hybrid systems coupled with advanced detection techniques. Many of these systems achieve excellent analytical sensitivity, in some cases reaching ng·L^–1 levels. , However, such high performance is often accompanied by increased material cost, complex synthesis procedures, or limited suitability for routine field deployment.

1. Comparison of Solid-Phase Extraction–Based Methods for Pb­(II) Determination.

	sorbent/extraction material	detection technique	LOD	main advantages	limitations	ref
1	oxidized carbon xerogel	FI–UV–vis (dithizone complex)	0.9 μg L–1	simple spectrophotometric detection; moderate enrichment factor (50×)	requires derivatization; limited sensitivity; off-line elution
2	AgNPs–cellulose composite (thin-film microextraction)	ICP–MS	0.0085 μg L–1	high sensitivity; good repeatability	use of noble metals; expensive instrumentation; multistep preparation
3	chitosan/alginate/PVA composite (syringe SPE)	FAAS	4.19 μg L–1	sustainable materials; high recovery; good precision	batch operation; multiple loading/elution cycles; higher acid consumption
4	ion-imprinted polymer on mesoporous silica/graphene oxide	FAAS	0.003 μg L–1	excellent selectivity and ultralow LOD	complex synthesis; long extraction time; synthetic polymers
5	magnetic Zr-MOF@Fe3O4	FAAS	3.9 μg L–1	high adsorption capacity; magnetic separation	MOF synthesis complexity; relatively high LOD
6	UiO-66-NH2@TA (SPE–LIBS)	LIBS	180 μg L–1	rapid detection; suitable for food matrices	requires LIBS system; moderate sensitivity in aqueous samples
7	TA/GE-modified cellulose membrane (membrane SPE)	SWASV (portable potentiostat)	2.7 μg L–1	simple fabrication; biodegradable materials; portable and low-cost detection	LOD higher than some works ,	this work

^aFI: flow system.

^bAgNP: silver nanoparticle.

^cICP-MS: inductively coupled plasma–mass spectrometry.

^dPVC: polivinyl alcohol.

^eFAAS: flame atomic absorption spectrometry.

^fZr-MOF@Fe₃O₄: zirconium-based metal organic framework.

^gUiO-66- NH₂@TA (MOF-TA): tannic acid-modified UiO-66- NH₂.

^hLIBS: laser-induced breakdown spectroscopy.

In contrast, the TA/GE-modified cellulose membrane developed in this work emphasizes analytical practicality and sustainability. Although the achieved LOD (2.7 μg·L^–1) is higher than that reported in some studies, , it remains well below the maximum allowable Pb concentration in drinking water established by the World Health Organization (10 μg·L^–1). This level of sensitivity is therefore sufficient for regulatory screening and environmental monitoring purposes. The membrane-based SPE format enables rapid preconcentration at high flow rates with minimal backpressure, while the use of tannic acid and gelatin offers a low-cost, biodegradable alternative to synthetic ligands. When coupled with portable SWASV detection, this approach provides a practical balance between analytical performance and real-world applicability for Pb­(II) monitoring in water samples.

3.6.2. Measurement Uncertainty

The intraday and interday precision of the method were evaluated by analyzing Pb­(II) standard solutions at a concentration of 0.05 mg·L^–1. Intraday precision was determined from five replicate measurements performed within a single day, while interday precision was assessed from nine replicate measurements conducted over 3 days. The relative standard deviations (%RSD) obtained were 3.1% for intraday precision and 2.6% for interday precision, demonstrating good repeatability and acceptable method precision. All measurements were carried out using the same SPGE to avoid potential variability arising from differences in electrode manufacturing.

A CRM, TRM-E-2002 (100 mL), was analyzed to evaluate the accuracy and potential bias of the developed method. The certified concentrations of the metal ions in this CRM are Cd­(II), 0.0050 ± 0.0003 mg·kg^–1; Cu­(II), 0.504 ± 0.011 mg·kg^–1; Pb­(II), 0.061 ± 0.004 mg·kg^–1; and Zn­(II), 0.632 ± 0.018 mg·kg^–1. Since the CRM contains Cu­(II) ions that can interfere with Pb­(II) determination under the applied SWASV conditions, 0.1 M KCN was added to selectively mask Cu­(II) through the formation of stable [Cu­(CN)₄]^3– complexes. Three 50.0 mL aliquots from three CRM bottles were analyzed using the developed method (Sections and ), yielding results of Pb­(II) of 61.1 ± 4.5 μg·L^–1, 60.6 ± 3.9 μg·L^–1, and 60.9 ± 4.9 μg·L^–1, respectively. These values are in good agreement with the certified concentration, thereby confirming the method’s accuracy and reliability.

The combined standard uncertainty, u _c, was calculated as $u_{\mathrm{c}}=\sqrt[2]{u_{\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{c}}^2+u_{\mathrm{b}\mathrm{i}\mathrm{a}\mathrm{s}}^2+u_{\mathrm{r}\mathrm{e}\mathrm{f}}^2}$ , where u _prec is the method precision, u _bias is the measurement bias, and u _ref is the standard uncertainty of the reference value. At 50 μg·L^–1 the precision is 3.2% (1.95 μg·L^–1), the measurement bias is 0 μg·L^–1, and the reference standard uncertainty is 4 μg·L^–1. Thus, the combined standard uncertainty (u _c) of a measured sample at 61.0 μg·L^–1 is 4.45 μg·L^–1. The expanded uncertainty (U), with a coverage factor of 2 for P = 0.95, is 8.9 μg·L^–1.

3.7. Analysis of Real Samples

Analysis of four bottled drinking water samples and four river water samples collected from the Mae Klong River in Kanchanaburi Province, Thailand did not detect any Pb­(II), with concentrations below the method LOD of 2.7 μg·L^–1. In the recovery experiments, spiked drinking water samples yielded recoveries between 82% and 100%, demonstrating good analytical accuracy and negligible matrix effects. Similarly, spiked river water samples provided recoveries ranging from 97% to 110%, confirming high accuracy despite the greater complexity of the natural matrix (data in Supporting Information Table S1). These results validate the suitability of the TA/GE modified filter paper membrane for the preconcentration and quantification of trace Pb­(II) in diverse water matrices. The consistent and satisfactory recovery values further support the applicability of the developed method for routine environmental monitoring and the assessment of drinking water safety.

4. Conclusion

In this study, a novel paper extraction membrane was successfully developed by modifying cellulose filter paper with a TA/GE composite, providing a simple, low-cost, and environmentally friendly platform for preconcentration of Pb­(II) from aqueous samples. The incorporation of TA provided abundant polyphenolic functional groups for Pb­(II) chelation, while GE acted as a biopolymeric matrix to immobilize TA onto the cellulose fibers and enhance the mechanical stability of the membrane. No additional cross-linking agents, such as glutaraldehyde or epichlorohydrin, were required. Electrochemical detection using square-wave anodic stripping voltammetry (SWASV) yielded a linear range of 10–100 μg·L^–1, with a limit of detection (LOD) of 2.7 μg·L^–1. The combined measurement uncertainty (U) at the 95% confidence level for a water sample containing 61 μg·L^–1 Pb­(II) was 9 μg·L^–1.

FTIR analysis confirmed successful TA/GE cross-linking and revealed characteristic spectral changes after Pb­(II) adsorption, supporting the proposed binding mechanism involving phenolic hydroxyl and amino functional groups. Key experimental parameters affecting adsorption and extraction performance, including TA and GE concentrations, cross-linking conditions, membrane dosage, solution pH, and eluent volume, were systematically optimized. Optimal Pb­(II) retention was achieved at pH 5, which is consistent with the determined point of zero charge (pHPZC = 4.74) and favors electrostatic attraction between Pb­(II) ions and the negatively charged membrane surface.

A key advantage of this approach is that the TA/GE membranes can be readily fabricated in a standard laboratory setting using simple procedures, without the variability issues often encountered in cartridge-based SPE systems where inconsistent packing of solid substrates can compromise reproducibility. The membrane format allows straightforward assembly, consistent batch-to-batch quality, and demonstrated low variability in extraction precision across different runs. Application to real drinking water and river water samples yielded recovery values (82–110%) within acceptable ranges, while validation against a certified reference material confirmed the accuracy of the method.

Overall, this work highlights the potential of natural biopolymers and sustainable materials for creating effective analytical tools for trace metal determination. The membrane-based extraction design, which combines operational simplicity, reproducibility, and compatibility with portable electrochemical detection, represents a practical and cost-effective alternative to conventional techniques for environmental monitoring applications.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c12214.

• Photograph of the experimental setup (Figure S1); micro X-ray fluorescence (XRF) images of the TA/GE membrane before ion loading, after metal loading, and after elution (Figure S2); effect of reaction time on GE–TA cross-linking (Figure S3); effect of pH on GE–TA cross-linking (Figure S4); interference study of coexisting metal ions in Pb­(II) analysis (Figure S5); recovery studies in spiked drinking water and river water samples (Table S1) (PDF)

Nuanlaor Ratanawimarnwong: conceptualization, methodology, writingreview and editing, Arinrada Rattanadechawan: investigation and validation, Nathawut Choengchan: validation, Suchao Donpudsa: methodology, Thitirat Mantim: review and editing, Chanpen Karuwan: conceptualization and methodology.

The authors declare no competing financial interest.