Preparation of EDTA-2Na-Fe₃O₄-Activated Carbon Composite and Its Adsorption Performance for Typical Heavy Metals

Abstract

This study developed a new magnetic adsorbent from waste coconut shells using high-temperature carbonization, EDTA-2Na chelation, and Fe₃O₄ magnetic loading. Response surface methodology optimized the preparation conditions to a mass ratio of activated carbon: EDTA-2Na:Fe₃O₄ = 2:0.6:0.2. Characterization (SEM, XRD, FT-IR, and EDS) showed that EDTA-2Na increased the surface carboxyl and amino group density, while Fe₃O₄ loading (Fe concentration 6.83%) provided superior magnetic separation performance. The optimal adsorption conditions of Cu²⁺ by EDTA-2Na-Fe₃O₄-activated carbon composite material are as follows: when pH = 5.0 and the initial concentration is 180 mg/L, the equilibrium adsorption capacity reaches 174.96 mg/g, and the removal rate reaches 97.2%. The optimal adsorption conditions for Pb²⁺ are as follows: when pH = 6.0 and the initial concentration is 160 mg/L, the equilibrium adsorption capacity reaches 157.60 mg/g, and the removal rate reaches 98.5%. The optimal adsorption conditions for Cd²⁺ are pH = 8.0 and an initial concentration of 20 mg/L. The equilibrium adsorption capacity reaches 18.76 mg/g, and the removal rate reaches 93.8%. The adsorption followed the pseudo-second-order kinetics (R² > 0.95) and Langmuir/Freundlich isotherm models, indicating chemisorption dominance. Desorption experiments using 0.1 mol/L HCl and EDTA-2Na achieved efficient desorption (>85%), and the material retained over 80% of its adsorption capacity after five cycles. This cost-effective and sustainable adsorbent offers a promising solution for heavy metal wastewater treatment.

1. Introduction

With the acceleration of the industrialization process, the problem of heavy metal pollution in water bodies has become increasingly severe. Among them, copper (Cu²⁺), lead (Pb²⁺), and cadmium (Cd²⁺) have been listed as priority pollutants for control due to their high toxicity, bio-accumulation, and difficulty in degradation. Industrial wastewater, mining, and agricultural activities lead to their large-scale release, posing a serious threat to the ecosystem and human health [1]. Traditional treatment methods such as chemical precipitation, membrane separation, and ion exchange have problems such as a high risk of secondary pollution, a high cost, and limited applicability [2]. Adsorption methods have become a current research hotspot due to their advantages such as simple operation, low cost, and wide availability of materials [3]. However, traditional activated carbon has defects such as poor adsorption selectivity and difficult recovery. Coconut shell activated carbon, as a representative of biochar, features a high specific surface area, porous structure, and chemical stability [4]. However, the chemical inertness of its original surface limits its adsorption performance. In recent years, the synergistic modification strategy of magnetic nanomaterials and chelating agents has attracted much attention: Fe₃O₄ endows materials with magnetic separation properties and enables rapid recovery [5]; EDTA-2NA enhances the chelating ability of heavy metals through hexagonal coordination. However, the combination of EDTA-2Na and Fe₃O₄ on biomass-derived activated carbon remains rarely explored. Through the collaborative design of high-temperature carbonization, EDTA-2Na chelating modification, and Fe₃O₄ magnetic loading, a new type of EDTA-2Na-Fe₃O₄-activated carbon composite material was prepared, aiming to break through the performance bottleneck of traditional activated carbon, clarify the synergistic adsorption mechanism of chemical coordination and physical diffusion, and develop functionalized materials with efficient adsorption, magnetic separation, and cycling stability. This work establishes a scalable, eco-adaptive paradigm for heavy metal removal—transforming agricultural residues into high-value functional materials aligned with zero-waste water treatment goals.

2. Materials and Methods

2.1. Materials

Coconut shell activated carbon is made from discarded coconut shells and purchased from local suppliers in Hainan. The reagents used in the experiment, including raw materials, modifiers, standard solutions for analysis, etc., were all purchased from professional chemical suppliers.

Standard solutions of copper (Cu(NO₃)₂), lead (Pb(NO₃)₂), and cadmium (Cd(NO₃)₂) elements, used for quantitative analysis and research on the adsorption performance of carbon materials, were purchased from the General Research Institute for Nonferrous Metals in Beijing. Sodium hydroxide (NaOH) and hydrochloric acid (HCl) were used for pH adjustment in the experiments. Methanol (CH₃OH), used as a desorption agent for desorption experiments, was purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China, Ethylenediamine tetraacetic acid disodium salt (C₁₀H₁₄N₂Na₂O₈) and ferrous sulfate (FeSO₄·7H₂O), purchased from Sinopharm Chemical Reagent Co., Ltd., were used to modify carbon materials. Citric acid (C₆H₈O₇) was purchased from Tianjin Comio Chemical Reagent Co., Ltd., Tianjin, China, Dopamine hydrochloride (C₈H₁₁NO₂·HCl), tartaric acid (C₄H₆O₆), and triethanolamine (C₆H₁₅NO₃) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., Shanghai, China.

2.2. Preparation of Activated Carbon Composite Materials

The preparation process of the activated carbon composite material involves four main steps: raw material pretreatment and carbonization, the preparation of chelating agent-modified activated carbon composites, the selection of chelating agents, and synthesis of C-Fe₃O₄-activated carbon composites.

Coconut shells were selected as raw materials. Surface impurities were removed via deionized water-assisted ultrasonic cleaning (frequency: 40 kHz; power: 100 W; duration: 30 min, three cycles). After drying, the shells were crushed and sieved to obtain particles with a size of 60–100 mesh. An amount of 20.0 g of pretreated raw material was placed in a crucible and carbonized in a muffle furnace by heating to 800 °C at a rate of 5 °C/min, and this temperature was maintained for 60 min. The resulting carbonized product was cooled, washed to neutrality (pH ≈ 7), and dried at 105 °C for 2 h to yield black carbonized material.

The activated carbon was mixed with chelating agent solutions at mass ratios of 1:1, 1:2, 1:3, 1:4, and 1:5. The mixtures were subjected to constant-temperature magnetic stirring for 4–6 h, washed to neutrality, and dried to obtain C-activated carbon composites (C denotes the chelating agent).

Five chelating agents—ethylenediamine tetraacetic acid disodium salt (EDTA-2Na), citric acid (CA), dopamine hydrochloride (DA), tartaric acid (TA), and triethanolamine (TEA)—were loaded onto the activated carbon surface via an impregnation method. After pH adjustment, the mixtures were oscillated, washed, and dried for subsequent use.

A total of 10.0 g of X-activated carbon composite (X denotes the selected chelating agent) was ultrasonically dispersed in 200 mL of ultrapure water. Subsequently, 27.81 mL of 0.1 mol/L FeSO₄ solution was added dropwise, followed by 0.1 mol/L NaOH to adjust the pH to ≈ 10. The reaction proceeded for 1 h to generate Fe₃O₄ nanoparticles, which were loaded onto the activated carbon surface. The product was magnetically separated, washed to neutrality, and dried at 105 °C for 2 h to obtain the final C-Fe₃O₄-activated carbon composite.

2.3. Optimization of Experimental Design by Response Surface Method

The Design-Expert response surface experimental design software was adopted to conduct response surface analysis experiments at three factors and three levels [6]. Taking activated carbon (A), chelating agent (B), and iron (III) oxide (C) as variables, the removal effect on copper, a heavy metal, was optimized.

The adsorption performance of the modified materials for Cu²⁺, Pb²⁺, and Cd²⁺ was evaluated through adsorption experiments. The equilibrium adsorption capacity (Q_e) was calculated using Equation (1) to identify the optimal chelating agent [7]:

$$Q_e={\displaystyle \frac{\left(C_0-C_e\right)V}{m}}$$

(1)

Q_e represents the equilibrium adsorption capacity (mg/g); C₀ and C_e represent the residual and initial concentrations of heavy metal ions in the solution (mg/L), respectively; V represents the initial solution volume (L); and m represents the mass of activated carbon adsorbent (g).

The pre-experiment concluded that the optimal value range of A was 1–3 (g), that of B was 0.4–0.8 (g), and that of C was 0.1–0.3 (g). According to the experimental design plan generated by the software, the corresponding masses of chelating agent and ferric oxide were accurately weighed, and they were mixed with activated carbon. The uniform dispersion of each component was ensured through ultrasonic treatment, and then the mixture was used for the adsorption experiment of the copper solution. The experimental design aims to optimize the variable ratio through the system and determine the optimal combination conditions to achieve the goal of efficient copper removal. The experimental design factors and levels are shown in

Table 1. Experimental factor coding and levels.

Value Level	Influence Factor
	A—Coconut Shell Activated Carbon (g)	B—Chelating Agent (g)	C—Fe3O4 (g)
−1	1	0.4	0.1
0	2	0.6	0.2
1	3	0.8	0.3

.

2.4. Characterization of Activated Carbon Composites

We used field emission scanning electron microscopy (SEM; JSM-7610F PLUS, JEOL, Tokyo, Japan) operating at an accelerating voltage of 10 kV to obtain high-resolution images. Before analysis, gold plating treatment should be carried out to reduce the influence of charge accumulation on the analysis results and improve surface conductivity [8]. The gold plating thickness was 5–10 nm, the magnification was ×3000, the working distance was 8–10 mm, and the scanning speed was medium. We used energy dispersive X-ray spectroscopy (EDS; N/A, Oxford Instruments, Abingdon, UK) integrated with SEM and operated at an acceleration voltage of 10 kV–20 kV to ensure the excitation of characteristic X-rays while increasing the beam current to enhance the signal strength and controlling the acquisition time to balance data accuracy and analysis efficiency. Fourier Transform Infrared Spectroscopy (FT-IR ALPHA, Thermo Fisher Scientific, Waltham, MA, USA) was used to analyze the surface chemical functional groups of the raw carbon and the modified composite materials. The samples were dried at 105 °C for 2 h, and the chemical functional groups were analyzed and identified within the wavenumber range of 400–4000 cm⁻¹ using the KBr pellet method. We used an X-ray diffractometer (XRD; D/max2500v/pc, Shimadzu Corporation, Kyoto, Japan) under the conditions of a tube voltage of 40 kV, a tube current of 150 mA, an angle range of 10–50°, a scanning speed of 2°/min, and a scanning step size of 0.05° to analyze the crystal structure, phase composition, and content of the raw carbon and the modified composite materials.

2.5. Single-Factor Adsorption Experiment

At 25 °C, a certain amount of standard solutions of copper, lead, and cadmium (1000 mg/L) was, respectively, placed into 100 mL volumetric flasks, and the volumes were made up with deionized water to obtain the standard solutions of copper, lead, and cadmium. Batch adsorption experiments were conducted. All samples were oscillated and reacted at a rotational speed of 200 rpm in a constant temperature shaker for 12 h to ensure that the adsorption process reached dynamic equilibrium. After the reaction was completed, the sample was centrifuged and separated at a speed of 3000 rpm for 10 min. The supernatant was taken, and the remaining concentrations of Cu²⁺, Pb²⁺, and Cd²⁺ were determined by atomic absorption spectrometry. By calculating the adsorption capacity and removal rate, we evaluated the influence of the dosage of adsorbent (1–5 g/L), adsorption time (0–240 min), pH value (3–9), initial concentrations of Cu²⁺ and Pb²⁺ (100–200 mg/L), and Cd²⁺ (5–30 mg/L) on the adsorption performance of EDTA-2Na-Fe₃O₄-activated carbon composites.

2.6. Regeneration Experiment

The adsorption-saturated EDTA-2Na-Fe₃O₄-activated carbon composite material (adsorbing Cu²⁺, Pb²⁺, and Cd²⁺) was mixed with different desorption reagents (0.1 mol/L HCl, 0.1 mol/L EDTA-2Na, methanol, and deionized water) at a solid–liquid ratio of 1:50. We placed it in a constant temperature shaker for oscillation. After the process was completed, the solid and liquid phases were separated. After centrifugal separation, the concentration of heavy metals in the desorption liquid was determined, the desorption efficiency was calculated, and the optimal desorption agent was screened. The material was dynamically eluted by the desorption agent. After desorption, the material was washed with an desorption agent—deionized water—selected from the above experiment until it became neutral, dried at 105 °C for 2 h, and used for the next round of the adsorption experiment. We repeated the above adsorption–desorption steps five times and recorded the adsorption capacity and desorption efficiency of each cycle.

2.7. Research on Adsorption Mechanism

2.7.1. Research on Adsorption Kinetics

Under the constant temperature condition of 25 °C, 10 g of EDTA-2Na-Fe₃O₄-activated carbon composite material was added to 2 L of Cu²⁺, Pb²⁺, and Cd²⁺ solutions with initial concentrations of 180 mg/L, 160 mg/L, and 20 mg/L, respectively. We adjusted the initial pH value of the solution, and the oscillate reaction was carried out using a shaker at 200 rad/min. Samples were taken at reaction times of 5 min, 15 min, 30 min, 45 min, 60 min, 120 min, 180 min, and 240 min. After centrifugal separation, the concentration of heavy metal ions in the supernatant was determined, and the adsorption capacity was calculated. To clarify the kinetic mechanism of the adsorption process, the following two commonly used kinetic models were adopted to fit and analyze the experimental data:

1.

Pseudo-First-Order Model

The pseudo-first-order kinetic model assumes that the adsorption rate is determined by the number of active sites on the surface of the adsorbent, and its formula is as follows [9]:

$$Q_t=Q_e\left(1-e^{-kt}\right)$$

(2)

Q_e and Q_t represent the equilibrium adsorption capacity and the adsorption capacity at time t (mg/g), respectively; k represents the pseudo-first-order kinetic rate constant (min⁻¹); and t represents the reaction time (min).

2.

Pseudo-Second-Order Model

The pseudo-second-order kinetic model assumes that the adsorption rate is determined by the chemical interaction between the surface active sites of the adsorbent and the adsorbate, and its formula is

$$Q_t={\displaystyle \frac{Q_e^2{k}_{2}t}{\left(1+Q_e{k}_{2}t\right)}}$$

(3)

Q_e and Q_t represent the equilibrium adsorption capacity and the adsorption capacity at time t (mg/g), respectively; k₂ represents the pseudo-second-order kinetic rate constant (g/(mg·min)); and t represents the reaction time (min).

3.

Intraparticle Diffusion Model

The intraparticle diffusion model is a key kinetic model that describes the diffusion of adsorbate from the solution to the internal pores of the adsorbent during the adsorption process. Its formula is as follows [10]:

$$Q_t=k_p{t}^{0.5}+C$$

(4)

Q_t represents the adsorption capacity (mg/g) at time t; k_p represents the intraparticle diffusion coefficient mg/(g·min^0.5); and t represents the reaction time (min).

2.7.2. Research on Adsorption Isotherms

Under the constant temperature condition of 25 °C, Cu²⁺ and Pb²⁺ solutions with different initial concentrations of 100, 120, 140, 160, 180, and 200 mg/L, as well as Cd²⁺ solutions with concentrations of 5, 10, 15, 20, 25, and 30 mg/L, were prepared. The pH was adjusted to the optimal adsorption conditions (based on single-factor optimization conditions). We added 10 g of modified carbon material to the solutions of each concentration, placed them in a constant-temperature shaker at 200 rad/min, and carried out oscillation and reaction for 240 min. After the reaction was completed, centrifugation was performed to separate the solid and liquid phases, and the supernatant was taken. The residual concentration of heavy metals in the supernatant was determined by atomic absorption spectrometry.

To explore the equilibrium adsorption capacity of EDTA-2Na-Fe₃O₄-activated carbon composites for other heavy metals, we clarified the single-layer or multi-layer characteristics of the adsorption process and processed the isothermal adsorption kinetics experimental data using the Langmuir and Freundlich isotherm models.

4.

Langmuir isothermal adsorption model:

$$Q_e={\displaystyle \frac{K_L{Q}_m{C}_e}{1+b{C}_e}}$$

(5)

Q_e and Q_m represent the equilibrium adsorption capacity and the maximum adsorption capacity (mg/g), respectively; C_e represents the concentration at equilibrium (mg/L); and K_L is the Langmuir constant.

5.

Freundlich isothermal adsorption model:

$$Q_e=K_F{C}_e^{{\displaystyle \frac{1}{n}}}$$

(6)

Q_e represents the equilibrium adsorption capacity (mg/g); C_e represents the concentration at equilibrium (mg/L); and K_F is the Freundlich constant.

3. Results and Discussion

3.1. Selection of Chelating Agents and Fe₃O₄ Range

By systematically investigating the effects of different types of chelating agents and the loading of Fe₃O₄ on the removal performance of Cu²⁺, Pb²⁺, and Cd²⁺, the chelating agent and the loading of Fe₃O₄ with the best adsorption effect were screened out.

It can be seen from Figure 1 that in (a), EDTA-2Na and TEA with a content of 0.2 g/g have a better adsorption effect on Cu²⁺, and in subfigure (b), EDTA-2Na and TEA with a content of 0.2 g/g have a better adsorption effect on Pb²⁺. Subfigure (c) shows that EDTA-2Na and TEA with contents of 0.4 g/g and 0.8 g/g have better adsorption effects on Cd²⁺. EDTA-2Na’s dominance stems from its hexadentate coordination structure, enabling robust chelation with multiple metal ions through high-stability constants (logβ: Cu²⁺ = 18.8, Pb²⁺ = 18.3, and Cd²⁺ = 16.5). In contrast, other agents exhibit limited coordination capacity: CA (tridentate ligands with weak complexation ability), DA (catechol has dentate coordination and is prone to oxidation and inactivation), TA (α-hydroxy acid has a bidentate coordination and weak complexation ability), and TEA (single-dentate ligands with low affinity) form weaker complexes. After comparison, EDTA-2Na was finally selected to modify the carbon material.

As shown in Figure 1, in order to increase the efficiency of recycling and utilization of modified materials and enhance the adsorption capacity, Fe₃O₄ was selected to modify carbon materials [11]. It can be seen from subfigure (d) that the content range of Fe₃O₄ with a better removal effect on Cu²⁺ is 0.1–0.3 g/g, the content range of Fe₃O₄ with a better removal effect on Pb²⁺ is 0.1-0.3 g/g, and the content range of Fe₃O₄ with a better removal effect on Cd²⁺ is 0.2–0.5 g/g. Finally, carbon materials with an Fe₃O₄ content of 0.1–0.3 g/g were selected for response surface analysis, and subsequent experiments were carried out.

3.2. An Analysis of the Experimental Results of Response Surface Optimization

The experiment was designed through the response surface method (RSM), and the experimental data were fitted using the experimental design software. The quadratic regression equation obtained is

Y = 2.94 − 0.031A − 0.066B + 0.07C − 0.055AB + 0.012AC + 0.037BC − 0.51A² − 0.17B² − 0.49C²

The reliability of the regression model determines its prediction accuracy. It is verified through analysis of variance and provides a scientific basis for experimental optimization [12].

As shown in

Table 2. Analysis of variance of response surface model.

Source	Sum of Squares	df	Mean Squares	F Value	Value Prob > F
Model	2.5458	9	0.2829	233.6379	<0.0001	significant
A	0.0078	1	0.0078	6.4528	0.0387
B	0.0351	1	0.0351	29.0015	0.0010
C	0.0392	1	0.0392	32.3776	0.0007
AB	0.0121	1	0.0121	9.9941	0.0159
AC	0.0006	1	0.0006	0.5162	0.4957
BC	0.0056	1	0.0056	4.6460	0.0681
A2	1.1059	1	1.1059	913.4451	<0.0001
B2	0.1181	1	0.1181	97.5718	<0.0001
C2	1.0109	1	1.0109	835.0008	<0.0001
Residual	0.0085	7	0.0012
Lack of Fit	0.0045	3	0.0015	1.4917	0.3447	not significant
Pure Error	0.0040	4	0.0010
Cor Total	2.5543	16

, the F-value of 233.64 implies that the model is significant. In the current model, items such as A, B, C, AB, AC, and BC have all reached the significance level, indicating that the contributions of these items to the model are statistically significant. Furthermore, the F-value of the misfitting test was 1.4917, and the corresponding p-value was 0.3447, indicating that misfitting was not significant compared to the pure error.

As shown in Figure 2 and

Table 3. Analysis of variance for quadratic regression equations.

Std. Dev.	0.034	R-Squared	0.9964
Mean	2.39	Adj R-Squared	0.9929
C.V.%	1.41	Pred R-Squared	0.9783
PRESS	0.055	Adeq Precision	44.978

, the closer the sum of squares of the residuals is to 0, the more it indicates that the model can fit the experimental data well and accurately predict the changes in the response values. In this study, the sum of squares of the predicted residuals of the model was calculated to be 0.055, indicating that the prediction error of the model is small and that it has a high prediction accuracy. The prediction coefficient R² = 0.9783 has a high degree of consistency with the correction coefficient R² = 0.9929. This indicates that there is a good match between the predictive ability of the model and the goodness of fit, further verifying the reliability of the model. Under normal circumstances, a signal-to-noise ratio (SNR) > 4 is considered acceptable; the current model has a signal-to-noise ratio as high as 44.978, which is much higher than this threshold. This indicates that the model can effectively distinguish signals from noise and has high prediction accuracy and stability.

The analysis was conducted using Design-Expert software. It was speculated that the optimal ratio for removing the three heavy metals was A:B:C = 1.98:0.56:0.21. To facilitate subsequent experiments, the ratio was adjusted to A:B:C = 2:0.6:0.2.

3.3. Material Characterization

It can be observed from Figure 3a that a certain number of pores and tiny concave–convex structures are distributed on the material surface. This is because the material underwent high-temperature pyrolysis at 800 °C during the preparation process, resulting in partial surface cracking and the formation of gully structures. Compared with subfigure (a), the pores in subfigure (b) are more developed, and the surface presents more irregular gully structures. This morphology is observed due to the fact that after the material is modified by EDTA-2Na, the surface is further damaged, forming more depressions and pores [13]. Compared with subfigures (a) and (b), the pore distribution in subfigure (c) is more uniform and the pore size is smaller, indicating that the surface structure of the material was optimized after further chemical treatment or loading with ferric oxide nanomaterials. This uniform pore structure provides the material with more adsorption sites, which helps to improve its adsorption performance. Comparing subfigure (d) and subfigure (e), after the original coconut shell activated carbon was carbonized at 800 °C, carbon (C) and oxygen (O) were the main elements, with atomic concentrations of 90.45% and 9.55%, respectively, and the weight concentration proportions were 87.67% and 12.33%. After modification by EDTA-2Na and Fe₃O₄, the atomic concentration of C significantly decreased to 81.86%, and the weight concentration dropped to 63.35%, while the atomic concentration of Fe reached 6.83%, and the weight concentration increased to 24.59%. Meanwhile, the atomic concentration of Na element was detected to be 0.92%. This indicates that Fe₃O₄ nanoparticles were successfully loaded onto the surface of the activated carbon, and the introduction of EDTA-2Na caused residue of Na element. In subfigure (f), the three samples all have obvious absorption peaks at 3430 cm⁻¹ and 1349 cm⁻¹, which can be attributed to the stretching vibration of hydroxyl groups in carboxyl groups, phenols, and alcohols. The peak at 1696 cm⁻¹ underwent a slight shift and an increase in intensity. This is because the amide HN-C=O that appears overlaps with the original C=O bond, indicating that the amidation reaction between the protocarbon and the carboxyl group (-COOH) of EDTA-2Na was successful and the introduction of EDTA-2Na was successful. The vibration peak of Fe-O appeared at 601 cm⁻¹ in the EDTA-2Na-Fe₃O₄-modified carbon material, indicating that Fe₃O₄ particles were loaded onto the surface of activated carbon. In subfigure (g), the characteristic diffraction peaks of coconut shell activated carbon molecules are at positions of 24.8° and 43.8° with 2θ, while the characteristic diffraction peaks of Fe₃O₄ crystals are at positions of 30.2°, 35.6°, 43.2°, 53.5°, 57°, and 62.9° with 2θ. The diffraction patterns of the EDTA-2Na-Fe₃O₄-modified carbon material were compared with those of the carbonized activated carbon. The modified activated carbon simultaneously showed characteristic diffraction peaks of Fe₃O₄ and the carbonized activated carbon, indicating that Fe₃O₄ was successfully loaded onto the surface of the activated carbon.

3.4. Single-Factor Adsorption Experimental Analysis

As shown in Figure 4, the following analysis can be performed:

(1)

The economic dosing thresholds of Cu²⁺, Pb²⁺, and Cd²⁺ are 3 g/L, 2 g/L, and 4 g/L, respectively. At this time, the adsorption capacity per unit mass and the removal rate reach equilibrium. Excessive addition leads to no significant change in the adsorption efficiency due to the particle aggregation effect, indicating that excessive addition has a limited effect on increasing the adsorption capacity.

(2)

The adsorption capacity increases with the increase in the concentration gradient. When the initial concentration reaches 180 mg/L (Cu²⁺), 160 mg/L (Pb²⁺), and 20 mg/L (Cd²⁺), the adsorption capacity approaches the maximum value. Further increasing the initial concentration will lead to a stabilization of the adsorption capacity, while the removal rate decreases due to the saturation of the adsorption sites. Therefore, the optimal initial concentrations are selected as Cu²⁺ 180 mg/L, Pb²⁺ 160 mg/L, and Cd²⁺ 20 mg/L. At this time, the adsorption capacity is close to the maximum value.

(3)

The removal rate of the adsorption process increased rapidly within the initial 60 min (Cu²⁺, Pb²⁺, and Cd²⁺ reached more than 85% of the equilibrium adsorption capacity, respectively), and then the adsorption rate slowed down significantly, basically reaching equilibrium after 240 min. From an economic perspective, the comprehensive benefit of adsorption efficiency and time cost is optimal at 60 min. Therefore, the adsorption time for subsequent experiments is uniformly set at 60 min.

(4)

At a low pH (3.0–5.0), H⁺ competes with heavy metal ions for adsorption sites, resulting in a relatively weak adsorption capacity. When the pH rises to 5.0–6.0, the deprotonation of surface functional groups enhances electrostatic attraction, and the adsorption capacity increases significantly. When pH further increases (>6.0), the adsorption capacity of Cu²⁺ and Pb²⁺ decreases due to the formation of hydroxide precipitates (Cu(OH)₂ and Pb(OH)₂) or hydroxyl complexes. Cd²⁺ begins to form Cd(OH)₂ precipitates and Cd(OH)₃⁻ and other complexes when pH>8, resulting in a decrease in adsorption capacity.

3.5. Regeneration Experiment Analysis

As shown in Figure 5, Cu²⁺ and Pb²⁺ were treated with 0.1 mol/L HCl as the desorption agent. Through the competition of H⁺ for adsorption sites, the coordination between heavy metals and the material surface was disrupted. The desorption rate was >85%, and the desorption efficiency was remarkable. An amount of 0.1 mol/L EDTA-2Na was selected as the desorption agent for Cd²⁺, and its desorption effect is the best because of its strong chelating ability, which forms a soluble complex with Cd²⁺. The desorption efficiency of methanol and deionized water is less than 20%, indicating that heavy metals mainly combine through chemical adsorption, and the physical desorption effect is limited. After five adsorption–desorption cycles, the retention rates of adsorption capacity of the material for Cu²⁺, Pb²⁺, and Cd²⁺ were 83%, 85%, and 80%, respectively, indicating that the material has good regeneration stability. The decline in adsorption capacity is mainly caused by two synergistic factors: physical corrosion (the desorbent (HCl/EDTA-2Na) dissolves Fe₃O₄ nanoparticles, destroying the pore structure and reducing the active surface area) and irreversible site occupation (the type of adsorption is mainly chemical adsorption, which consumes a large number of adsorption sites).

3.6. Analysis of Adsorption Mechanism

Adsorption Kinetics Analysis

As shown as Figure 6,

Table 4. Adsorption kinetics parameters of heavy metal ions in EDTA-2Na-Fe₃O₄-activated carbon composites.

	Pseudo-First-Order Model			Pseudo-Second-Order Model
	K1 (min−1)	Qe (mg/g)	R2	K2 (g/(mg·min))	Qe (mg/g)	R2
Cu	0.012	168.82	0.9241	0.00591	169.49	0.9986
Pb	0.018	131.2	0.9852	0.00525	131.58	0.9984
Cd	0.016	16.88	0.9124	0.059019	16.95	0.9522

and

Table 5. Intra-particle diffusion model parameters of heavy metal ions in EDTA-2Na-Fe₃O₄-activated carbon composites.

	Intraparticle Diffusion Model
	K1 (g/mg/min0.5)	C1 (mg/g)	R12	K2 (g/mg/min0.5)	C2 (mg/g)	R22	K3 (g/mg/min0.5)	C3 (mg/g)	R32
Cu	10.55071	81.3309	0.99679	2.73737	136.57061	0.87643	0.41067	161.51277	0.81805
Pb	14.85224	38.76026	0.98086	4.35142	94.7643	0.87643	0.27721	126.81259	0.8659
Cd	0.80018	7.99424	0.97641	0.43794	10.75559	0.99222	0.03254	16.30941	0.89743

, Kinetic studies of Cu²⁺, Pb²⁺, and Cd²⁺ by EDTA-2Na-Fe₃O₄-activated carbon composites show that this adsorption process is more in line with the pseudo-second-order kinetic model (R² is 0.9986, 0.9984, and 0.9522, respectively), indicating that chemical adsorption is the main rate-controlling step. The equilibrium adsorption capacities can reach 169.49, 131.58, and 16.95 mg/g, respectively. This is due to the combination of the active sites on the surface of the composite material with Cu²⁺, Pb²⁺, and Cd²⁺ through coordination bonds or ion exchange. The intra-particle diffusion model further reveals the three-stage mechanism of the adsorption process: the initial stage is the rapid adsorption stage, mainly dominated by membrane diffusion of heavy metal ions migrating to the active sites on the material surface, and the mid-term adsorption rate decreases. At this time, heavy metal ions diffuse into the internal pores of the particles. This process is affected by the pore structure and concentration gradient of the material. The final adsorption is close to equilibrium. At this point, the surface active sites are basically saturated, and residual adsorption may be driven by a small number of unoccupied sites or weak interactions. The comprehensive analysis of the kinetics and diffusion models indicates that in the early stage of adsorption, surface chemical adsorption is dominant, while in the later stage, it is jointly regulated by intraparticle diffusion and surface reactions [14,15].

3.7. Adsorption Isotherm

As shown as Figure 7 and

Table 6. Parameters of isothermal adsorption curves of the Langmuir model and the Freundlich model.

	Langmuir			Freundlich
	Qmax (mg/g)	KL (L/mg)	R2	KF (mg/g)	1/n	R2
Cu	181.84	0.00207	0.96619	3.10221	0.76574	0.99959
Pb	162.31	0.00825	0.93661	12.87275	0.45627	0.90977
Cd	20.36	0.04187	0.97062	2.40471	0.62008	0.94158

, through the fitting analysis of the Langmuir and Freundlich isotherm models, the differences in adsorption characteristics of EDTA-2Na-Fe₃O₄-activated carbon composite material for different heavy metals were revealed. The adsorption behavior of Cu²⁺ is more in line with the Freundlich model (R² = 0.9996), and its non-uniform multilayer adsorption characteristics result from the synergistic coordination effect of carboxyl groups (-COOH) and amino groups (-NH₂) on the material surface. The Freundlich constant K_F = 3.10. And the index 1/n = 0.7657 indicates that the adsorption process is spontaneous and that the distribution of active sites is uneven [16,17,18]. The adsorption of Pb²⁺ and Cd²⁺ was more in line with the Langmuir model (R² > 0.93), and the maximum adsorption capacities were Q_max = 157.60 mg/g and 18.76 mg/g, respectively, indicating that monolingual chemical adsorption was dominant, and Cd²⁺ showed a stronger affinity for the hexatentate chelating structure of EDTA-2Na.

4. Conclusions

In this study, by using waste coconut shells as raw materials, through the synergetic strategy of high-temperature carbonization, EDTA-2Na chelating modification, and Fe₃O₄ magnetic loading, EDTA-2Na-Fe₃O₄-activated carbon composite material (AC-EDTA-Fe₃O₄) with both efficient adsorption and magnetic separation characteristics was successfully prepared. And its adsorption performance and mechanism regarding the heavy metals Cu²⁺, Pb²⁺, and Cd²⁺ in water bodies were systematically explored. Through the processes of high-temperature carbonization at 800 °C combined with ultrasonic cleaning and mechanical crushing, and characterized by SEM, XRD, and FT-IR, it was confirmed that EDTA-2Na modification significantly enhanced the densities of functional groups such as carboxyl groups (-COOH) and amino groups (-NH₂) on the material surface. Fe₃O₄ is loaded by the co-precipitation method (XRD shows characteristic peaks of 30.2° and 35.6°, and the Fe atomic concentration detected by EDS reaches 6.83%), endowing the material with magnetic separation ability (separation efficiency > 90%). Among the five chelating agents, namely EDTA-2Na, CA, DA, TA, and TEA, EDTA-2Na had the best comprehensive adsorption performance for Cu²⁺, Pb²⁺, and Cd²⁺, with removal rates reaching 86.73%, 94.45%, and 93.83%, respectively. The response surface method indicates that the optimal preparation ratio is activated carbon (EDTA-2Na:Fe₃O₄ = 2:0.6:0.2), and the model is highly significant (F = 233.64, p < 0.0001).

Adsorption experiments showed that the adsorption performance of AC-EDTA-Fe₃O₄ is significantly affected by pH. The adsorption of Cu²⁺ is optimal at pH = 5.0 (initial concentration of 180 mg/L, equilibrium adsorption capacity of 174.96 mg/g, and removal rate of 97.2%). The adsorption of Pb²⁺ was optimal at pH = 6.0 (initial concentration of 160 mg/L, equilibrium adsorption capacity of 157.60 mg/g, and removal rate of 98.5%). The adsorption of Cd²⁺ was optimal at pH = 8.0 (initial concentration of 20 mg/L, equilibrium adsorption capacity of 18.76 mg/g, and removal rate of 93.8%). Kinetic studies show that adsorption reaches equilibrium within 60 min, with a removal rate of >85%. The intra-particle diffusion model reveals that it is divided into three stages: In the initial stage (0–40 min), membrane diffusion is dominant (the diffusion coefficient (K₁) values of Cu²⁺, Pb²⁺, and Cd²⁺ were 10.55, 14.85, and 0.80 mg/(g·min^0.5), respectively). In the middle stage (40-60 min), pore diffusion restriction occurs (K₂ decreases to 2.74, 4.35, and 0.44 mg/(g·min^0.5) respectively), and in the final stage (>60 min), it tends to be in dynamic equilibrium. The pseudo-second-order kinetic model (R² > 0.95) indicates that the type of adsorption is mainly chemical adsorption, involving the coordination and complexation of functional groups and metal ions. The isothermal adsorption experiment shows that Cu²⁺ adsorption conforms to the Freundlich model (R² = 0.99959) and presents non-uniform multilayer adsorption characteristics. However, Pb²⁺ and Cd²⁺ are more in line with the Langmuir model (R² > 0.93), and monolingual chemical adsorption is dominant.

Furthermore, the material exhibits excellent regeneration performance: a concentration of 0.1 mol/L HCl can effectively desorb Cu²⁺ and Pb²⁺, and EDTA-2Na is suitable for the desorption of Cd²⁺. After three cycles, the regeneration efficiency is >85%, after five cycles, the retention rate of adsorption capacity is >80%, and the magnetic separation efficiency is always >95%, which significantly reduces the treatment cost. In conclusion, the AC-EDTA-Fe₃O₄ composite material achieves efficient adsorption and convenient recovery of heavy metals through the synergistic effect of functional group modification and magnetic loading, and it has broad application prospects in wastewater treatment.