Adsorption performance and mechanism of MgAl-LDHs@FRBCs/Fe₃O₄ composite for multiple heavy metal ions in the aqueous extract of Ligusticum chuanxiong Hort

Abstract

Heavy metal contamination in traditional Chinese medicines (TCMs) and their preparations is an urgent quality and safety issue to be addressed in the industrialization of TCMs. Herein, a novel magnetic composite of MgAl-LDHs@FRBCs/Fe₃O₄ was prepared as the adsorbent, which was fabricated by using the leaves (Isatidis Folium) and roots (Isatidis Radix) of Isatis indigotica as the carbon source for biochar (FRBCs), followed by in situ growth of MgAl-layered double hydroxides (MgAl-LDHs). In the aqueous extract of Ligusticum chuanxiong Hort. (CX), the composite exhibited high simultaneous removal efficiencies for multiple heavy metal ions: 96.53% for Pb²⁺, 78.49% for Cu²⁺, 83.82% for Cd²⁺, and 35.48% for Ni²⁺. The adsorption process fitted better with the pseudo-second-order (PSO) kinetic model and Langmuir isotherm, indicating that the adsorption was dominated by monolayer chemisorption. Meanwhile, a certain correlation with the pseudo-first-order (PFO) kinetic model was also observed, reflecting the participation of the physisorption. Thermodynamic parameters (ΔH = 25.84–105.42 kJ mol⁻¹, ΔS = 122.15–386.15 J mol⁻¹ K⁻¹, ΔG = −17.37 to −9.65 kJ mol⁻¹) confirmed a spontaneous and endothermic process with increased disorder. Further mechanism investigation by pH-dependent experiments, zeta potential measurements, and XPS revealed that the adsorption occurred mainly through electrostatic interactions and chemical complexation. Importantly, HPLC analysis revealed no significant alteration of the major bioactive components in the CX extract after adsorption. The composite also exhibited fast adsorption kinetics, high removal efficiency under competitive conditions, good stability and reusability, and easy magnetic separation. The current work provides a high-performance adsorbent for heavy metal removal from TCM aqueous extracts, which also offers a sustainable strategy for valorizing TCM residues.

Herein, a novel magnetic composite of MgAl-LDHs@FRBCs/Fe₃O₄ was prepared, and the adsorption performance and mechanism of this composite for multiple heavy metal ions in the aqueous extract of Ligusticum chuanxiong Hort. were investigated.

1. Introduction

Due to the obvious bioaccumulation and toxicological characteristics of heavy metals, the long-term intake of them may lead to liver and kidney damage, nervous system disorders, endocrine disorders and immune dysfunction, thus posing potential health risks.¹ In recent years, the application scale of TCMs and their preparations has been continuously expanding worldwide, and the attention to the quality and safety of TCMs has been growing.² Heavy metal contamination is considered as a common concern in the quality control of TCMs. Multiple studies have shown that heavy metal ions can be introduced during the planting, harvesting, processing and storage of Chinese medicinal materials.³ Therefore, the effective removal of heavy metals from TCM extracts and preparations is important for the modernization development of TCMs.

At present, the common removal methods for heavy metal contamination include adsorption, ion exchange, precipitation and membrane separation.^4,5 Among them, the adsorption method is considered as a promising approach due to its simple operation, economic feasibility, mild conditions and potential for recycling.^6–8 However, the existing commonly used adsorbents, such as activated carbon, resins and metal oxides, still have drawbacks such as limited specific surface area, insufficient active sites and difficulty in recovery, which make them difficult to meet the treatment requirements of multi-metal and multi-component systems.⁹

Traditional Chinese medicine residue, a waste generated during the extraction process of TCM industrial production, is a kind of underutilized biomass resource with the characteristics of stable source, abundant output and sufficient carbon content.^10,11Isatidis Folium and Isatidis Radix are the main raw materials of Banlangen Granules, a commonly used TCM preparation. The active ingredients in Isatidis Radix mainly include indole alkaloid glycosides, polysaccharides, isothiocyanates and thiocyanates, while Isatidis Folium is rich in numerous bio-active compounds such as indican, indigo, indirubin, tryptanthrin and flavonoids.^12,13 Carbonization of the above herb residues can not only retain their porous structure, but also form abundant functional groups including hydroxyl, carboxyl and amino groups on the surface of biochar, providing a variety of complexation and adsorption sites for metal ions.^14,15 Therefore, as compared with other conventional biomass such as corn stalks and waste wood, biochar from TCM residue has unique advantages in terms of the types of functional groups, the number of active sites and adsorption potential.^16,17

Layered double hydroxides (LDHs) are formed by metal cations and feature adjustable interlayer structure, strong anion exchange ability and abundant surface hydroxyl groups, and thus exhibit good performance in the adsorption of heavy metal ions.¹⁸ In addition, they have good chemical and thermal stability and can be recycled in aqueous systems, making them important candidates for constructing high-efficiency multi-metal adsorbents.¹⁹ Magnetic Fe₃O₄ nanoparticles can endow materials with magnetic response characteristics when used as adsorbents, facilitating rapid separation and recovery after adsorption.²⁰ The construction of multi-component composites using TCM residue biochar, LDHs and Fe₃O₄ is expected to realize increased specific surface area, abundant active sites, synergistic adsorption mechanisms and magnetic separation advantages.

In this study, biochar (FRBCs) was prepared from the residues of Isatidis Folium and Isatidis Radix, which was then doped with Fe₃O₄ magnetic nanoparticles, followed by the in situ growth of MgAl-LDHs to form the MgAl-LDHs@FRBCs/Fe₃O₄ composite (Fig. 1). The composite was expected to achieve efficient adsorption of multiple heavy metal ions through the synergy of multiple components, providing an effective method for the adsorption of heavy metals in aqueous extract of TCM.

Fig. 1. Schematic illustration of (A) the synthesis process of the MgAl-LDHs@FRBCs/Fe₃O₄ composite, and (B) the adsorption performance study of heavy metal ions.

2. Experimental section

2.1. Materials and reagents

All the used reagents were of analytical grade (AR) or chromatographic grade (GR). Anhydrous MgCl₂, AlCl₃·6H₂O, FeCl₃·6H₂O, anhydrous sodium acetate, phosphoric acid, ethylene glycol and NaOH were purchased from Aladdin (Beijing, China). The mixed herb residues of Isatidis Folium and Isatidis Radix were provided by a Chinese medicinal material production enterprise (Chongqing, China), the medicinal material of Ligusticum chuanxiong Hort. was purchased from the local Chinese medicinal material market (Chongqing, China). Concentrated nitric acid (HNO₃) and hydrogen peroxide (H₂O₂) were purchased from Chuandong Chemical Co., Ltd (Chongqing, China). Standard solutions of metal ions including Pb²⁺, Cd²⁺, Ni²⁺, Cu²⁺, Fe³⁺, Mg²⁺ and Al³⁺ were all purchased from NCS Testing Technology Co., Ltd (Beijing, China). Acetonitrile (GR) and methanol (GR) were purchased from Honeywell (Charlotte, USA). In addition, ultrapure water (18.2 MΩ cm⁻¹) was used for the preparation of the solutions.

2.2. Instruments and conditions

The crystal structure of the materials was analyzed by X-ray diffractometer (XRD, Rigaku Ultima IV, Japan), and the surface functional groups were determined by Fourier transform infrared spectrometer (FTIR, Thermo Fisher Nicolet iS50, USA). The microscopic morphology and lattice structure were observed by transmission electron microscope (TEM, JEOL JEM-2100F, Japan). The element composition and chemical states of the materials were determined by X-ray photoelectron spectrometer (XPS, Thermo Scientific K-Alpha, USA), and the specific surface area and pore structure characteristics were analyzed by Automatic specific surface area and porosity analyzer (Micromeritics ASAP 2460, USA).

The concentration of metal ions in the solution was determined by inductively coupled plasma mass spectrometry (ICP-MS, PerkinElmer NexION 1100 G, USA). The content determination of active ingredients and fingerprint analysis of Ligusticum chuanxiong Hort. were conducted by high performance liquid chromatography (HPLC, waters HPLC-2998 PDA, USA).

2.3. Synthesis of adsorbents

Firstly, the Isatidis Folium and Isatidis Radix residues were cleaned and dried. Then the residues were crushed with a grinder and sieved through a 100-mesh sieve. The obtained powder was pyrolyzed at 600 °C for 2 h in a tube furnace under nitrogen atmosphere to form the carbonized material. After that, the carbonized material was washed with 10 wt% HCl for 3–5 times, then washed with ultrapure water until pH = 7, and then dried at 80 °C for 24 h, which was designated as FRBCs.

Secondly, magnetic Fe₃O₄ nanoparticles were prepared by a solvothermal method. Specifically, FeCl₃·6H₂O (16.2 g) and CH3COONa (32.4 g) were dissolved in 240 mL of ethylene glycol, stirred at 25 °C for 1 h, then transferred to a Teflon-lined autoclave and kept at 200 °C for 10 h. After natural cooling to room temperature, the mixture was thoroughly washed with water and ethanol. Subsequently, it was dried at 55 °C for 15 h.²¹

Finally, MgAl-LDHs were in situ grown on the FRBCs/Fe₃O₄ composite by a hydrothermal method. The FRBCs (1 g) and Fe₃O₄ (1 g) were ultrasonically dispersed in 100 mL of deionized water. Then, anhydrous MgCl₂ (0.03 mol) and AlCl₃·6H₂O (0.01 mol) were added to the suspension, followed by vigorous stirring and ultrasonic treatment at room temperature. Subsequently, the mixture was adjusted to pH = 10 with NaOH aqueous solution (1 M), and then subjected to hydrothermal treatment at 180 °C for 24 h.²² The resulting MgAl-LDHs@FRBCs/Fe₃O₄ composite was collected by centrifugation, washed repeatedly, and then freeze-dried. MgAl-LDHs were prepared in the same way without adding biochar and Fe₃O₄.

2.4. Preparation of aqueous extract of Ligusticum chuanxiong Hort

The dried medicinal material of Ligusticum chuanxiong Hort. (80 g) was crushed and sieved (100-mesh), transferred into a 2000 mL round-bottom flask with 10 times the amount of water, then refluxed for 2 h, and filtered while hot. Suspended particles were separated by centrifugation (25 °C, 8000 rpm, 5 min), and the resulting supernatant was passed through a membrane filter (0.22 µm) and kept at 4 °C until use.

2.5. Adsorption behavior study

To carry out the adsorption behavior study, the optimization of adsorption condition, adsorption isothermal experiments and adsorption kinetics experiments were performed. A control group without spiked standard ions (non-spiked) and experimental groups with spiked metal ion standard solutions (Pb²⁺, Cu²⁺, Ni²⁺, Cd²⁺) were involved, and all the experiments were performed in triplicate. Following the adsorption experiments, the samples were centrifuged (8000 rpm, 10 min) for collecting the supernatant, then treated by a 0.22 µm polytetrafluoroethylene (PTFE) filter membrane. The obtained solutions were transferred to a PTFE microwave digestion tank, and the mixture of HNO₃ and H₂O₂ (1 : 1, v/v) was added for digestion to ensure the complete release of metal ions. After being diluted to an appropriate concentration, the digested solution was determined by ICP-MS.

The removal efficiency and maximum adsorption capacity of the materials were determined by eqn (1) and (2):²³

1

2

where η (%) is the adsorption efficiency of the material, C₀ (µg mL⁻¹) is the initial concentration of metal ion, C_e (µg mL⁻¹) is the equilibrium concentration of metal ion, q_e (mg g⁻¹) is the equilibrium adsorption capacity of the material to metal ion, m (mg) is the dosage of the adsorbent, and V is the volume of the solution (mL).

To achieve the optimal adsorption conditions, the initial pH, the adsorption time and the dosage of adsorbent were optimized. The effect of initial pH on adsorption was further evaluated at pH value of 2.3, 3.3, 4.3, and 5.3, and the pH was adjusted using dilute HNO₃ or NaOH before adsorption. The adsorption time was optimized with a fixed adsorbent dosage, with a contact time varying from 0 to 720 min. The dosage of adsorbent was optimized in the range of 0 to 100 mg, with a fixed contact time.

The adsorption kinetics studies were carried out by adding 80 mg of adsorbent into 40 mL of metal ion solution (1 µg mL⁻¹). The mixture was then placed in a thermostatic shaker and agitated at 298 K and 220 rpm. The adsorption times were preset at 30, 60, 120, 180, 240, 360, and 720 min. The pseudo first order (PFO), pseudo second order (PSO) and Elovich models were used to fit the adsorption kinetic data, as shown in eqn (3)–(5).²⁴ The competition adsorption measurements were conducted under conditions similar to those of the adsorption kinetics studies using the optimized adsorption time.

ln(q_e − q_t) = ln q_e − k₁t 3

4

5

where q_t (mg g⁻¹) is the adsorption capacity at time t; k₁ (1 min⁻¹) is the first-order adsorption rate constant; k₂ (g mg⁻¹ min⁻¹) is the second-order adsorption rate constant; α (mg g⁻¹ min⁻¹) is the initial adsorption rate and β (g mg⁻¹) is a constant related to surface heterogeneity.

The adsorption isothermal experiments were performed by adding 80 mg of adsorbent into 40 mL of metal ions solution with metal ion concentration in the range of 0.25–150 µg mL⁻¹. Under optimal adsorption conditions, the equilibrium adsorption behavior was recorded by varying the initial metal ion concentration. The adsorption isothermal experiments were carried out at 298 K, 308 K and 318 K, respectively. The resulting data were fitted using the Langmuir, Freundlich, and Temkin models, as shown in eqn (3)–(5), respectively.²⁵ The Langmuir model assumes that adsorption occurs on a homogeneous surface and forms a monolayer coverage, the Freundlich model is used to describe multilayer adsorption on a heterogeneous surface and the Temkin model considers that the adsorption heat decreases linearly with the increase of surface coverage.²⁵

6

7

q_e = B ln K_T + B ln C_e 8

where K_L (L mg⁻¹) is the Langmuir adsorption equilibrium constant; q_m (mg g⁻¹) is the maximum monolayer adsorption capacity; K_F is the Freundlich complexation capacity constant; 1/n is the surface heterogeneity factor; K_T (L mg⁻¹) is the Temkin adsorption equilibrium constant; and B is the Temkin constant related to adsorption heat.

The thermodynamic adsorption parameters were also measured at 298 K, 308 K and 318 K, and the Gibbs free energy change (ΔG), enthalpy change (ΔH), entropy change (ΔS) in the process of adsorption were calculated by the following eqn (9)–(12):²³

9

ΔG = −RT ln K_c 10

11

ΔG = ΔH − TΔS 12

where K_c is the thermodynamic equilibrium constant, ΔG (kJ mol⁻¹) is the standard Gibbs free energy change, R (8.314 J (mol⁻¹ K⁻¹)) is the molar gas constant, T (K) is the absolute temperature, ΔH (kJ mol⁻¹) is the enthalpy change, and ΔS (J (mol⁻¹ K⁻¹)) is the entropy change.

2.6. Stability and reusability studies

To evaluate the stability of the adsorbent, the XRD determinations were carried out on the adsorbent before adsorption, after adsorption and desorption. Specifically, after the adsorption of metal ion on adsorbent, the desorption was performed by 1 wt% HNO₃, and the material was washed with deionized water several times and then freeze-dried. To further investigate the reusability of the adsorbents, the adsorption–desorption cycle was repeated five times to determine the change of adsorption efficiency.

2.7. Metal leaching rate studies

To investigate the metal leaching rate of the adsorbent, 80 mg of the adsorbent was added to 40 mL of Ligusticum chuanxiong Hort. extract (unspiked) and shaken under 298 K for 12 h. Then, the mixture was treated as described in Section 2.5 to determine the content of Mg²⁺, Al³⁺ and Fe³⁺ by ICP-MS.

2.8. Content determination of active ingredients and similarity evaluation of fingerprints of CX extract

Ferulic acid, senkyunolide A and senkyunolide I were used as the representative active ingredients of the Ligusticum chuanxiong Hort. The content of these compounds and the fingerprint of CX extract before and after the removal of heavy metal ions were measured by HPLC.

The HPLC equipped with a 2998 photodiode array (PDA) detector was employed to measure the active ingredients and fingerprint of CX extract. The chromatographic column used was Waters SunFire C18 Column (150 × 4.6 mm, 5 µm), the mobile phase was methanol (organic phase) – 0.1% phosphoric acid aqueous solution (aqueous phase), the detection wavelength was 276 nm, the flow rate was 0.8 mL min⁻¹, and the injection volume was 10 µL. The content of each active ingredient was calculated according to the standard curve from the chromatographic peak area, and the fingerprint similarity of each group was calculated by the relative peak area normalization method.

3. Results

3.1. Characterization of adsorbents

3.1.1. SEM characterization

To analyze the microscopic morphology of each material, SEM characterization was performed on MgAl-LDHs, Fe₃O₄, FRBCs and the MgAl-LDHs@FRBCs/Fe₃O₄ composite. As shown in Fig. 2A, MgAl-LDHs exhibited a typical layered structure. Fe₃O₄ presented a nearly spherical structure with a diameter of about 100–800 nm (Fig. 2B). As presented in Fig. 2C, the FRBCs featured complex pore structures on the surface and inside, which provided good support for the subsequent loading.

Fig. 2. SEM images of (A) MgAl-LDHs, (B) Fe₃O₄, (C) FRBCs, and (D) MgAl-LDHs@FRBCs/Fe₃O₄.

As shown in Fig. 2D, the MgAl-LDHs nanosheets were densely grown on the surface of FRBCs/Fe₃O₄. As Compared to FRBCs, MgAl-LDHs and Fe₃O₄, the surface of the composite was more complex, which may provide more active sites, thus laying a foundation for its better adsorption performance. On the whole, the SEM results clearly confirmed that both Fe₃O₄ and MgAl-LDHs were successfully anchored on FRBCs.

3.1.2. TEM characterization

The TEM images (Fig. 3A) showed that MgAl-LDHs presented a typical layered structure with a certain degree of stacking between the nanosheets. The lattice fringes with an average spacing of 0.265 nm were observed, which were ascribed to the (012) crystal plane of MgAl-LDHs (PDF card #35-0965). The TEM image of FRBCs (Fig. 3B) showed an irregular block structure. Lattice fringes with an average lattice spacing of 0.236 nm could be observed in its HRTEM image, which could be attributed to the in-plane lattice of the graphitic carbon-like structure, corresponding to the (100) crystal plane.²⁶ At the same time, there were areas in the HRTEM image where no obvious lattice fringes were observed, indicating that the FRBCs contained both amorphous carbon and carbon structural units with short-range order characteristics, which was consistent with the typical structural characteristics of biomass char.²⁶ Fe₃O₄ nanoparticles (Fig. 3C) presented an approximately spherical morphology. Lattice fringes with a spacing of about 0.248 nm were observed in the HRTEM image, which could be ascribed to the (311) crystal plane of Fe₃O₄ (PDF card #65-3107), demonstrating that the prepared Fe₃O₄ nanoparticles had good crystallinity.

Fig. 3. TEM images of (A) MgAl-LDHs, (B) FRBCs, (C) Fe₃O₄, (D) MgAl-LDHs@FRBCs/Fe₃O₄ and (E–L) EDS mapping images of the MgAl-LDHs@FRBCs/Fe₃O₄.

Obvious sheet-like structures, spherical nanoparticles and irregular block structures were presented in the TEM images of MgAl-LDHs@FRBCs/Fe₃O₄ composite (Fig. 3D). Lattice fringes with spacings of about 0.200 nm and 0.255 nm were displayed simultaneously in the HRTEM image, corresponding to the (018) plane of MgAl-LDHs (PDF card #35-0965) and the (311) plane of Fe₃O₄ (PDF card #65-3107), respectively. The EDS mapping images (Fig. 3E–L) revealed that the C element was mainly distributed in the FRBCs region. Mg and Al elements were uniformly distributed, consistent with the distribution characteristics of MgAl-LDHs. The Fe element signal was concentrated in the spherical particle region, corresponding to Fe₃O₄ nanoparticles. The O element was widely present in the entire field of view, which was contributed by the oxygen-containing groups in each component. The above results indicated that the components and the main elements (C, N, O, Mg, Al, Cl, Fe) coexisted in the prepared composite.

3.1.3. FTIR characterization

To further determine the main functional groups in materials, FT-IR tests were performed, as shown in Fig. 4A and B. A broad and obvious absorption band appeared at 3564.49 cm⁻¹ in the FTIR spectrum of MgAl-LDHs, corresponding to the vibration of –OH and H₂O. The absorption peak at 1633.73 cm⁻¹ could be ascribed to the vibration of interlayer water molecules. The vibration of CO₃²⁻ resulted in the peak at 1372.46 cm⁻¹, indicating the presence of CO₃²⁻ as balancing anions between the MgAl-LDHs layers.²⁷ The vibration peaks of M–O (Mg–O/Al–O) could be observed in the low wave number region (600–500 cm⁻¹), which were attributed to the interlayer structure of MgAl-LDHs.²⁸ In the FTIR spectrum of FRBCs, the absorption peaks in the range of 1380–1537 cm⁻¹ were related to the C C skeleton vibration of aromatic structures or carboxylate groups,²⁹ the absorption peak at 1037.28 cm⁻¹ was attributed to C–O. The peaks in the range of 400–900 cm⁻¹ were dense and sharp, indicating the presence of aromatic rings with multiple chemical states in the carbon layer structure.³⁰ In the FTIR spectrum of Fe₃O₄, a strong absorption peak appeared at 582.15 cm⁻¹, corresponding to the stretching vibration of Fe–O bonds, indicating the formation of Fe₃O₄.³¹

Fig. 4. (A and B) FT-IR spectra, (C) XRD patterns, (D) XPS survey spectra of FRBCs, Fe₃O₄, MgAl-LDHs, and the composite; (E–J) high-resolution XPS spectra of the MgAl-LDHs@FRBCs/Fe₃O₄.

For MgAl-LDHs@FRBCs/Fe₃O₄ composite, the typical peaks of the three components could be observed simultaneously in its FTIR spectrum, indicating that all the functional groups were maintained in the composite. An obvious vibration peak of C C aromatic rings or carboxylate groups appeared at about 1537.62 cm⁻¹ in the composite, and the C–O stretching vibration at 1037.65 cm⁻¹ was still visible, indicating the successful loading of FRBCs. The absorption peaks at 1633.50 cm⁻¹ and 1371.67 cm⁻¹ were still clearly visible, indicating the layered structure and interlayer carbonate of MgAl-LDHs were maintained during the formation of the composite. In addition, an obvious absorption peak appeared at 609.06 cm⁻¹ in the low wave number region, corresponding to the vibration of M–O (Mg–O, Al–O, Fe–O) bonds. In summary, FTIR analysis indicated that MgAl-LDHs, FRBCs and Fe₃O₄ were effectively combined in the composite, and the original groups of the three components were all retained.

3.1.4. XRD characterization

Fig. 4C shows the XRD patterns of MgAl-LDHs, Fe₃O₄ and FRBCs. Two remarkable diffraction peaks of MgAl-LDHs appeared at 11.040° and 22.220°, which were attributed to the (003) and (006) planes, respectively. In addition, several weaker diffraction peaks were also observed at around 34.600°, 39.010°, 45.500° and 60.420°, corresponding to the (012), (015), (018) and (110) planes, respectively. The XRD pattern is consistent with the standard card (#35-0965), indicating the formation of MgAl-LDHs. In the XRD pattern of Fe₃O₄, the diffraction peaks appeared at 30.060°, 35.420°, 43.020°, 53.360°, 56.940° and 62.500°, which were attributed to the (220), (311) (400), (422), (511) and (440) crystal planes, respectively. The above peaks are in accordance with the PDF standard card (#65-3107), indicating the successful synthesis of Fe₃O₄. The diffraction peaks of FRBCs appeared at 26.62°, 40.480° and 50.140°, attributed to the ordered graphitic carbon, corresponding to the crystal planes of (002), (100) and (102).³²

The main characteristic diffraction peaks of MgAl-LDHs, Fe₃O₄ and FRBCs were retained in the XRD pattern of the MgAl-LDHs@FRBCs/Fe₃O₄ composite, and the positions of the diffraction peaks showed no obvious shift, indicating that the preparation process of the composite did not damage the original crystal structures of each component. Compared to the single component, the intensity of the characteristic diffraction peaks of some components in the composite was weakened, which may be due to their high dispersion in the composite or to being covered by other components. In summary, XRD patterns confirmed the successful construction of the composite.

3.1.5. XPS characterization

The XPS measurement was employed to analyze the element composition and chemical states of the prepared materials, and the survey spectra were presented in Fig. 4D. The MgAl-LDHs were mainly composed of Mg, Al and O elements, Fe₃O₄ was composed of Fe and O elements, and FRBCs mainly contained C, O and N elements. The MgAl-LDHs@FRBCs/Fe₃O₄ composite contained C, O, N, Mg, Al and Fe elements, indicating that MgAl-LDHs and Fe₃O₄ had been successfully loaded on the FRBCs surface. Furthermore, high-resolution XPS spectra were collected on MgAl-LDHs@FRBCs/Fe₃O₄ (Fig. 4E–J), Fe₃O₄ (Fig. S1), MgAl-LDHs (Fig. S2) and FRBCs (S3), to confirm the successful preparation of the materials.

The C 1s spectrum of the MgAl-LDHs@FRBCs/Fe₃O₄ composite is shown in Fig. 4E, which can be deconvoluted into three peaks at 284.7 eV (C–C), 286.1 eV (C–O–C) and 289.0 eV (O–C O). The Fe₃O₄ (Fig. S1A), MgAl-LDHs (Fig. S2A) and FRBCs (Fig. S3A) presented similar C 1s spectra, indicating the close chemical environment of carbon. However, the carbon-containing groups of Fe₃O₄ and MgAl-LDHs mainly come from the carbon contamination of air.³² Therefore, the C–C, C–O–C and O–C O bonds in the composite mainly originate from FRBCs. As presented in Fig. 4F, the O 1s spectrum of the MgAl-LDHs@FRBCs/Fe₃O₄ composite presented four peaks attributed to Fe₃O₄ (529.9 eV),³³ C–O/O–H (531.7 eV),³⁴ C O (532.5 eV) and O C–O (533.8 eV).³⁵ The O 1 s spectrum of Fe₃O₄, MgAl-LDHs, FRBCs were presented in Fig. S1C, Fig. S2E and S3C, respectively. The results indicate that the O–H of MgAl-LDHs@FRBCs/Fe₃O₄ mainly contributed by MgAl-LDHs, the metal–oxygen bonds (M–O) of MgAl-LDHs@FRBCs/Fe₃O₄ mainly contributed by Fe₃O₄. The oxygen-containing groups in MgAl-LDHs@FRBCs/Fe₃O₄ composite may serve as potential hydrogen bond donors/acceptors and metal ion coordination sites, which provide a variety of chemical interaction sites for the subsequent adsorption process.^36,37

The N 1s spectrum (Fig. 4G) shows the presence of pyridinic N (398.3 eV), pyrrolic N (400.0 eV), graphitic N (401.3 eV)³⁸ and N-oxide in MgAl-LDHs@FRBCs/Fe₃O₄. FRBCs present similar N 1s spectrum (Fig. S3B), indicating that the nitrogen-containing functional groups were retained in the composite. Pyridinic N, pyrrolic N and graphitic N mainly originate from nitrogen-containing alkaloids, indole compounds and protein residues generated during the carbonization process. These nitrogen-containing functional groups provided active sites for the adsorption of metal ions, and graphitic N could adjust the electron density to synergistically enhance the adsorption performance.^37,39 The peak with high binding energy was nitrogen oxide, which might be introduced by the adsorption of a small amount of nitrogen oxide in the air.

The Fe 2p spectrum (Fig. 4H) of MgAl-LDHs@FRBCs/Fe₃O₄ composite confirmed the simultaneous presence of Fe²⁺ and Fe³⁺, including Fe²⁺ 2p_3/2 (710.2 eV), Fe³⁺ 2p_3/2 (713.2 eV), satellite peak (718.0 eV), Fe²⁺ 2p_1/2 (723.2 eV), Fe³⁺ 2p_1/2 (726.0 eV) and satellite peak (732.0 eV).^40,41 Compared with Fe 2p spectrum of Fe₃O₄ (Fig. S1B), the peak positions have no obvious shift, indicating the chemical state of Fe₃O₄ remained stable during the formation process of composite. Thus providing stable metal active sites in composite, which are favorable for the adsorption of metal ions.⁴² Mg 1s spectrum of MgAl-LDHs@FRBCs/Fe₃O₄ composite including one peaks at 1304.2 eV, ascribed to Mg–OH (Fig. 4I).^43,44 Al 2p spectrum of the composite can be fitted into two peaks at 75.0 eV and 75.5 eV, which correspond to Al 2p_3/2 and Al 2p_1/2 (Fig. 4J).⁴⁵ The Mg 1 s and Al 2p spectra are similar to those of MgAl-LDHs.

3.1.6. BET characterization

The pore structure characteristics of MgAl-LDHs, Fe₃O₄, FRBCs and their composite were characterized by N₂ adsorption–desorption method (Fig. 5A–B). According to the IUPAC classification standard,⁴⁶ all samples exhibited typical Type IV isotherms, indicating that the materials were mainly mesoporous. In the region of relatively low pressure (P/P₀ < 0.1), the adsorption capacity of all samples increased slowly with the increase of pressure, indicating the presence of a certain number of micropores in the materials. An obvious hysteresis loop (H3 type) appeared in the medium and high relative pressure range, which was usually related to capillary condensation occurring in mesopores. The specific surface areas of MgAl-LDHs and Fe₃O₄ were 48.42 m² g⁻¹ and 22.48 m² g⁻¹, respectively, and their pore structures mainly originated from the mutual stacking of flake MgAl-LDHs or spherical Fe₃O₄ nanoparticles. The specific surface area of FRBCs was low (11.60 m² g⁻¹), which might be related to the collapse of some pore channels and the formation of disordered structures during the carbonization of TCM residues. Compared with FRBCs, the specific surface area of the MgAl-LDHs@FRBCs/Fe₃O₄ composite increased to 31.35 m² g⁻¹, which was not exceed that of MgAl-LDHs, indicating that the growth of MgAl-LDHs and the loading of Fe₃O₄ occupied part of the pore channels of FRBCs.

Fig. 5. N₂ adsorption–desorption isotherms (A and B) and BJH pore size distribution curves (C–F) of MgAl-LDHs, Fe₃O₄, FRBCs, and MgAl-LDHs@FRBCs/Fe₃O₄.

To further analyze the pore structure characteristics of the materials, the BJH method was employed to confirm the mesoporous and macroporous distribution of the materials, and the HK method was used to calculate the microporous distribution (Fig. 5C–F). Among them, MgAl-LDHs showed a wide range of mesopores and macropores (2–200 nm), relatively concentrated at about 120 nm, and a small number of micropores were relatively concentrated at 0.7 nm and 1.2 nm. The pore size of FRBCs was widely distributed in the range of 2–224 nm, relatively concentrated at 22 nm, 34 nm and 53 nm, and the micropores were distributed in the range of 0.8–1.9 nm. The mesopores and macropores of Fe₃O₄ were widely distributed in the range of 2–208 nm, the mesopores were relatively concentrated at about 15 nm and 19 nm, and the micropores were distributed in the range of 0.8–1.9 nm. Finally, the pore size of MgAl-LDHs@FRBCs/Fe₃O₄ was widely distributed in the range of 2–188 nm, with macropores relatively concentrated (about 104 nm). The micropores were distributed in the range of 0.8–1.9 nm, relatively concentrated at about 0.9 nm.

3.1.7. VSM characterization

To investigate the magnetic recovery performance of the composite material, the magnetic properties of equal masses of MgAl-LDHs@FRBCs/Fe₃O₄ and Fe₃O₄ were characterized using a vibrating sample magnetometer (VSM) over an applied magnetic field range of −25 kOe to 25 kOe. The results are presented in Fig. S4. The saturation magnetization of Fe₃O₄ was determined to be 78.7 emu per g, whereas that of the MgAl-LDHs@FRBCs/Fe₃O₄ composite was 17.8 emu per g. This reduction in magnetization is primarily attributed to the introduction of non-magnetic components (MgAl-LDHs and FRBCs). The results demonstrate that the composite material retains a significant magnetic response, indicating its suitability for magnetic separation under an applied magnetic field.⁴⁷

3.2. Adsorption behavior investigation results in the aqueous extract of Ligusticum chuanxiong Hort

3.2.1. Comparison of heavy metal removal performance of different adsorbents

As presented in Fig. 6A, the adsorption efficiencies of FRBCs, Fe₃O₄, MgAl-LDHs and MgAl-LDHs@Fe₃O₄@FRBCs were compared. Among them, FRBCs show adsorption capability for Pb²⁺, Cu²⁺, Cd²⁺ and Ni²⁺, with the efficiency of 59.71%, 14.59%, 27.92% and 1.99%, respectively. Fe₃O₄ show adsorption capability for Pb²⁺, Cd²⁺ and Ni²⁺, with the efficiency of 53.22%, 63.20%, 2.16%, respectively. MgAl-LDHs show adsorption capability for Pb²⁺ and Cd²⁺, with the efficiency of 86.36%, 82.78%, respectively. MgAl-LDHs@Fe₃O₄@FRBCs show adsorption capability for Pb²⁺, Cu²⁺, Cd²⁺ and Ni²⁺, with the efficiency of 96.53%, 78.49%, 83.82%, 35.48%, respectively.

Fig. 6. Effect of (A) material type, (B) pH value (C) adsorption time and (D) adsorbent dosage on the removal efficiency of Pb²⁺, Cd²⁺, Cu²⁺, and Ni²⁺; The adsorption kinetics of Pb²⁺, Cd²⁺, Cu²⁺, and Ni²⁺ onto MgAl-LDHs@FRBCs/Fe₃O₄ (E) and fitted lines with PFO (F), PSO (G), and Elovich (H) models.

The above results indicate that the MgAl-LDHs@Fe₃O₄@FRBCs composite exhibits the strongest adsorption performance, compared with the other three materials. Specifically, the composite exhibits excellent adsorption capability for Pb²⁺, Cu²⁺, Cd²⁺, with relatively limited adsorption efficiency on Ni²⁺. The MgAl-LDHs@Fe₃O₄@FRBCs composite presented the highest adsorption rates as compared to the FRBCs, Fe₃O₄ and MgAl-LDHs, which may be due to the composite have more diverse active adsorption sites, more complex pore structure, and suitable specific surface area.

3.2.2. Effect of initial pH on heavy metal adsorption

To investigate the effect of initial pH on the adsorption performance of heavy metals, adsorption experiments were conducted at pH values of 2.3, 3.3, 4.3, and 5.3. As shown in Fig. 6B, the removal efficiency of most metal ions exhibited an increasing trend with increasing pH. Under strongly acidic conditions, the presence of a large number of H⁺ ions may lead to excessive protonation of functional groups on the material surface, thereby reducing the affinity of the adsorbent for metal ions.⁴⁸ As the pH increases, partial deprotonation of the functional groups occurs, reducing the occupancy of negatively charged sites and thus promoting electrostatic interactions with metal cations.⁴⁹ As the pH increased from 4.3 to 5.3, the resulting adsorption efficiency of Pb²⁺, Cd²⁺, and Ni²⁺ exhibited a slight increase, while the adsorption efficiency of Cu²⁺ decreased slightly. Considering that pH = 4.3 is closer to the natural state of the CX extract, and taking into account the simplicity and operability of subsequent procedures, subsequent adsorption experiments were mainly conducted at pH = 4.3.

3.2.3. Effect of adsorption time and adsorbent dosage on heavy metal adsorption

As presented in Fig. 6C, the adsorption efficiency of the MgAl-LDHs@FRBCs/Fe₃O₄ composite for ions improved with increasing time. It displayed a rapid growth trend in the range of 0–120 min and slowed down after 120 min, thus the optimal adsorption time was confirmed to be 120 min. The possible reason for this phenomenon might be that there were a large number of adsorption sites on the material surface at the initial stage of adsorption, and the adsorption sites were gradually occupied with the passage of time, leading to a slow change in adsorption efficiency until it remained unchanged.⁵⁰

As shown in Fig. 6D, the effect of the adsorbent dosage on adsorption efficiency was investigated. When the dosage increased from 10 mg to 80 mg, the number of available adsorption sites in the system added, thus promoting the removal of ions in the solution and significantly increasing the adsorption efficiency. However, when the dosage continuously increased from 80 mg to 120 mg, the adsorption efficiency showed no obvious change, which might be attributed to the saturation of adsorption sites.⁵¹ In addition, excessive dosage might lead to poor dispersion, reducing the specific surface area, thus inhibiting further removal of metal ions.⁵²

3.2.4. Adsorption kinetics

As shown in Fig. 6E, the adsorption capacity (q_e) increased rapidly at the initial stage of adsorption, then the growth rate slowed down gradually and reached adsorption equilibrium after about 120 min. To further explore the kinetic characteristics of the adsorption, PFO, PSO and Elovich models were used to fit the experimental data. The fitting results from the three models are presented in Fig. 6F–H and Table 1. The linear relationship between t/q_t and t indicates that, the adsorption kinetics of Pb²⁺, Cu²⁺, Cd²⁺, and Ni²⁺ on MgAl-LDHs@FRBCs/Fe₃O₄ are mostly consistent with PSO model (R² = 0.9803–0.9997), which reflects that the adsorption process may be mainly ascribed to chemisorption (Fig. 6G).⁵³ In addition, the fitting results also exhibit a certain correlation with PFO model (Fig. 6F, R² = 0.9270–0.9968), implying that the adsorption process is also accompanied by physical adsorption.⁵⁴ In contrast, the Elovich model showed a relatively poor fit, with R² range from 0.5952 to 0.8602 (Fig. 6H).

Table 1. The fitting results of Kinetic models for the adsorption process of heavy metal ions onto the MgAl-LDHs@FRBCs/Fe₃O₄ composite in CX extract. (k represents k₁ (min⁻¹) for PFO model or k₂ (g mg⁻¹ min⁻¹) for PSO model).

Kinetic models	Metal ions	R 2	q e (mg g−1)	k (min−1 or g mg−1 min−1)
Pseudo-second-order	Pb2+	0.9955	0.79	1.98
	Cu2+	0.9997	1.40	0.36
	Cd2+	0.9890	0.87	1.50
	Ni2+	0.9803	0.76	2.22
Pseudo-first-order	Pb2+	0.9685	1.00	0.01795
	Cu2+	0.9270	2.18	0.01749
	Cd2+	0.9968	2.19	0.03016
	Ni2+	0.9517	1.07	0.01591
Elovich	Pb2+	0.8602	—	—
	Cu2+	0.6949	—	—
	Cd2+	0.5952	—	—
	Ni2+	0.6813	—	—

To provide a more comprehensive evaluation of the three adsorption kinetic models described above, a further analysis was conducted using four error indicators including SSE (Sum of Squared Errors), ARE (Average Relative Error), SAE (Sum of Absolute Errors), and HYBRID (Hybrid Fractional Error Function), the corresponding results are presented in Table S1.^55,56 Among the three models, the PSO model exhibits the lowest values of SSE (0.008–0.096), ARE (5.75–12.71), SAE (0.22–0.63), and HYBRID (0.31–2.11). The PFO model presented the highest values of SSE(0.188–5.176), ARE(40.06–162.66), SAE(0.80–4.49), and HYBRID(18.03–361.76). The Elovich model show the moderate error values among the investigated models, with the SSE, ARE, SAE, and HYBRID values ranging from 0.015–0.133, 7.25–12.10, 0.29–0.85, and 0.48–2.52, respectively. Based on the comprehensive analysis of R² and the error functions, the PSO model is the most suitable for describing the adsorption kinetics of the heavy metal ions on the prepared composite material in this study, while the PFO model can only describe part of the kinetic behavior. Overall, the adsorption process of the investigated metal ions on the MgAl-LDHs@FRBCs/Fe₃O₄ composite may be dominated by chemisorption, which may also be accompanied by physical adsorption.

3.2.5. Adsorption isotherms

The adsorption isotherms of Cd²⁺, Ni²⁺, Cu²⁺ and Pb²⁺ on the composite were displayed in Fig. 7A–D. The q_e of ions on the composite increased rapidly with the increasing initial concentration (C₀) and gradually reached equilibrium in the high-concentration region, indicating that the adsorption process reached equilibrium within the experimental concentration range. The adsorption isotherm data were fitted by the Freundlich, Langmuir and Temkin models (Fig. 7E–H, S5A–H). The adsorption of all the four metal ions on MgAl-LDHs@FRBCs/Fe₃O₄ composite fitted well with the Langmuir model (R² = 0.9968–0.9993), which was significantly better than that with the Freundlich (R² = 0.3178–0.8224) and Temkin (R² = 0.2975–0.9552) models.

Fig. 7. (A–D) The adsorption isotherms of Cd²⁺, Cu²⁺, Ni²⁺ and Pb²⁺ onto MgAl-LDHs@FRBCs/Fe₃O₄, (E–H) fitted lines with the Langmuir model.

The three adsorption models were further evaluated through error analysis, and the corresponding results are presented in Table S2. Compared with the Freundlich and Temkin models, the Langmuir model exhibited lower values of SSE (7.88–29.67), ARE (6.16–16.23), SAE (4.32–8.48), and HYBRID (9.90–38.40), indicating that the Langmuir model can more accurately describe the adsorption isotherm behavior of each metal ion in this study.⁵⁵ The Freundlich model showed generally higher error function values, with SSE, ARE, SAE, and HYBRID ranging from 73.75 to 3859.85, 40.1 to 368.2, 16.62 to 139.77, and 274.25 to 3372.80, respectively, which are significantly higher than those of the Langmuir model. The Temkin model also exhibited high error values, suggesting its poor applicability to the present system. Based on the comprehensive evaluation of R² and the error analysis results, the isothermal adsorption process of the metal ions in this study is better described by the Langmuir model.

The poor fitting of the Freundlich model indicated that the surface heterogeneity of the material and the contribution of multi-layer adsorption were limited; the insufficient fitting of the Temkin model suggested that the adsorption heat changed slightly with the surface coverage during the adsorption process, and the interaction between adsorption sites was weak.⁵⁷ The above results demonstrated that the adsorption of heavy metal ions by the composite conformed to a typical monolayer adsorption model, and the adsorption process might involve chemical complexation, hydrogen bonding and electrostatic interaction.⁵⁸

The maximum adsorption capacity (q_m) could be further achieved from the fitting results of the Langmuir model, and the order was q_m (Cu²⁺) > q_m (Pb²⁺) > q_m (Ni²⁺) > q_m (Cd²⁺). Among them, q_m of Cu²⁺ and Pb²⁺ was obviously higher than that of Cd²⁺ and Ni²⁺, implying that the composite material had an obvious selective adsorption behavior for heavy metals. The q_m, K_L and R_L for metal ions calculated by the Langmuir model were listed in Table 2. Compared with previously reported adsorbents for heavy metal ions (Table 3), the MgAl-LDHs@FRBCs/Fe₃O₄ composite prepared in this work exhibits a moderate adsorption capacity for each ion. However, it should be noted that this adsorption capacity represents the theoretical value obtained under non-competitive conditions, and may differ from the actual value in practical applications due to the competitive adsorption among different ions.

Table 2. The fitting results of Langmuir isotherm for Pb²⁺, Cd²⁺, Ni²⁺, and Cu²⁺ adsorption onto the MgAl-LDHs@FRBCs/Fe₃O₄ in ultrapure water.

Langmuir constant	Heavy metal ions
	Pb2+	Cd2+	Ni2+	Cu2+
q m (mg g−1)	82.24	47.35	9.82	135.87
K L(L mg−1)	5.43	1.43	4.95	1.52
R 2	0.9989	0.9967	0.9982	0.9974
R L	0.002–0.424	0.007–0.737	0.002–0.447	0.006–0.725

Table 3. Comparison of maximum adsorption capacities of reported adsorbents for Pb²⁺, Cd²⁺, Ni²⁺, and Cu²⁺.

Adsorbent	Ions and qmax (mg g−1)	Isotherm model	Ref.
ZnAl-LDH-BC	Cd2+(99.8)	Langmuir, Freundlich	59
	Pb2+(128. 4)
Mn-BC	Cd2+(47.13)	Langmuir	60
	Pb2+(128. 4)
	Ni2+(18.92)
Ac-FeMnLDH-BC	Pb2+(457.10)	Langmuir	61
	Cu2+(81.25)
	As5+(5.39)
nFeS-CG	Cd2+(10.52)	Langmuir	62
	Pb2+(12.61)
CMCMW-Bent	Cd2+(44.97)	Langmuir	63
Zr-BADS MOFs	Cd2+(53.58)	Langmuir	64
	Pb2+(60.75)
Red clay (RC)	Cu2+(24.90)	Langmuir	65
White clay(WC)	Cu2+(68.02)	Langmuir	65
MgAl-LDHs@FRBCs/Fe3O4	Cd2+(47.13)	Langmuir	This work
	Pb2+(82.24)
	Ni2+(9.82)
	Cu2+(135.87)

The q_m values of the four ions was calculated in the range of 9.82–135.87 mg g⁻¹, which was basically consistent with the actually measured value of 9.83 mg g⁻¹ −134.79 mg g⁻¹. The Langmuir constant (K_L) reflects the affinity of the material for metal ions under low-concentration conditions. The relatively high K_L values of Pb²⁺ and Ni²⁺ indicate that the material possesses strong adsorption affinity for these two ions at low initial concentrations. The calculated separation factor (R_L) was in the range of 0.002–0.737 within the studied concentration range, indicating that the adsorption processes of the material for Pb²⁺, Cd²⁺, Ni²⁺and Cu²⁺ were all favorable. Especially at low initial concentrations, the small value of R_L indicated a strong adsorption driving force. This result was consistent with the high value of K_L, further proving that the Langmuir model had good applicability in the studied concentration range.⁶⁶

To investigate the effect of temperature on the adsorption process, the thermodynamic adsorption parameters were calculated by at different temperatures (298 K, 308 K, 318 K), and shown in Table 4.

Table 4. Thermodynamic parameters and Langmuir isotherm constants for heavy metal ions adsorption on the MgAl-LDHs@FRBCs/Fe₃O₄ at different temperatures^a.

Temperature (K)	Constants	Heavy metal ions
		Pb	Cd	Ni	Cu
298 K	ΔG (kJ mol−1)	−13.05	−10.56	−9.65	−11.72
	ΔH (kJ mol−1)	+42.18	+25.84	+105.42	+31.56
	ΔS (J mol−1 K−1)	+185.34	+122.15	+386.15	+145.22
	q m (mg g−1)	82.24	47.35	9.82	135.87
	R 2	0.9989	0.9967	0.9982	0.9974
308 K	ΔG (kJ mol−1)	−14.90	−11.78	−13.51	−13.17
	ΔH (kJ mol−1)	—	—	—	—
	ΔS (J mol−1 K−1)	—	—	—	—
	q m (mg g−1)	112.61	51.55	19.67	162.07
	R 2	0.9945	0.9901	0.9998	0.9985
318 K	ΔG (kJ mol−1)	−16.76	−13.01	−17.37	−14.62
	ΔH (kJ mol−1)	—	—	—	—
	ΔS (J mol−1 K−1)	—	—	—	—
	q m (mg g−1)	137.17	53.36	36.95	163.93
	R 2	0.9997	0.9961	0.9995	0.9978

^aAt the investigated temperatures, the changes of Gibbs free energy (ΔG) for metal ions were all negative, indicating the adsorption process could proceed spontaneously under the experimental conditions. The absolute value of ΔG increased gradually with the rise of temperature, reflecting that the increase in temperature favors the adsorption process.

The enthalpy change (ΔH) of all metal ions was positive, revealing that the adsorption process was endothermic, which was in accordance with that the q_m increased with the rise of temperature. Meanwhile, the entropy changes (ΔS) were all positive, indicating that the adsorption process increased the degree of disorder, which may be due to the release of water molecules accompanied by the adsorption of metal ions and the rearrangement of the interface structure.⁶⁷

Different metal ions exhibited differentiated adsorption capacities in the single-ion system, and the order of the q_m was Cu²⁺ > Pb²⁺ > Cd²⁺ > Ni²⁺. Among them, Cu²⁺ showed a relatively high q_m, indicating that the material had a strong adsorption capacity for Cu²⁺ under ideal conditions. However, in the aqueous extract of CX with multiple ions coexisting, the adsorption efficiency of the composite for Cu²⁺ (78.49%) was obviously lower than that for Pb²⁺ (96.53%), indicating that the coexistence of multiple ions significantly altered their adsorption behaviors. This phenomenon might be due to the fact that Pb²⁺ could occupy the limited adsorption sites preferentially, thus producing an obvious inhibitory effect on the adsorption of Cu²⁺.68 In addition, Ni²⁺ had relatively high ΔH and ΔS, which might be because Ni²⁺ formed a stable hydration layer in the aqueous solution. Energy was required to break these hydration interactions during the adsorption process, which resulted in high ΔH. Moreover, the release of water molecules increased the disorder of the system, which led to high ΔS. Therefore, increasing the temperature was helpful to break the hydration interaction and promote the combination of Ni²⁺ with the composite material, thus enhancing the adsorption effect.⁶⁹

3.3. Adsorption mechanism

To obtain the surface charge of the heavy metal adsorption environment and to reveal the electrostatic interactions between the adsorbent and heavy metal ions, zeta potential measurements were performed on the CX extract containing MgAl-LDHs@FRBCs/Fe₃O₄ (MgAl-LDHs@FRBCs/Fe₃O₄ + CX extract), the CX extract containing MgAl-LDHs@FRBCs/Fe₃O₄ after heavy metal adsorption (MgAl-LDHs@FRBCs/Fe₃O₄-M + CX extract), and the blank CX extract. The results are presented in Fig. 8A. In the pH range of 2.3–5.3, both the extract containing MgAl-LDHs@FRBCs/Fe₃O₄ and the blank extract exhibited negative zeta potential values, indicating that the above solution presented a negative charge state. At the same pH, the absolute zeta potential value of the extract containing MgAl-LDHs@FRBCs/Fe₃O₄ was larger than that of the blank extract, suggesting that MgAl-LDHs@FRBCs/Fe₃O₄ provides a favorable negative charge environment for the adsorption of metal cations in the extract. After the adsorption of heavy metal, the absolute zeta potential value of the material decreased, indicating that the adsorbed metal cations bound to part of the negative charges on the composite, further demonstrating the existence of electrostatic interactions between the heavy metal ions and the material surface.

Fig. 8. (A) pH-dependent zeta potential of different systems: the CX extract containing MgAl-LDHS@FRBCs/Fe₃O₄ (MgAL-LDHs@FRBCs/Fe₃O₄ + CX extract), the CX extract containing MgAl-LDHs@FRBCs/Fe₃O₄ after heavy metal adsorption (MgAl-LDHs@FRBCs/Fe₃O₄-M + CX extract) and the blank CX extract (CX extract). (B) XPS survey spectrum of MgAl-LDHs@FRBCs/Fe₃O₄ after adsorption; (C) high-resolution XPS spectrum of Pb 4f; (D) high-resolution XPS spectra of O 1s before and after adsorption; and (E) relative area percentages of oxygen-containing components before and after adsorption.

As shown in Fig. 8B, a distinct Pb 4f peak was observed in the XPS survey spectrum of the composite after heavy metal adsorption, whereas no obvious characteristic peaks for Cd, Cu, or Ni were detected. However, the quantitative results have already demonstrated that the composite exhibited considerable adsorption capacity for Cd²⁺, Cu²⁺, and Ni²⁺. Therefore, the absence of these metal elements in the XPS spectra may be attributed to the amount of adsorbed Cu²⁺, Cd²⁺, and Ni²⁺ being below the detection limit of XPS, or to the low sensitivity of XPS for these ions. Further analysis was performed on the high-resolution XPS spectra of Pb 4f and O 1s to obtain the bonding information of Pb ions on the adsorbent surface. The Pb 4f spectrum, shown in Fig. 8C, can be deconvoluted into two spin–orbit doublets. The peaks at 138.50 eV and 143.40 eV are assigned to Pb 4f_7/2 and Pb 4f_5/2 of Pb–O, respectively; the peaks at 139.20 eV and 144.10 eV are assigned to Pb 4f_7/2 and Pb 4f_5/2 of Pb–O–C, respectively.^70,71 These results further confirm that Pb adsorption is associated with the formation of Pb–O bonds.

As shown in Fig. 8D, the high-resolution XPS spectra of O 1s before and after the adsorption of heavy metal ions on MgAl-LDHs@FRBCs/Fe₃O₄ composite were measured and deconvoluted, and Fig. 8E presents the corresponding peak area percentages of each component. The high-resolution O 1s spectrum of MgAl-LDHs@FRBCs/Fe₃O₄-M can be fitted into five peaks, corresponding to M–O (530 eV),⁷² M–OH (531.0 eV),⁷³ C–O/O–H (531.7 eV),³⁴ C O (532.6 eV), ⁷⁴ and O C–O (533.8 eV).³⁵ In terms of binding energy, slight decreases were observed for the C–O/O–H, C O, and O C–O components after the adsorption of heavy metal ions, which may be attributed to the increased electron cloud density around these oxygen-containing functional groups following the integration of metal ions.⁷⁵ Regarding the peak areas, compared with those before adsorption, M–O increased by 1.58%, M–OH increased by 9.75%, while C–O/O–H decreased by 9.71%, C O decreased by 1.07%, and O C–O decreased by 0.55%. The results further indicate the adsorption of metal ions on composite mainly through the chemical complexation with oxygen-containing functional groups (C–O/O–H, C O, and O C–O).⁷⁶ Based on the magnitude of the changes in peak areas, metal ions primarily undergo chemical complexation with C–O and O–H groups.

From the zeta potential and XPS characterization results, it can be concluded that the adsorption mechanism of heavy metal ions onto MgAl-LDHs@FRBCs/Fe₃O₄ mainly involves electrostatic adsorption and chemical complexation.

3.4. Distribution coefficient (K_d) and relative adsorption selectivity of heavy metals

The selectivity of MgAl-LDHs@FRBCs/Fe₃O₄ toward multiple metal ions in the CX extract was evaluated through competitive adsorption experiments. The distribution coefficient (K_d) and selectivity coefficient (α) for each metal ion were calculated using eqn (13) and (14), respectively.⁷⁷

13

14

where C₀ and C_e represent the initial and equilibrium concentrations of the metal ion (mg L⁻¹), respectively, V is the solution volume (L), and m is the mass of adsorbent (g). A larger K_d value indicates a stronger affinity of the adsorbent for the metal ion. When α_A/B > 1, it indicates that the adsorbent exhibits greater selectivity for metal ion A than for metal ion B.⁷⁸

As shown in Table 5, the K_d and α values for the four metal ions are presented. The calculated K_d values were 27.84 (Pb²⁺), 5.18 (Cu²⁺), 3.65 (Cd²⁺), and 0.55 (Ni²⁺), indicating that the adsorbent exhibits the strongest affinity for Pb²⁺, and therefore presents the most pronounced adsorption performance for this ion. Furthermore, using the K_d value of Ni²⁺ as the reference, the α values for Pb²⁺, Cu²⁺, Cd²⁺, and Ni²⁺ were calculated to be 50.63, 9.42, 6.63, and 1, respectively. These results demonstrate that the selectivity of the composite toward the four metal ions follows the order of Pb²⁺ > Cu²⁺ > Cd²⁺ > Ni²⁺. This difference in selectivity may be attributed to multiple factors, including the ionic radius, electronegativity of the heavy metals, and the structure of the composite.⁷⁹ Among these factors, the ionic radius order of Pb²⁺ (1.20 Å) > Cd²⁺ (0.97 Å) > Cu²⁺ (0.73 Å) > Ni²⁺ (0.69 Å) shows the closest correlation with the selectivity.⁸⁰ This suggests that ionic radius may play an important role in the adsorption of metal ions onto the composite surface. For example, the larger Pb²⁺ ion may simultaneously bind to multiple surface hydroxyl groups of the LDHs in a multidentate coordination manner, forming a stable chelate structure. Regarding electronegativity, the values for the four ions are Pb²⁺ (2.33) > Ni²⁺ (1.91) > Cu²⁺ (1.90) > Cd²⁺ (1.69),⁸¹ which differs considerably from the selectivity order. Nevertheless, the strongest adsorption capacity of the composite toward Pb²⁺ may also be related to its relatively high electronegativity, which enables stronger complexation ability with oxygen-containing groups within composite.⁸¹

Table 5. Distribution coefficients and selectivity coefficients of metal ions in the multi-metal system.

Metal ion	K d (L g−1)	α A/Ni
Pb2+	27.84	50.63
Cu2+	5.18	9.42
Cd2+	3.65	6.63
Ni2+	0.55	1

3.5. Effect of the adsorbent on the aqueous extract of Ligusticum chuanxiong Hort

To evaluate the effect of the adsorbent on the main active components of Ligusticum chuanxiong Hort., the determination of content was carried out. The calibration curves of ferulic acid, senkyunolide I and senkyunolide A were established, by plotting the peak area against the mass concentration (mg mL⁻¹) of the reference solution, and the relevant linear regression parameters were calculated. As shown in Fig. 9A–C and Table 6, the peak area of each component had a good linear relationship with the mass concentration within the investigated concentration range.

Fig. 9. (A–C) Linear calibration curves for ferulic acid, senkyunolide I, and senkyunolide A. (D–F) HPLC chromatograms of the control group, experimental group, and mixed standard solution. (G) Similarity evaluation of HPLC fingerprints of CX extract with (E1–E5) and without (C1–C5) the heavy metals removal. (H) XRD patterns of MgAl-LDHs@FRBCs/Fe₃O₄ before and after use under different conditions. (I) Reusability of the fabricated MgAl-LDHs@FRBCs/Fe₃O₄.

Table 6. Calibration curves and linear ranges of reference compounds.

Analyte	Regression equation	R 2	Linear range (mg mL−1)
Ferulic acid	y = 2.12 × 107 × +42 320.3	0.9999	0.02–0.69
Senkyunolide I	y = 1.05 × 107 × −209098.1	0.9983	0.03–1.07
Senkyunolide A	y = 3.41 × 106 × −9915.3	0.9991	0.017–0.56

As shown in Fig. 9D–F, under the described chromatographic conditions, the resolution met the analysis requirements, and no obvious interference peak was observed. In the chromatogram of the standard solution, clear characteristic peaks appeared at the corresponding retention times for all target components, with symmetrical peak shapes and good repeatability. In the chromatograms of the samples, the corresponding characteristic peaks could be detected at the same retention times as the standard solution, indicating that the representative components were present in the samples. Further quantitative analysis of the content of components was carried out (n = 6), and the results were shown in Table S3. Student's t-test was used to compare the contents of these components before and after adsorption treatment, and the results showed no statistically significant differences (P > 0.05) in the contents of ferulic acid (P = 0.51), senkyunolide I (P = 0.79) and senkyunolide A (P = 0.21). These results indicated that the adsorbent had no obvious effect on the contents of the main active components of the samples.

As shown in Fig. 9G, the effect of the adsorbent on the chemical components in the CX extract was evaluated by fingerprint similarity. The fingerprint similarity was calculated by the relative peak area normalization method, and the results showed that the average fingerprint similarity of the CX extract before and after the adsorption experiments was 99.8%. The results indicated that after the heavy metal adsorption treatment, the number of main chromatographic peaks and the relative peak area of the samples did not change significantly (n = 6), which further confirmed that the treatment with the composite material had no significant effect on the contents of the main active components.

3.6. Stability and reusability of adsorbent

The stability of the MgAl-LDHs@FRBCs/Fe₃O₄ composite was investigated by XRD analysis. The XRD patterns of the composite before adsorption, after adsorption and after elution with 1% HNO₃ are shown in Fig. 9H, respectively. Compared with that before adsorption, the composite after the removal of metal ions presented a similar XRD pattern, indicating that the adsorption process had no obvious effect on the crystal structure of the composite material, and the layered structure of MgAl-LDHs and the crystal phase of Fe₃O₄ remained stable. In addition, some diffraction peaks attributed to FRBCs were weakened to a certain extent, which might be due to the adsorption of heavy metal ions resulted in a reduction in the XRD peak intensity.⁸² After elution treatment with HNO₃, the XRD pattern of the composite was similar to that of the sample before adsorption, indicating that the adsorbed metal ions could be effectively removed by acid washing, and the main crystal structure of the material was not significantly damaged during the elution process, showing good structural stability. In summary, the XRD results indicated that the MgAl-LDHs@FRBCs/Fe₃O₄ composite material had good crystal structure stability during the adsorption and regeneration processes, and the adsorption process mainly occurred on the surface or in the interlayer structure of the material, which provided a structural basis for its practical application.

To evaluate the regeneration and reusability of the MgAl-LDHs@FRBCs/Fe₃O₄ composite material, continuous adsorption–desorption cycle experiments were carried out. After adsorption, 1 wt% HNO₃ was used for elution for the next cycle. As shown in Fig. 9I, the removal rates of the composite material for Pb²⁺, Cd²⁺, Ni²⁺ and Cu²⁺ in the first cycle were 96.53%, 83.82%, 35.48% and 78.49% respectively, and remained basically stable within five cycles. The corresponding removal rates in the fifth cycle were 92.60%, 81.28%, 32.62% and 76.89% respectively, with no obvious performance attenuation, indicating that the material had good cyclic utilization capacity. The ability to maintain cyclic performance in a complex competitive system further reflected the strong structural stability and the practical application value of the material.

3.7. Study on metal leaching rate

To further evaluate the structural stability of the composite material during the adsorption process, the leaching behavior of Mg²⁺, Al³⁺ and Fe³⁺ in the adsorption system was measured. The detection results showed that the leaching rates of Mg²⁺, Al³⁺ and Fe³⁺ were 0.00032%, 0.00033% and 0.0043%, respectively, all at an extremely low level, indicating that the MgAl-LDHs@FRBCs/Fe₃O₄ had good structural stability in the adsorption system. The leaching behavior had no significant effect on the system and did not cause potential secondary pollution, proving that the material had good application safety.

4. Discussion

The published adsorbents such as layered double hydroxides (LDHs),⁶⁸ biochar, ⁸³ and magnetic composites⁸⁴ have been applied to aqueous solutions, wastewater, or soil with simulated specific matrices, focusing on single or a few metal ions under relatively simple conditions. In contrast, this study prepared a MgAl-LDHs@FRBCs/Fe₃O₄ composite using Isatidis Folium and Isatidis Radix residues and applied it to the removal of multiple heavy metal ions from the aqueous extract of CX, which contains a complex mixture of bioactive compounds. This complex matrix can significantly influence adsorption behavior, which considerably differs from that in general aqueous solution systems.

The prepared composite exhibited satisfactory adsorption performance for heavy metal ions under both non-competitive conditions and competitive conditions. Mechanistic investigations by pH-dependent experiments, zeta potential measurements, and XPS indicated that chemisorption involving oxygen-containing functional groups (C–O/O–H, C O, and O C–O) played a major role, with the C–O/O–H bond contributing most significantly. Meanwhile, electrostatic interactions also facilitated the removal of metal ions, because the MgAl-LDHs@FRBCs/Fe₃O₄ provides a favorable negative charge environment for the adsorption of metal cations in the extract. Most importantly, HPLC analysis revealed no remarkable change in major chemical components of CX extract after adsorption. The negligible effect on active compounds may be attributed to the mild adsorption conditions and the specific surface chemistry of the composite, which preferentially targets metal ions over organic molecules under the experimental pH.

Overall, this work distinguishes itself from prior studies by demonstrating effective multi-metal removal in a real aqueous extract of TCM, while preserving the active components. The use of TCM residues as the carbon source further aligns with circular economy principles, offering a sustainable and cost-effective approach for heavy metal control in TCM preparations.

5. Conclusions

In this study, the MgAl-LDHs@FRBCs/Fe₃O₄ magnetic composite was successfully prepared using residues of Isatidis Folium and Isatidis Radix as the biochar carbon source. This composite exhibited a porous structure, high specific surface area, abundant functional groups, and good magnetic separability. It demonstrated efficient simultaneous removal of multiple heavy metal ions (Pb²⁺, Cu²⁺, Cd²⁺, Ni²⁺) from the aqueous extract of Ligusticum chuanxiong Hort., with fast adsorption kinetics, high capacity, and excellent reusability. The adsorption process fitted better with the pseudo-second-order (PSO) kinetic model and Langmuir isotherm, indicating that the adsorption was dominated by monolayer chemisorption. Meanwhile, a certain correlation with the pseudo-first-order (PFO) kinetic model was also observed, reflecting the participation of the physisorption. Thermodynamic analysis confirmed a spontaneous, endothermic process with increased entropy. Mechanistically, the removal was mainly governed by chemical complexation with oxygen-containing functional groups and electrostatic interactions. Importantly, the adsorption did not alter the major active components of the CX extract. This work not only provides a practical adsorbent for heavy metal decontamination in TCM aqueous preparations but also establishes a sustainable strategy for valorizing TCM residues, offering both theoretical support and application potential for quality control in TCM industrialization.

Author contributions

J. F.: methodology, data curation, writing – original draft. H. G.: methodology, data curation. C. X.: data curation. D. J.: data curation. S. W.: data curation. Q. H.: methodology, supervision, B. Y.: methodology, supervision. W. C.: methodology, supervision, writing – review and editing. H. H.: project administration, supervision, writing – review and editing.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

All data generated or analyzed during this study are included in the supplementary information (SI) accompanying this article. Supplementary information: high-resolution XPS spectra of Fe₃O₄, MgAl-LDHs, FRBCs; magnetic hysteresis curves of Fe₃O₄ and MgAl-LDHs@FRBCs/Fe₃O₄; the adsorption isotherms of Cd²⁺, Cu²⁺, Ni²⁺ and Pb²⁺ fitted lines with the Freundlich model, and fitted lines with the Temkin model; error function analysis of adsorption kinetic models and adsorption isotherm models for different metal ions; the contents of ferulic acid, senkyunolide A, and senkyunolide I before and after adsorption. See DOI: https://doi.org/10.1039/d6ra01513a.