Facile preparation of a novel Hg(II)-ion-imprinted polymer based on magnetic hybrids for rapid and highly selective removal of Hg(II) from aqueous solutions

Abstract

A novel magnetic ion-imprinted polymer (MIIP) with incorporated Fe₃O₄@SiO₂ nanoparticles was synthesized via a surface imprinting technique using allylthiourea (ATU) as a functional monomer. The MIIP has an excellent saturation magnetization of 15.9403 emu g⁻¹. Solution pH values had an immediate influence on the Hg(II) adsorption capacity with optimal removal occurring at pH 6. The adsorptivity of MIIP towards Hg(II) ions (78.3 mg g⁻¹) was approximately twice that of a magnetic non-ion-imprinted polymer (39.5 mg g⁻¹). The adsorption indicated that the adsorption mechanism closely agreed with a pseudo-second-order adsorption process, with a correlation coefficient (R²) of 0.998, which could be attributed to the soft acid–soft base interaction of Hg(II) and allylthiourea. The relative selectivity coefficients of MIIP for Hg(II)/Ni(II), Hg(II)/Cu(II), Hg(II)/Co(II) and Hg(II)/Cd(II) were 623, 355, 623 and 155, respectively. The recovery in rebinding capacity for Hg(II) was found to be decreased by only approximately 10.4% up to the 6th cycle, which suggests an excellent stability of MIIP. The application of MIIP in real samples shows a removal rate for Hg(II) of more than 99% to well below U.S. EPA mercury limits for wastewater, indicating that the MIIP has wide application prospects for Hg(II) removal in environmental water samples.

---

1. Introduction

Commonly, mercury contamination has been a focus of attention due to its persistent and highly toxic ability to bioaccumulate, producing progressive or irreversible nervous system injury, in acute cases, birth defects, disease and brain damage in humans and other species.^1,2 Due to its potential hazards, effective removal and recovery of mercury ions from aqueous solutions is significant for humans, the environment, and other resources. Consequently, strict limits have been mandated by the U.S. Environmental Protection Agency (EPA) and the World Health Organization such that an upper limit of 2 μg L⁻¹ is mandated for drinking water, and a mandatory release limit of 10 μg L⁻¹ for wastewater is mandated to protect ecosystems.^3,4

Among various processes that have been developed, adsorption is promising due to its design flexibility, simplicity, relative cost-effectiveness and effectiveness to purify water.^5–8 However, conventional adsorbents, which are not eco-friendly, and would make pollutants transfer, leaving potentially hazardous residues, hold disadvantages of low capacities, weak binding affinities and poor selectivities for mercury.^9–11

Recently, ion-imprinting polymer (IIP), which provides a versatile and powerful method of remarkable recognition for template ions in three-dimensional cross-linked polymers, have been extensively developed with regard to their adsorption properties.^12,13 This imprinting technique has been widely explored as a new type of adsorbent for the treatment of Cu(II),¹⁴ Li(I),¹⁵ Pb(II),¹⁶ Au(III)¹⁷ from waste water. However, traditional IIP, which was prepared via the entrapment technique, suffers several disadvantages, such as weak binding affinities and the embedding of recognition sites, which further hinder the application of IIP.¹⁸

The surface ion imprinted polymers, which can effectively overcome those drawbacks, provide accessible recognition sites that are created on a material's surface, where complex ligands are arranged to match the size, charge, coordination number of the target ion. The silicon-based mesoporous polymers can be effectively imprinted on the surfaces of SiO₂ nanoparticles, and can increasingly improves the stability and the high graft ratio to immobilize specific ligands to fabricate a variety of functional shells due to the presence of additional hydroxyl groups, which can react with organosilane-coupling agents.^19,20 However, the surface ion imprinted polymers can't be separated from wastewater in a convenient and economical way.

For more convenient separation and recycling of adsorbents, if a magnetic surface ion imprinting polymer, which is combined with magnetic properties, can be prepared. Magnetic nanoparticles Fe₃O₄ possess unique magnetic response, high dispersability, low toxicity, relatively large surface area and easy surface modification.^21,22 The superparamagnetic microspheres of Fe₃O₄@SiO₂@IIP (MIIP) as a magnetic surface ion imprinting polymer by encapsulating a magnetic nanoparticles core Fe₃O₄ and a mesoporous SiO₂ shell in organic polymers, achieve a remarkable combination of the surface imprinting technique and magnetic properties of Fe₃O₄, avoiding the complicated process of filling column and problems of pressure drop. MIIP nanoparticles could possess the advantages of the surface ion imprinted polymers, such as large surface areas, excellent metal ion transfer kinetics, high metal adsorption capacity and high selectivity, meanwhile, it also has several potential advantages of favourable physical properties, preeminent thermal and chemical stabilities, easy separation and recycling of the adsorbents from liquid systems by magnetically.^23,24

For Hg(II) adsorption, among recent studies, M. Monier et al.²⁵ used thiourea modified cellulosic cotton fibers achieved a good adsorption capacity; however, this processes still exist slow adsorption kinetics, making the process tedious and time-consuming. Xu et al.²⁶ proposed bifunctional ligands for thymine–Hg(II)–thymine interactions to enrich mercury, where it suffers from low adsorption capacity and relatively low selectivity for Hg(II). To solve these problems, considering the rapid combination capability and high selectivity, a thiol/thio-functionalized monomer is typically preferred over other monomers for use as remarkable recognition of binding sites based on the soft acid–soft base interaction of Hg(II) and allylthiourea. An additional double bond is essential for polymerization; therefore, allylthiourea is a suitable functional monomer to be used to avoid additional grafts of double bonds to make the material effective at selectively removing mercury.

In this study, a novel magnetic ion-imprinted polymer (MIIP), which is a composite structure that integrates magnetic separation and imprinting technique, was prepared using allylthiourea as an effective functional monomer towards Hg(II) ions based on the surface imprinting technique. To demonstrate this material's performance, the resulting MIIP was compared with magnetic non-ion-imprinted polymer (MNIP) in terms of adsorption capacity, dynamic adsorption, selectivity and regeneration tests; the parameters of the removal process were also optimized.

2. Experimentation

2.1. Materials and reagents

Ferric chloride hexahydrate (FeCl₃·6H₂O) and ferrous sulphate (FeSO₄·7H₂O) were obtained from Yufeng Chemical Reagents Company (Changsha, China). Tetraethyl orthosilicate (TEOS) was supplied by Pure Crystal Shanghai Reagent Co., Ltd. (Shanghai, China). Polyethylene glycol was received from J&K Scientific. Allylthiourea (ATU), ethyleneglycol dimethacrylate (EGDMA), and azodiisobutyronitrile (AIBN) and 3-methacryloxypropyltrimethoxysilane (KH570) were purchased from Aladdin Reagent Co. Ltd. (Shanghai, China). Mercury(II) chloride was obtained from Tianjin Chemical Reagent Company, China. All reagents were of analytical grade and used without further purification.

2.2. Materials and instruments

Fourier transmission infrared spectra (FTIR) in KBr were recorded using a VERTEX70 FTIR apparatus (Germany) at 4000–400 cm⁻¹. Scanning electron microscopy (SEM, Japan, Shimadzu) was employed to assess the morphologies. The Brunauer–Emmett–Teller (BET) methods were used to calculate the specific surface area of materials based on N₂ adsorption isotherms (Micromeritics, America). A ConteAA 700 (Analytik Jena, Germany) flame atom absorption (FAAS) spectrophotometer and an atomic fluorescence spectrometer with a limit of detection (LOD) of 0.005 μg L⁻¹ (AFS, AFS-2202E, Beijing Haiguang, China) were used to determine the concentration of metal ions. Thermogravimetric analysis (TGA) was performed with a DSC/DTA-TG device (STA 449C Jupiter Netzsch, Germany).

2.3. Preparation of Hg–MIIP

2.3.1. Preparation and modification of Fe₃O₄. Fe₃O₄ magnetic nanoparticles were prepared via co-precipitation with modifications.²⁷ FeCl₃·6H₂O (13.5 g) and polyethyleneglycol (10 g) were first dissolved in 200 mL of deionized water. After purging the solution with high-purity N₂ gas for at least 30 min, FeSO₄·7H₂O (8 g) in a molar ratio of Fe³⁺/Fe²⁺ of 2 : 1 was added. The resulting precursor solution was vigorously stirred under an N₂ atmosphere for 20 min at room temperature, where the pH was adjusted to 10 using the addition of 40 mL NH₃·H₂O solution (25%, w/v) dropwise at 60 °C. The suspension was maintained at 90 °C for 30 min. Then the black iron oxide nanoparticles were obtained and dried under vacuum at 60 °C.

2.3.2. Preparation and modification of Fe₃O₄@SiO₂. Fe₃O₄@SiO₂ magnetic nanoparticles were created using the procedure described by H. He et al. with minor modifications.²⁸ For the subsequent coating with silica dioxide, the as-prepared Fe₃O₄ magnetic nanoparticles (5.0 g) were homogeneously resuspended in 220 mL 90% (v/v) of the isopropanol–deionized water mixed solvent. Then, 20 mL of NH₃·H₂O solution was promptly added to this suspension, and a predetermined amount of 30 mL TEOS in large excess was added dropwise under continuous stirring for 12 h. Then, the Fe₃O₄@SiO₂ nanoparticles was obtained after vacuum drying at 50 °C.

2.3.3. Preparation of vinyl-functionalized Fe₃O₄@SiO₂. For better polymerization, C=C groups were grafted onto the coating silica surface via the reaction of KH570 with surface silicon hydride bonds of Fe₃O₄@SiO₂.²⁹ The obtained silica-coated sample (0.8 g) was fully dispersed in 100 mL of methyl and then deoxygenated with nitrogen gas before the immediate addition of 2 mL of ammonia solution. Then, 10 mL of KH570 was added dropwise at 65 °C, and the solution was then left overnight under rotatory stirring. The magnetic vinyl-functionalized Fe₃O₄@SiO₂ was then dried under vacuum at 50 °C overnight.

2.3.4. Synthesis of Fe₃O₄@SiO₂@IIP. Magnetic ion-imprinted polymers were synthesized using ATU as a functional monomer. HgCl₂ (1 mmol) and ATU (2 mmol) were dissolved in 60 mL of methyl; the mixture was then stirred vigorously under nitrogen for prepolymerization at 35 °C over a period of 4 h. A mixture of 2 g of vinyl-functionalized Fe₃O₄@SiO₂ nanoparticles that were previously suspended in 20 mL of methyl, and approximately 0.95 mL of EGDMA that were dissolved in 20 mL of DMF was prepared. Then, the mixed solution and AIBN (50.0 mg) were added into the pre-polymerization solution in sequence, purged with N₂ for 15 min and then sealed. The polymerization reaction thus proceeded at 65 °C for an additional 12 h. For comparison, the corresponding non-imprinted polymer was synthesized analogously without the addition of Hg(II) ions.

Leached MIIP nanoparticles were finally acquired by leaching the retained Hg(II) ions out from the polymeric network using 5% thiourea in 0.5 mol L⁻¹ HNO₃. The adsorbent was then washed with distilled water, soaked with ammonia–deionized water mixture 1 : 9 (v/v) overnight, and dried under 60 °C. The scheme for preparing Hg(II)-ion-imprinted polymers is shown in Fig. 1.

Fig. 1 (a) The pre-complex of Hg(II)–ATU; (b) schematic procedure for the preparation of Hg(II)–MIIP via surface imprinting technique.

2.4. Batch adsorption experiments

2.4.1. Adsorbent dose studies. MIIP with a dry mass of 5, 15, 20, 25 and 30.0 mg was dispersed in 20 mL of dilute aqueous solutions containing 60 mg L⁻¹ Hg(II) ions and was shaken for 3 h at 25 °C, respectively. MIIP was then collected magnetically for determining the residual Hg(II) concentration by atomic fluorescence spectroscopy (AFS).

2.4.2. Static adsorption. To evaluate the effectiveness of MIIP at removing Hg(II) from aqueous solutions, adsorption isotherm experiments were performed. MIIP (MNIP) with a dry mass of 20.0 mg was dispersed in 20 mL of dilute aqueous solutions containing Hg(II) ion ranging from 20 to 300 mg L⁻¹ at optimum pH values, respectively. The solution was shaken for 6 h at 25 °C to ensure sufficient adsorption of the Hg(II) ions. MIIP (MNIP) were then collected magnetically for determining the residual Hg(II) concentration by atomic fluorescence spectroscopy (AFS). The adsorption capacity of Hg(II) can be calculated based on the following equation:³⁰where Q_e (mg g⁻¹) is the adsorption capacity of Hg(II); C₀ and C_e (mg L⁻¹) represent the initial and the final equilibrium concentrations of Hg(II), respectively; V (L) is the volume of solution; and m (g) is the dry mass of the MIIP or MNIP nanoparticles.

2.4.3. Dynamic adsorption. For the dynamic adsorption experiment, MIIP was examined by investigating the mercury kinetics. An appropriate quantity of the adsorbent (300 mg) of MIIP (MNIP) was added to 300 mL of Hg(II) solution at a concentration of 200 mg L⁻¹. The mixture solution was sampled at designated time intervals while undergoing continuous stirring; the concentration of Hg(II) remaining in the samples was determined by AFS. Also, binding kinetics were analysed by monitoring the amounts of Hg(II) ions in the solutions over time.

2.4.4. Effect of pH. To investigate the effect of pH on Hg(II) adsorption, 0.1 M HNO₃ and 0.1 M NaOH were used to adjust the values of the system pH. In this study, the effect of pH on Hg(II) uptake was evaluated in batch experiments by adding 20 mg of MIIP (MNIP) to 20 mL of the solutions containing 30 mg L⁻¹ of Hg(II) within the pH values of 2 to 7.

2.4.5. Adsorption–desorption studies. For the adsorbent, its regeneration is likely an important factor in achieving sustainability and is an economic necessity; therefore, the reusability and stability of MIIP was repeatedly performed by desorption–regeneration cycles. The influence of different concentrations in the elute solution on desorption was investigated. Hg(II)–MIIP was regenerated by treatment with 0.2 and 0.5 mol L⁻¹ HNO₃ and 5% thiourea in 0.5 mol L⁻¹ HNO₃, respectively. The concentration of residual Hg(II) ions in samples was determined via FAAS. The regenerated MIIP was soaking with ammonia–deionized water mixture overnight, and was reused to adsorb Hg(II) in subsequent desorption and regeneration runs.

2.4.6. Selectivity study. The selective recognition of MIIP for Hg(II) ions was implemented by competitive adsorption studies. The interference ions were used in mixtures with similar ionic radii and same charge for comparison, such as Ni(II), Cu(II), Co(II) and Cd(II). 50 mg of MIIP (MNIP) was added into 50 mL mixture solution with co-existing ions. The concentrations of metal ions were measured by FAAS after shaking for 2 h. Meanwhile, the selective effectiveness of MIIP for Hg(II) was also studied in the presence of 30 mg L⁻¹ hexadecylpyridinium bromide hydrate (CPB) as surfactant and Ni(II), Cu(II), Co(II), Cd(II) as co-existent ions with an initial concentration of 30 mg L⁻¹.

2.4.7. Application to real samples. To validate the applicability of MIIP for the selective removal of Hg(II) in samples, river water samples (Ganjiang River, Nanchang, China), lake water samples (Tianyi lake, Nanchang, China), wastewater 1 (mainly containing Fe ion) and wastewater 2 (mainly containing Li ion) were selected and analysed. In addition, trace Ni(II), Cu(II) and Cd(II) are also detected in the wastewater by FAAS. 25 and 50 mg L⁻¹ Hg(II) were added in above samples, respectively, due to trace Hg(II) in the solution. 50 mL solution was then treated with 50 mg of MIIP for 3 h at pH of 6.0 ± 0.2.

3. Results and discussion

3.1. Characteristics

3.1.1. FT-IR results. To further characterize the surface microstructure of MIIP, FT-IR spectra was determined and is shown in Fig. 2. A characteristic peak at 463 cm⁻¹ that corresponds to the stretching of Fe–O bonds is observed in all spectra. The strong bands at 1097 cm⁻¹ for the stretching vibration of Si–OH is due to the presence of intermediate amorphous silica shells.³¹

Fig. 2 FT-IR spectra of MIIP, MNIP and Fe₃O₄@SiO₂ nanoparticles.

For the MIIP, the absence of C–S and the thiol (S–H) characteristic peaks at 2300 and 1200 cm⁻¹, respectively, in addition to the peak at 1464 cm⁻¹ are related to the stretching vibration of C=S bonds in the thioamide groups (NH–C=S), which implies that active thiourea units exist principally in the form of thion. The additional peak at 1726 and 1387 cm⁻¹ correspond to the characteristic vibration frequencies of carbonyl functional groups (i.e., –COO) and the covalent bond between Fe₃O₄@SiO₂ and floating –COOH, which indicates the formation of imprinting polymerization layer. In addition, similar spectral profiles are observed in the MNIP.

3.1.2. SEM and TEM analysis. A preliminary observation of the surface morphology and topography of the synthesized nanoparticles was performed using SEM and TEM, as shown in Fig. 3. All samples were found to be spherical. Fe₃O₄ nanoparticles (Fig. 3(A)) can be characterized by a smooth surface, and more rough surfaces of MIIP (Fig. 3(C)) can attributed to the presence of imprinted cavities on the surface of Fe₃O₄@SiO₂.

Fig. 3 SEM/TEM images of (A/D) Fe₃O₄, (B/E) Fe₃O₄@SiO₂ and (C/F) MIIP nanoparticles.

To further characterize the diameter and dispersity of particles, TEM was carried out. The Fe₃O₄ presents an excellent property of uniform size (9–13 nm) and good dispersion in (Fig. 3(D)). An average diameter of Fe₃O₄@SiO₂ particles (Fig. 3(E)) was observed of about 54.5 nm, without obvious reunion, due to weak aggregation of Fe₃O₄ nanoparticles. The composite MIIP nanoparticles (Fig. 3(F)) exist a certain extent of polymer chain which changes the state of dispersion of the composite nanoparticles.

3.1.3. Magnetic properties. The magnetic properties of nanoparticles has been measured and the special saturation magnetization loops are shown in Fig. 4. The saturation magnetization of Fe₃O₄, Fe₃O₄@SiO₂ and MIIP nanoparticles at an external field of 12 kOe is 50.0251 emu g⁻¹, 22.8814 emu g⁻¹, 15.9403 emu g⁻¹, respectively, which generally represent the weakening trend. This result shows that the coated SiO₂ layer and the functionalization of organic layer on the surface of Fe₃O₄@SiO₂ can slightly affect the magnetism. However, the magnetization of MIIP is enough for regeneration under magnetic field, and the process of separation of MIIP is also shown in Fig. 4.

Fig. 4 VSM curves of (A) Fe₃O₄, (B) Fe₃O₄@SiO₂ and (C) MIIP nanoparticles; and the below inset shows the magnetic separation of MIIP.

3.1.4. Thermogravimetric analyses. The thermogravimetric analysis (TGA) curves were determined under nitrogen gas ranging from the room temperature to 800 °C (10 °C min⁻¹) and shown in Fig. 5.

Fig. 5 TGA curves of (A) Fe₃O₄@SiO₂, (B) MIIP, (C) MNIP, (D) Hg–MNIP and (E) Hg–MIIP nanoparticles.

The pure matrix material Fe₃O₄@SiO₂ (A) was found to exhibit a marginal weight loss of about 5% up to 800 °C. For the leached MIIP (B), the TGA curves are composed of a two-stage degradation mechanism. In the first stage, weight loss occurred from 100 to 300 °C due to the desorption of adsorbed water. The second stage was present between 300 and 450 °C, where the weight decreased significantly due to the thermal decomposition of the imprinted polymer. The TGA curves of Hg–MIIP (D) revealed a rapid and three-stage degradation mechanism. Different from the MIIP, the degradation stage occurred at lower temperatures (i.e., 150 to 300 °C) due to the decomposition of the oligomer, in addition to the residual volatile components of free small organic solvent molecules. The third degradation stage was followed by a large decline of 4.6%, which resulted in the thermal degradation of mercury and its strong complex with sulphur-generated groups in the polymers. Similar TGA profiles were also detected for Hg–MNIP (E).

In short, the leached MIIP was stable up to 300 °C with 36.6% by weight, where the polymers are approximately 31.6% and Fe₃O₄@SiO₂ nanoparticles are of about 63.4% in addition to 5% adsorbed water. For comparison, MNIP (C) exhibited a similar curves and weight loss. The weight loss of Hg–MIIP and Hg–MNIP were found to be more than approximately 5.65% and 3.93% larger compared with to the MIIP and MNIP, respectively. Based on these results, the calculated adsorption capacity of MIIP and MNIP for Hg(II) were found to be 69.7 and 45.1 mg g⁻¹, which is correspond with the adsorption isotherm. Also, the Hg(II)/ATU linker ratio in the MIIP can be figured out to be approximately 1 : 2.

3.1.5. BET. The porosity of the MIIP (MNIP) was determined using the BET method. Based on this analysis, a specific BET surface area of MIIP (97.32 m² g⁻¹) was calculated, which was significantly higher than that of MNIP (57.95 m² g⁻¹). All profiles belonged to a typical type IV isotherm with a distinct hysteresis H2-type loop. A larger cumulative pore volume of MIIP was probably caused by the mesoporosity on their surface, this phenomenon also particularly provided evidence for a large amounts of imprinting cavity sites.

3.2. Adsorption experiments

3.2.1. Adsorbent dose studies. To study the efficient removal rate for Hg(II) ions per mass, a proper adsorbent dose was taken after, as shown in Fig. 6. The results show that 20 mg MIIP is optimum. Thus we conducted other studies by adding 20 mg MIIP in solution.

Fig. 6 Adsorption of various mass of MIIP towards Hg(II) ions at a pH of 6.

3.2.2. Influence of pH. The effect of pH on the efficiency of MIIP for the Hg(II) adsorption process was investigated over pH values ranging from 2 to 7. As shown in Fig. 7, the binding capacity increases with increasing pH from 2 to 6 but declines marginally from 6 to 7. The results revealed that a low adsorption capacity was maintained at low pH values between 2 and 3; this is primarily caused by more positively charged surface of MIIP due to the combination of sulphur in the polymer matrix with large amounts of H⁺ competing with Hg(II) for the imprinted sites at low pH. At higher pH levels (i.e., 7), the formation of speciation, such as Hg(OH)₂ in aqueous solution, cannot match the imprinted cavities and functional groups in the polymer matrix via electrostatic interaction due to variations in size, shape; which thus lead to a decline of the adsorption capacity. We hence conducted the studies at a pH range of 6.

Fig. 7 Effects of pH on the adsorption capacity of MIIP and MNIP for Hg(II) ions.

3.2.3. Adsorption isotherm. The adsorption isotherms of MIIP and MNIP for Hg(II) removal from water samples are shown in Fig. 8(A), which clearly shows that the adsorption capacity of MIIP towards Hg(II) increased significantly with an increasing initial concentration of Hg(II) ions in the low concentration range; this increase in adsorption capacity then gradually slowed and eventually saturated (78.3 mg g⁻¹), which was significantly higher than that of MNIP by approximately 2 fold. The higher effectiveness of MIIP to form complexes with Hg(II) ions due to the imprinting process was likely caused by an energy benefit related to the formation of non-distorted complexes compared to the MNIP nanoparticles. The adsorption capacity of MIIP towards Hg(II) ions was higher than that of previous reports, as shown in Table 1.

Fig. 8 (A) Adsorption isotherms for Hg(II) ions on MIIP and MNIP at a pH of 6.0 ± 0.2 at 25 ± 0.2 °C; (B) Langmuir isotherms and (C) Freundlich isotherms for Hg(II) for Hg(II) adsorption on MIIP and MNIP.

Table 1 Comparison of the maximum adsorption capacities for Hg(II) ions of different adsorbents

Sorbent	Monomer or ligand	Preparation technique	Qmax (mg g^−1)	Equilibrium time (min)	Ref.
MIIP	Allylthiourea	Surface imprinting	78.3	5	This study
Hg-imprinted copolymer	N-Methacryloyl-2-mercaptoethylamine	Radical copolymerization	28	50	32
Hg–IIPs	3-Aminopropyltriethoxysilane	Sol–gel process	4.5	80	33
Hg–IIPs–T	3-Isocyanatopropyltriethoxysilane	Sol–gel process	5.3	60	34
P(TAR–Hg)	Methacrylic acid	Thermally copolymerizing	0.025	10	35

---

It is important to demonstrate the Hg(II) ions distribution between the adsorbent phases and the liquid; therefore, Langmuir and Freundlich isotherm models were used to fit the adsorption process in this study. These models were defined by the following equations:³⁶

where Q_m (mg g⁻¹) is the maximum adsorption capacity at equilibrium; C_e (mg L⁻¹) and Q_e (mg g⁻¹) represent the equilibrium concentration and adsorption capacity toward Hg(II) in solution at adsorption equilibrium, respectively; K_L is the Langmuir constant, which is related to the affinity between the adsorbent and solution (L mg⁻¹); K_F (L mg⁻¹) and 1/n are the Freundlich constants for the adsorption capacity and adsorption intensity, respectively.

The fitting results are shown in Fig. 8(B) and (C), and the isotherm parameters are summarized in Table 2. It is significant that a plot of C_e/Q_e vs. C_e was found to be nearly linear, which indicates that the adsorption fits the Langmuir model with higher correlation coefficients of 0.998 and 0.997; this results indicates that the active adsorption sites of the adsorbent towards Hg(II) ions on MIIP are effectively homogeneous, leading to a monolayer binding. The separation factors of K_L and 1/n are in the range of 0 to 1, which suggests that the adsorption process is feasible and favourable. The calculated maximum adsorption capacity of MIIP (78.3 mg g⁻¹) and MNIP (39.5 mg g⁻¹) are consistent with the experiment dates.

Table 2 Freundlich, Langmuir and D–R models constants of Hg(II) adsorption on MIIP and MNIP

Sorbent	Langmuir isotherm parameters				Freundlich isotherm parameters			Duninin–Radushkevich parameters
	Qm.cal (mg g^−1)	Qm.exp (mg g^−1)	KL (L·mg^−1)	R^2	n (mg^(1−(1/n))·L^(1/n)·g^−1)	KF (mg^(1−(1/n))·L^(1/n)·g^−1)	R^2	KDR ((mol J^−1)^2)	E (J mol^−1)	R^2
MIIP	79.4	78.3	0.268	0.998	6.974	38.51	0.928	0.005	10.2	0.996
MNIP	43.6	39.5	0.070	0.997	3.658	10.05	0.879	0.006	8.9	0.995

---

To further distinguish the physisorption of metal ions from chemisorption, the Dubinin–Radushkevich (D–R) isotherm model, which assumes a non-homogenous surface on the adsorbent and a complete adsorption process with the adsorbate, was used to analyse the adsorption energy and can be expressed as follows:³⁷

ln Q_e = ln Q_m − K_DRε² (4)

ε² = RT ln(1 + (1/C_e)) (5)

E = (2K_DR)^−1/2 (6)

where K_DR ((mol J⁻¹)²) represents the activity coefficient related to the free energy of adsorption; ε is the Polanyi potential; R is the ideal gas constant (8.314 J mol⁻¹ K⁻¹); T (K) is the absolute temperature; C_e is the equilibrium concentration; and E (J mol⁻¹) is the average adsorption energy.

The D–R equation is one of the best features based on the material's temperature dependence and its balance between the adsorption energy and the adsorption process. Physisorption is the principal process that occurs when the value of E is less than 8 kJ mol⁻¹; otherwise, chemisorption is the dominant process. Fitting curves for these processes are shown in Fig. 9. Similarly, the large correlation coefficients R² (>0.995) show that the D–R model describes the adsorption accurately. For the mean adsorption energy (E) of MIIP and MNIP, as shown as Table 2, the parameters are larger than 8 kJ mol⁻¹, which indicated that the adsorption of Hg(II) onto MIIP primarily proceeds by binding the surface functional groups of the polymers; thus, chemisorption is the predominant adsorption mechanism.

Fig. 9 D–R model for Hg(II) adsorption on MIIP and MNIP.

3.2.4. Adsorption kinetics. To identify the adsorption mechanism of the kinetics, the dynamic binding curves as functions of time are used to evaluate the ionic diffusion on MIIP (MNIP), as shown in Fig. 10(A). Extremely fast kinetics are observed with MIIP, which can achieve 91.4% of total capacity within the initial 3 min in the reaction. This high adsorption rate can be attributed to the material's sufficient number of adsorption sites, which is more favourable for the strong chelation between ATU and Hg(II) ions. The adsorption gradually approached equilibrium approximately 5 min. While the adsorption rate of MNIP with Hg(II) is considerably slower than that of MIIP, indicating a slower and more constrained diffusion process for Hg(II) adsorption on MNIP.

Fig. 10 (A) Adsorption kinetics for Hg(II) ions on MIIP and MNIP at a pH of 6.0 ± 0.2; and (B) pseudo-second-order kinetics model for Hg(II) adsorption on MIIP and MNIP, respectively.

Considering that a high reliability represents the adsorption's dynamic processes, the experimental data was used to fit the pseudo-second-order kinetic model by assuming a proportional relationship between the adsorption capacity of the adsorbent and the amounts of sites on its surface, which can be defined based on the following equations:³⁸

h₀ = K₂Q_e² (8)

where Q_t and Q_e (mg g⁻¹) and are the adsorption capacities at time t and at equilibrium, respectively; K₂ (g mg⁻¹ min⁻¹) is the constant rate of the pseudo-second-order equation; and h₀ (mg min⁻¹ g⁻¹) is the initial adsorption rate.

The fitting curves of the models are shown in Fig. 10(B), and the related kinetics parameters and regression values are listed in Table 3. The value of the initial adsorption rate h₀ was calculated to be 17 mg g⁻¹ min⁻¹ of MIIP, which was significantly higher than that of MNIP (5 mg g⁻¹ min⁻¹). Such fast kinetics for MIIP can be attributed to their imprinted pore size, which can facilitate the diffusion of Hg(II) ions. Additionally, the estimated theoretical adsorption parameter Q_e (82.7 mg g⁻¹) with an extremely high correlation coefficient (R² 0.998) is significantly closer to the experimental data (80.3 mg g⁻¹). Thus, the results demonstrate that the pseudo-second-order mechanism is predominant, and the rate-limiting factor is the chemical adsorption with regard to the chemical binding sites on the surface of the MIIP during Hg(II) absorption.

Table 3 Pseudo-second-order kinetic constants of Hg(II) adsorption on MIIP and MNIP

Adsorbent	Pseudo-second-order kinetics
	K2 (min^−1)	Qe (mg g^−1)	h0 (mg g^−1 min^−1)	R^2
MIIP	0.025	82.7	173	0.998
MNIP	0.034	38.5	51	0.997

---

3.2.5. Adsorption selectivity. The potential for the selective separation of Hg(II) ions from an aqueous solution was then evaluated by performing selectivity binding studies. As shown in Fig. 11, it can be seen that MIIP exhibits a much higher selective adsorption for Hg(II) ions than that of other ions in the mixture solutions of co-existing ions. And the Hg(II) adsorption capacity and the selectivity did not reduce in the presence of CPB. Moreover, the excellent selectivity towards Hg(II) ions is more superior than the previous report by F. X. Luo.³⁹

Fig. 11 Selectivity of MIIP and MNIP for Hg(II) ions.

The selectivity parameters of the distribution coefficient (K_d), selectivity coefficient (K) and relative selectively coefficients (K′) are also calculated.

One measure of an adsorbent's affinity for target metal ions is the distribution coefficient (K_d), which is defined as:

The selectivity coefficient (K) can be calculated as:

To measure the imprinting effect on the selectivity of the resulting MIIP, the relative selectivity coefficients (K′) are used, which can be expressed as follows:

where K_d is the distribution ratio of Hg(II) ions between the MIIP/MNIP nanoparticles and the aqueous solution; K^Hg_d and K^M_d are the distribution ratios of Hg(II) and competitive ions, respectively; K_Hg/M are the selectivity coefficients for Hg(II) ions relative to other competitive ions; K′ is the relative selectivity coefficient; and K^MIIP_Hg/M and K^MNIP_Hg/M are the selectivity coefficients of the MIIP and MNIP, respectively.

A comparison of the selectivity parameters of MIIP and MNIP is summarized in Table 4. The selectivity coefficients (K_Hg/M) obtained for MIIP reached 355, and the distribution coefficient (K_d) of MIIP for Hg(II) ions is significantly larger than other ions, which proves a higher affinity of restricted-access. With regard to size, shape, and coordination geometries, Hg(II) ions can match well with the amounts of specific binding sites in the polymer layer on the matrix, which were complementary to Hg(II) ions. Also, the stability constant of the coordination compounds between Hg(II) ion and the ligand of ATU is significantly higher than that for other ions: more stable constants tend to yield more stable coordination compounds. In addition, the protonated sulphur of the MIIP matrix simultaneously weakened its interactions with other interfering ions, which contributes to the selectivity effectiveness.

Table 4 Selectivity parameters of MIIP and MNIP for Hg(II) ions

Metal ions	MIIP			MNIP
	Kd	KHg/M	K′	Kd	KHg/M
Hg(II)	1667			743
Cd(II)	11	155	7	33	22
Co(II)	3	623	30	36	20
Cu(II)	5	355	29	60	12
Ni(II)	3	623	22	26	29

---

3.2.6. Adsorption mechanism. D–R isotherm model with parameters of adsorption energy (E) (10.2 kJ mol⁻¹) indicated Hg(II) adsorption on MIIP is chemisorption. This verified also the soft acid–soft base interaction of Hg(II) and allylthiourea.

In order to further analyze the interaction ratio between sulphur and mercury ions. TGA curve was carried out, and the results show that the Hg(II)/ATU linker ratio in MIIP can be figured out to be approximately 1 : 2, according to different degradation values of MIIP and Hg(II)–MIIP. Meanwhile, EDS analysis was also performed, as shown in Fig. 12. The EDS spectra of Hg(II)–MIIP proved the existence of sulphur and mercury in MIIP, and the element percent of sulphur and mercury was 0.45 and 0.21, respectively, indicating a bonding ratio of 2 : 1 during the process of adsorption for Hg(II) ions on MIIP, which is consist with the calculated ratio in TGA.

Fig. 12 Element analysis of Hg(II)–MIIP after adsorption.

3.2.7. Regeneration. To evaluate the elution efficiency based on the concentration of different elution solutions, Hg(II)–MIIP was treated with 0.2 and 0.5 mol L⁻¹ HNO₃, which resulted in desorption percentages of 86.5% and 99.1%, respectively. However, 5% thiourea in 0.5 mol L⁻¹ HNO₃ could desorb pre-adsorbed Hg(II) more easily compared to pure 0.5 mol L⁻¹ HNO₃; thus, 5% thiourea in 0.5 mol L⁻¹ HNO₃ was used as the elute solution for Hg(II)–MIIP.

Each regeneration cycle is shown in Fig. 13. MIIP can be effectively regenerated up to the 6th desorption–regeneration cycle, and its rebinding capacity for Hg(II) decreased by only approximately 10.4%, showing that the structural of MIIP is stable and its regeneration ability is excellent.

Fig. 13 Desorption–regeneration cycles of Hg(II)–MIIP at a pH of 6.0 ± 0.2.

3.2.8. Application to real samples. The proposed MIIP was applied to validate their potential practicability; the results of this analysis are shown in Table 5. Among the samples tested, a removal rate of more than 99% has effectively reduced the Hg(II) concentration to well below U.S. EPA mercury limits for wastewater, which further confirms the validity of MIIP for the removal of Hg(II) in environmental water samples, especially in wastewater samples.

Table 5 Removal rate of Hg(II) in natural water and wastewater samples

Sample	Hg(II) added (μg L^−1)	Hg(II) found (μg L^−1)	Residual Hg(II) (μg L^−1)	Removal rate (%)
Tianyi Lake	—	0.00
	25	25.03	0.19	99.23
	50	49.94	0.33	99.35
Ganjiang River	—	0.15
	25	25.61	0.16	99.36
	50	50.23	0.41	99.18
Wastewater 1	—	0.94
	25	25.87	0.22	99.14
	50	51.13	0.51	99.01
Wastewater 2	—	2.43
	25	28.35	0.28	99.02
	50	54.72	0.53	99.04

---

4. Conclusions

In this study, MIIP was prepared by enveloping magnetic nuclei in the presence of intermediate amorphous silica shells using the surface imprinting technique, which produced improved kinetics, favourable selectivity, excellent reusability, high stability and enhanced binding ability towards Hg(II). FTIR and TGA analyses revealed that MIIP was based on the chemical chelation between ATU and Hg(II). The isotherm for Hg(II) adsorption on MIIP was dominated by monolayer chemical adsorption, and the adsorption kinetics featured a particularly rapid initial step. MIIP exhibited prominent effectiveness towards Hg(II) ions that were approximately twice that of MNIP and a higher selectivity for Hg(II) ions in the presence of interfering ions. The regeneration of Hg(II)–MIIP can be up to six regeneration cycles without any significant reduction in their rebinding capacity. Thus, the novel MIIP can be used for potential utilization in the removal of Hg(II) ions for industrial and environmental sustainability, mitigating pollutant transfer.

Acknowledgements

This study was financially supported by the Natural Science Foundation of China (51178213, 51238002, 51272099) and the National Science Fund for Excellent Young Scholars (51422807).

References

1. D. G. Streets, M. K. Devane, Z. Lu, T. C. Bond, E. M. Sunderland and D. J. Jacob, Environ. Sci. Technol., 2011, 45, 10485.
2. L. Magos and T. W. Clarkson, Ann. Clin. Biochem., 2006, 43, 257.
3. K. H. Nam, S. Gomez-Salazar and L. L. Tavlarides, Ind. Eng. Chem. Res., 2003, 42, 1955.
4. U.S.Environmental Protection Agency, 2001.
5. L. Sun, Y. X. Li, M. D. Sun, H. Q. Wang, S. F. Xu, C. Q. Zhang and Q. B. Yang, New J. Chem., 2011, 35, 2697.
6. A. Oehmen, D. Vergel, J. Fradinho, M. A. M. Reis, J. G. Crespo and S. Velizarov, J. Hazard. Mater., 2014, 264, 65.
7. M. Soylak and R. S. Cay, J. Hazard. Mater., 2007, 146, 142.
8. H. T. Wang and C. Z. Na, ACS Appl. Mater. Interfaces, 2014, 6, 20309.
9. Y. Z. Niu, R. J. Qu, H. Chen, L. Mu, X. G. Liu, T. Wang, Y. Zhang and C. G. Sun, J. Hazard. Mater., 2014, 278, 267.
10. E. K. Faulconer, N. V. H. Reitzenstein and D. W. Mazyck, J. Hazard. Mater., 2012, 199–200, 9.
11. H. Javadian and M. Taghavi, Appl. Surf. Sci., 2014, 289, 487.
12. L. X. Chen, S. F. Xu and J. H. Li, Chem. Soc. Rev., 2011, 40, 2922.
13. Y. T. Liu, R. H. Liu and C. B. Liu, J. Hazard. Mater., 2010, 182, 912.
14. X. B. Luo, S. L. Luo, Y. C. Zhan, H. Y. Shu, Y. N. Huanga and X. M. Tu, J. Hazard. Mater., 2011, 192, 949.
15. X. B. Luo, B. Guo, J. M. Luo, F. Deng, S. Y. Zhang, S. L. Luo and J. Crittenden, ACS Sustainable Chem. Eng., 2015, 3, 460.
16. X. B. Luo, L. L. Liu, F. Deng and S. L. Luo, J. Mater. Chem. A, 2013, 1, 8280.
17. H. Ebrahimzadeh, E. Moazzen, M. M. Amini and O. Sadeghi, Anal. Methods, 2012, 4, 3232.
18. J. M. Pan, X. H. Zou, Y. S. Yan, X. Wang, W. Guan, J. Han and X. Y. Wu, Appl. Clay Sci., 2010, 50, 260.
19. S. K. Li, F. Z. Huang, Y. Wang, Y. H. Shen, L. G. Qiu and A. J. Xie, J. Mater. Chem., 2011, 21, 7459.
20. A. Bagheri, M. Behbahani and M. M. Amini, Talanta, 2012, 89, 455.
21. R. K. Gautam, S. K. Sharma, S. Mahiya and M. C. Chattopadhyaya, Heavy Metals In Water: Presence, Removal and Safety, 2014.
22. M. Kaur, H. Han, A. Johnson, J. Kaczor, A. Paszczynski and Y. Qiang, MRS Online Proc. Libr., 2010, 1265, 1265-AA09-05.
23. H. Han, A. Johnson, J. Kaczor, M. Kaur, A. Paszczynski and Y. Qiang, J. Appl. Phys., 2010, 107, 09B520.
24. M. Kaur, H. Zhang and Y. Qiang, IEEE Magn. Lett., 2013, 4, 4000204.
25. M. Monier, I. M. Kenawy and M. A. Hashem, Carbohydr. Polym., 2014, 106, 49.
26. S. F. Xu, L. X. Chen, J. H. Li, Y. F. Guan and H. Z. Lu, J. Hazard. Mater., 2012, 237–238, 347.
27. Y. Xi, Y. T. Luo, J. M. Luo and X. B. Luo, J. Chem. Eng. Data, 2015, 60, 3253.
28. H. He, D. Xiao, J. He, H. Li, H. He, H. Dai and J. Peng, Analyst, 2014, 139, 2459.
29. L. Sun, Y. X. Li and M. D. Sun, New J. Chem., 2011, 35, 2697.
30. L. Y. Chai, T. Wang, L. Y. Zhang and H. Y. Wang, Carbon, 2015, 81, 748.
31. F. Q. Zhuang, R. Q. Tan and W. F. Shen, J. Chem. Eng. Data, 2014, 59, 3873.
32. M. Firouzzaren and Q. Q. Wang, Talanta, 2012, 101, 261.
33. Z. Zhang, J. H. Li, X. L. Song, J. P. Ma and L. X. Chen, RSC Adv., 2014, 4, 46444.
34. S. F. Xu, L. X. Chen, J. H. Li, Y. F. Guan and H. Z. Lu, J. Hazard. Mater., 2012, 237–238, 347.
35. D. K. Singh and S. Mishra, Desalination, 2010, 257, 177.
36. R. K. Gautam, A. Mudhoo, G. Lofrano and M. C. Chattopadhyaya, J. Environ. Chem. Eng., 2014, 2, 239.
37. Q. Z. Li, L. Y. Chai and Z. H. Yang, Appl. Surf. Sci., 2009, 255, 4298.
38. S. Banerjee, R. K. Gautam, A. Jaiswal, M. C. Chattopadhyaya and Y. C. Sharma, Advances, 2015, 5, 14425.
39. F. X. Luo, S. Y. Huang, X. D. Xiong and X. Q. Lai, RSC Adv., 2015, 5, 67365.

---