Adsorption of Hg²⁺ by thiol functionalized hollow mesoporous silica microspheres with magnetic cores

Abstract

Novel hollow mesoporous silica spheres with magnetic cores (HMSMCs) were successfully synthesized by using hybrid magnetic carbon (Fe₃O₄/C) spheres as templates. The microspheres were further functionalized with (3-mercaptopropyl)trimethoxysilane (MPTS) to produce thiol functionalized HMSMCs (SH-HMSMCs), and their ability to absorb traces of toxic Hg²⁺ was evaluated. The characterization results revealed that the hollow microspheres were 250–300 nm in diameter. The thickness of the shell was about 50 nm, in which contained an inner core of Fe₃O₄ crystallites with a size of about 10 nm. It was also found that the saturation magnetization of the sample was 62.5 emu g⁻¹ and the BET surface area was 421 m² g⁻¹. These magnetic hybrid silica microspheres with thiol functional groups were found to have a high affinity to Hg²⁺, and were able to reduce even a low concentration of Hg²⁺ (<1 mg L⁻¹) down to about 0.53 μg L⁻¹, which was less than the Hg²⁺ content in the drinking water standard. The super strong affinity towards Hg²⁺ was attributed to the synergistic effect of the thiol groups and the unique structure of the microspheres. Moreover, the microspheres as adsorbents could be easily separated by an external magnetic field, and the adsorbed Hg²⁺ on the adsorbents could be removed by using hydrochloric acid, thus the adsorbents are readily reusable.

---

Introduction

Human activities related to gasoline combustion, mining and the electronic industry, have made mercury (Hg) a widespread pollutant in the atmosphere, water, soil and food.¹ Especially, inorganic Hg²⁺ pollutants in water impose a serious threat because of their mobility and complexity, showing features of wide distribution, low concentration and being difficult to control.² Furthermore, even a trace of Hg²⁺ contaminant is a serious threat to humans because its intrinsic toxicity is reinforced by its bioaccumulation capacity through the metabolic processes in the food chain leading to the destruction of brain tissue, lungs, kidney and the nervous system.³ Currently, for many countries, 1 μg L⁻¹ as the maximum Hg²⁺ concentration is allowed in drinking water.⁴ At the same time, there is a pressing need for developing methods to deal with pollution problems related to Hg²⁺. Conventional Hg²⁺ removal techniques, such as ion exchange, solvent extraction, precipitation and membrane separation,⁵ are generally effective for high concentrations of Hg²⁺ in solutions, but they are not capable of eliminating the remaining traces of Hg²⁺ selectively. Attempts have been made using adsorption methods to remove traces of Hg²⁺ effectively.⁶

A good adsorbent for water treatment not only needs to possess specific sites for capturing heavy metal ions, but it should also have a large surface area to accommodate the binding sites.^7,8 Silica-based mesoporous adsorbents such as MCM-41,^9–11 HMS^12,13 and SBA-15,^14–16 are robust inorganic solids which possess large specific surface areas, large pore volumes and easily modified pore surfaces,¹⁷ and have shown good potential in Hg²⁺ removal applications. In order to improve the adsorption capacity and the selectivity towards Hg²⁺, these silica-based adsorbents are usually modified by incorporating sulphur containing organic groups, such as thioureas,¹⁸ thioethers¹⁹ and mercapto coupling agents.²⁰ Considering the strong bonding ability between sulphur and Hg²⁺,^21,22 the adsorbents may capture Hg²⁺ to achieve national drinking water standards. Among them, thiol functionalized silica-based mesoporous adsorbents²³ have attracted extensive attention since they are expected to exhibit a specific binding ability toward highly toxic Hg²⁺, as a consequence of soft Lewis acid–base interactions.²⁴ MPTS is a widely used source for thiol groups, because it contains short chains of alkane thiols that do not block the pore channels.²⁵ On the other hand, by modifying these high surface area mesoporous materials with short chains of alkane thiol groups, the number of adsorbed binding sites could effectively increase, which results in a higher Hg²⁺ loading onto the surface of the adsorbents. At the same time, materials with unique pore structures greatly facilitate the access of Hg²⁺ to the thiol binding sites, resulting in higher metal-loading capacities as compared to those of substrates exhibiting no or less porosity, such as silica gels.²⁶ It has been reported that thiol functionalized mesoporous adsorbents with uniform pore structures have been found to demonstrate a high selectivity for Hg²⁺.²⁷

However, the recovery of these adsorbents after use is also important for practical applications. Considering the shortcomings of traditional recovery methods which are often time consuming, and that many of the used adsorbents are no longer usable and become secondary pollutants, we propose magnetic separation as a practical and effective means of adsorbent recovery.²⁸ Magnetic particles have been introduced into silica-based adsorbents, which usually assemble as solid yolk–shell structures,²⁹ hollow rattle-type structures³⁰ or with a structure of dispersed magnetic particles in the channels of mesoporous silica.³¹ Compared with the hollow rattle-type structure, the solid yolk–shell structure magnetic mesoporous silica spheres have no hollow cavity, which would have a relatively low surface area and adsorption capacity. The magnetic particles dispersed in the channels of mesoporous silica would block the pores, and lead to high mass transfer resistance in the process of metal ion diffusion. Obviously, rattle-type magnetic mesoporous silica is more promising for heavy metal ion removal. It has the advantages of (i) having capillary action which facilitates metal ions to diffuse into the binding sites on the shells,³² and (ii) possessing larger open surface areas and cavity volumes, which enable high heavy metal loading. Thus, numerous attempts³³ are being made to synthesize rattle-type magnetic mesoporous silica microspheres.

In this context, we report a new and facile route to fabricate hollow mesoporous silica spheres with magnetic cores (HMSMCs) using Fe₃O₄/C spheres as the templates. Scheme 1 illustrates the synthesis procedure. First, Fe₃O₄/C microspheres were synthesized by a solvothermal reaction. Second, the silica shells were deposited on the Fe₃O₄/C spheres through the simultaneous sol–gel polymerization of tetraethyl orthosilicate (TEOS) and cetyltrimethylammonium bromide (CTAB). Third, after the samples had undergone calcination to remove organic components followed by reduction in a hydrogen atmosphere, scattered Fe₃O₄ cores within hollow mesoporous silica spheres were obtained. Finally, these magnetic hollow mesoporous silica spheres were surface modified with appropriate organic functional thiol groups. The synthesized functional hybrid microspheres are composed of three components: (i) the core (inner layer) with the dispersed Fe₃O₄ particles, (ii) the mesoporous silica shell with large surface area, and (iii) the surface with thiol groups. The functional microspheres with thiol groups show highly selective and specific binding for Hg²⁺, and the large surface area silica spheres facilitate metal ions to diffuse into the binding sites. Due to the ferromagnetic properties of the Fe₃O₄ cores, the hollow mesoporous silica adsorbents could be easy separated by an externally applied magnetic field. Therefore, the adsorbent particles with Hg²⁺ after adsorption could be removed rapidly from the aqueous media.

Scheme 1 Synthesis of HMSMCs and the removal of Hg²⁺.

Experimental

Raw materials

All chemicals were analytical grade and used without further purification. Ferric chloride (FeCl₃·6H₂O), urea (CO(NH₂)₂), tetraethyl orthosilicate (TEOS, 98%), cetyltrimethylammonium bromide (CTAB, 99%), glucose, ethylene glycol, ammonia solution (28%), sodium hydroxide, hydrogen chloride (36–38%), hydrogen nitrate, anhydrous ethanol, mercuric nitrate (Hg(NO₃)₂·0.5H₂O), PBS buffer, cupric nitrate (Cu(NO₃)₂·3H₂O), zinc nitrate (Zn(NO₃)₂·6H₂O) and lead nitrate (Pb(NO₃)₂) were purchased from the Sinopharm Chemical Reagent Co. Ltd. (3-Mercaptopropyl)trimethoxysilane (MPTS, 90%) was purchased from Aladdin reagent (Shanghai) Co. Ltd.

Synthesis of Fe₃O₄/C microspheres

The Fe₃O₄/C microspheres were synthesized by a facile one-step solvothermal method.³⁴ In a typical sample preparation, 0.67 g FeCl₃·6H₂O was dissolved in 30 mL ethylene glycol to form a clear solution. Then 1.5 g urea and 0.08 g glucose were added into the above solution and vigorously stirred for 1 h. This mixed solution was transferred into a Teflon-lined stainless steel autoclave and heated at 200 °C for 10 h. Finally, the obtained solid products were magnetically separated by a magnet before washing with deionized water and anhydrous ethanol, and dried under vacuum at 40 °C for 24 h.

Synthesis of HMSMCs microspheres

0.2 g Fe₃O₄/C microspheres was dispersed in 100 mL anhydrous ethanol by sonication, and 1.8 mL concentrated ammonia solution was added to form solution A. Next, 0.3 g CTAB was dispersed in 105 mL deionized water under stirring to form solution B. Solution B was mixed with solution A under vigorous stirring for 6 h. Then 0.4 mL TEOS diluted with 5 mL ethanol was added drop-by-drop into the reaction mixture under stirring at room temperature for 4 h. The resultant particles were separated by a magnet before being washed with deionized water repeatedly and dried in a vacuum oven at 30 °C for 24 h. The as-prepared products were then calcined in air at 550 °C for 3 h at a heating rate of 5 °C min⁻¹, and deoxidized in 10% H₂/90% Ar at 350 °C for 2 h. The sample transformed into hollow mesoporous silica spheres with magnetic cores (denoted as HMSMCs). To rehydroxylate the samples, 0.2 g HMSMCs was suspended in PBS buffers of different pH levels (100 mL, pH = 6.5 and 7.5), then taken to reflux for 4 h at 110 °C. After being cooled to room temperature, the samples were separated by a magnet and were dried in a vacuum oven at 30 °C for 48 h. The obtained samples prepared at pH 7.5 and 6.5 were labeled as HMSMCs-A and HMSMCs-B, respectively.

Synthesis of SH-HMSMCs microspheres

0.5 g HMSMCs (HMSMCs-A or HMSMCs-B), 30 mL anhydrous toluene and 600 μL MPTS were added into a flask, and refluxed under a N₂ atmosphere for 8 h. Then, the modified HMSMCs were separated and then dried in a vacuum oven at 30 °C for 24 h. The sample product was denoted as thiol functionalized HMSMCs (SH-HMSMCs).

Evaluation of Hg²⁺ adsorption

For all the Hg²⁺ adsorption experiments, Hg(NO₃)₂·0.5H₂O was used as the Hg²⁺ source. For the Hg²⁺ selective adsorption experiment, 50 mg SH-HMSMCs were used to treat a 100 mL mixed metal solution containing Zn²⁺, Pb²⁺, Hg²⁺ and Cu²⁺ while stirring for 12 h. The adsorption capacity and the lowest absorption level of the samples were determined using a batch method. For each batch, 5 mg of adsorbent was stirred (using a thermostat stirrer at a speed of 300 rpm) in a vial containing 100 mL of various concentrations of mercuric nitrate solution (with pH 6.5 ± 0.05) at room temperature for 8 h. The different Hg²⁺ concentration solutions were prepared by diluting a 1000 mg L⁻¹ mercury solution with deionized water. The pH was adjusted with the appropriate amount of HNO₃ or NaOH. The adsorbent was separated from the adsorption medium using an external magnet bar. Langmuir and Freundlich equations were used to normalize the adsorption isotherms. The Langmuir equation is expressed as:

q_e = BQC_e/(1 + BC_e) (1)

where Q and B are the maximum adsorption capacity (mg g⁻¹) and the equilibrium adsorption constant (L mg⁻¹). q_e is the adsorption capacity at equilibrium (mg g⁻¹), C_e is the equilibrium concentration of the heavy metal (mg L⁻¹).

The Freundlich equation is expressed as:

q_e = K_FC_e^1/n (2)

where K_F and n are the Freundlich constants.

For the adsorption kinetic study, 5 mg SH-HMSMCs was suspended in a 100 mL mercury solution (1 mg L⁻¹ and 6 mg L⁻¹). The pseudo-second-order kinetic model was used to evaluate the sorption kinetics and calculate the rate constants, initial sorption rates and adsorption capacities:

t/q_t = 1/K_adq_e² + t/q_e (3)

where K_ad (g mg⁻¹ min⁻¹) is the rate constant, q_e (mg g⁻¹) is the equilibrium adsorption capacity, q_t (mg g⁻¹) is the amount of heavy metal adsorbed at time t.

To regenerate the used absorbents, the SH-HMSMCs with adsorbed Hg²⁺ were treated with concentrated HCl (1 M) under sonication for 30 min, and then separated by an external magnetic field. After washing several times with deionized water, the absorbents were dried in a vacuum oven at 40 °C for 12 h. The dried absorbents were re-used for a second and a third time. Similar experimental procedures were repeated.

Characterization

The crystalline structures of the samples were identified by X-ray diffraction analysis (XRD, Philips X’pert PRO) using Ni-filtered monochromatic CuKα radiation at 40 kV and 40 mA. The morphology and structure of the products were characterized using a field emission scanning electron microscope (FESEM, Sirion 200 FEI) using an accelerating voltage of 10 kV, and transmission electron microscopy (TEM, JEOL-2010, 200 kV) with an energy dispersive X-ray spectrometer (EDS Oxford, Link ISIS). The powders were ultrasonically-dispersed in ethanol. Then, the suspensions were dropped onto the SEM stub and holey-carbon grid for SEM and TEM examination, respectively. Fourier transform infrared (FT-IR) spectroscopy, performed with KBr disks of powdered samples, was used to identify the functional groups of the SH-HMSMCs. The specific surface areas and the pore distributions of the samples were calculated by nitrogen adsorption (Micrometrics ASAP 2020M) at 77 K using the Brunauer–Emmett–Teller (BET) equation. X-ray photoelectron spectroscopy (XPS) analysis was performed using an ESCALAB 250 X-ray photoelectron spectrometer (Thermo, America) equipped with Al Kα_1,2 monochromatized radiation at 1486.6 eV as an X-ray source. Thermo-gravimetric analysis (TGA, PerkinElmer Pyris Diamond) was carried out at a heating rate of 10 °C min⁻¹ in air. The magnetic measurements were performed using a superconducting quantum interference device (SQUID) magnetometer (Quantum design, MPMS XL). Quantitative determination of the metal ion content was performed by inductively coupled plasma (ICP) or inductively coupled plasma mass spectrometry (ICP-MS).

Results and discussion

Microstructures

Fig. 1 shows the SEM images of Fe₃O₄/C, Fe₃O₄/C@SiO₂, HMSMCs and the XRD patterns of Fe₃O₄/C and HMSMCs. It could be clearly observed that the Fe₃O₄/C microspheres had a rough surface but excellent monodispersity with a size of 180–220 nm in diameter (Fig. 1a). After coating with an organic silica, the size of the Fe₃O₄/C@SiO₂ microspheres became larger (250–300 nm) and the surface appeared smooth (Fig. 1b). After calcination to remove the organic components, the size of the HMSMCs microspheres (Fig. 1c) had no obvious change as compared to those shown in Fig. 1b. The patterns in Fig. 1d show that the diffraction peaks that appear at 2θ of 30.2°, 35.5°, 43.2°, 53.4°, 57.3° and 62.7° correspond to the (220), (311), (400), (422), (511), and (440) planes of Fe₃O₄ (JCPDS card no. 19-0629).³⁵ All of the diffraction peaks of the HMSMCs sample are similar to those of Fe₃O₄/C, indicating that Fe₃O₄ existed in the sample and had not been changed during the entire preparation process, and the broad diffraction peaks at around 23° corresponded to amorphous silica shells. The results obtained were in good agreement with the XRD patterns of Fe₃O₄@SiO₂ microspheres previously reported by others.³⁶

Fig. 1 SEM images of (a) Fe₃O₄/C, (b) Fe₃O₄/C@SiO₂, and (c) HMSMCs. (d) The XRD patterns of Fe₃O₄/C and HMSMCs.

TEM was used to further examine the microstructures of the samples. Fig. 2a showed that the Fe₃O₄/C spheres were a similar size and hollow. The high-resolution TEM (HRTEM) image (Fig. 2b) showed that the Fe₃O₄/C microspheres were actually an aggregation of small primary crystals with a size of 10–12 nm. The black Fe₃O₄ particles were encapsulated within gray carbon shells (3–5 nm thickness, seen in the enlarged image inset in Fig. 2b). The inter-planar spacing of about 0.29 nm corresponded to the (220) lattice planes of a cubic Fe₃O₄ crystal, in agreement with the XRD analysis. Fig. 2c and d clearly show that the synthesized hybrid microspheres have a core–shell structure with an outer light silica shell with a thickness of 50 nm and a dark inner core of Fe₃O₄/C particles, indicating the successful preparation of the core–shell structured hybrid mesoporous silica precursor. After removing the carbon and organic surfactant, rattle-type hollow mesoporous silica spheres with magnetic cores were formed. Fig. 2e shows that the dark Fe₃O₄ particles comprising the core structure are evenly dispersed in the hollow mesoporous silica spheres. The HRTEM image (Fig. 2f) indicated that the silica shells possessed wormlike mesopores (about 2 nm) derived from the CTAB-templating,³⁷ and that the coated magnetite cores mainly retained the original structure and morphology. The above results confirmed that the magnetic Fe₃O₄ particles had been successfully loaded into the hollow mesoporous silica spheres by using carbon and CTAB as cavity and pore forming agents.

Fig. 2 TEM images: (a and b) Fe₃O₄/C, (c and d) Fe₃O₄/C@SiO₂, and (e and f) HMSMCs.

Fig. 3a displays the thermo-gravimetric analysis (TGA) results for Fe₃O₄/C and Fe₃O₄/C@SiO₂. The data suggested that the initial weight loss of approximately 2 wt% between 40 and 200 °C was due to the loss of surface hydroxyls and water content in the hybrid microspheres. The maximum weight loss observed at the second step ranging from 200 to 400 °C was approximately 5 wt% in Fe₃O₄/C and 15 wt% in the Fe₃O₄/C@SiO₂ sample. These suggested that the combustion of amorphous carbon began at a temperature higher than 200 °C, and was nearly completed at around 400 °C. There was a little rise between 400 and 600 °C in the curve of Fe₃O₄/C, which might be attributed to the transformation of Fe₃O₄ to Fe₂O₃ due to oxidation, while carbon would be further eliminated with increasing temperature.³⁸ When the temperature was higher than 600 °C, the final weight loss could be a response to the reduction of the residual carbon. Comparing the 8% weight loss of Fe₃O₄/C, the weight loss of Fe₃O₄/C@SiO₂ was approximately 20 wt% from 20 °C to 700 °C. This reflected that considerable amounts of the CTAB surfactant had been deposited on the magnetic carbon spheres successfully, which contributed to the meso-channels of the magnetic hollow mesoporous silica hybrid microspheres.³²

Fig. 3 (a) TGA curves of Fe₃O₄/C and Fe₃O₄/C@SiO₂. (b) FTIR spectra of HMSMCs and SH-HMSMCs.

FTIR spectroscopy was performed to investigate the functional groups in the samples. As shown in Fig. 3b, the minimum of the transmittance curve at around 1080–1100 cm⁻¹ was assigned to the asymmetric stretching vibration of the Si–O–Si bonds of the inorganic framework. The appearance of a characteristic vibration band at 578 cm⁻¹ attributed to the Fe–O of Fe₃O₄ in both the HMSMCs and SH-HMSMCs samples,³⁹ also confirmed the existence of Fe₃O₄ nanoparticles in the hybrid microspheres. The broad peak at 3200–3600 cm⁻¹ could be assigned to the stretching vibration of –OH in the lattice and the water molecules. The peak at 1600 cm⁻¹ was assigned to the bending vibration of H₂O and –OH, which implied that hydroxyl groups existed in the as-synthesized microspheres. Compared with the spectrum of HMSMCs microspheres, that of the SH-HMSMCs microspheres revealed that the weak characteristic peak at 2560 cm⁻¹ of –SH, the peaks of C–H (2800–3000 cm⁻¹), and C–C, C–H₂ (1300–1500 cm⁻¹) were also found. The stretching vibration bands at 2855 cm⁻¹ and 2923 cm⁻¹ were assigned to the C–H stretching vibrations of the propyl chains of the integrated MPTS moieties and the weak peak at 2560 cm⁻¹ successfully proved the presence of –SH groups in the mesoporous silica shell of the SH-HMSMCs microspheres. In order to further investigate the elemental distribution of the sulphur (Fig. S1, ESI†), the elemental mapping of sulphur was also performed. It can be seen in Fig. S1c–e, ESI† that the S, Si and O elements were distributed throughout the entire microsphere. Moreover, the EDS spectrum of SH-HMSMCs (Fig. S1a, ESI†) confirmed the presence of S. All these results indicated that MPTS had been grafted onto the surface of the HMSMCs.

The porosities of Fe₃O₄/C@SiO₂, HMSMCs and SH-HMSMCs were investigated using a N₂ adsorption desorption technique. The N₂ adsorption desorption isotherms of all of the samples were of type IV with a distinct hysteresis loop in the high-pressure region (P/P₀ = 0.4–1.0) as shown in Fig. 4, indicating that the mesoporous characteristics included capillary condensation. The results (Table S1, ESI†) also showed that the BET specific surface areas of HMSMCs and SH-HMSMCs were determined to be 420.66 m² g⁻¹ and 380.75 m² g⁻¹, respectively. The BET pore sizes of the HMSMCs and SH-HMSMCs samples were 2.15 nm and 1.91 nm, respectively, which were consistent with the results from the HRTEM analysis (Fig. 2f). The decrement of the surface area and pore size of the SH-HMSMCs microspheres were attributed to the modification of the MPTS functionality in the mesopore channels.

Fig. 4 (a) N₂ adsorption desorption isotherms of SH-HMSMCs, HMSMCs and Fe₃O₄/C@SiO₂. (b) The size distribution of SH-HMSMCs, HMSMCs and Fe₃O₄/C@SiO₂.

The magnetization curves of Fe₃O₄/C, HMSMCs and SH-HMSMCs (Fig. 5a) were obtained using a vibrating sample magnetometer at room temperature. The saturation magnetization values of Fe₃O₄/C, HMSMC and SH-HMSMCs were 98.5, 62.5 and 55.4 emu g⁻¹, respectively. The SH-HMSMC microspheres dispersed in water showed a rapid separation in 30 seconds soon after the magnet was placed near the glass vial (Fig. 5b). This suggested that the synthesized SH-HMSMCs could be separated easily from a solution mixture by an external magnetic field.

Fig. 5 (a) Magnetization curves of SH-HMSMCs, HMSMCs and Fe₃O₄/C. (b) Pictures of the aqueous suspension of SH-HMSMCs (left) and the same suspension magnetically separated after 30 s (right).

Effects of pH

To investigate the effects of the pH level on the removal of Hg²⁺, the pH level in the solution was adjusted between the range of 2.0 and 7.0 with a Hg²⁺ concentration of 50 mg L⁻¹. As shown in Fig. 6a, the maximum Hg²⁺ removal (131 mg g⁻¹) was observed in the case of SH-HMSMCs in a pH 6.2 solution. In contrast, the HMSMC sample showed only about 15 mg g⁻¹ of Hg²⁺ removal in a pH 6.2 solution. Moreover, the SH-HMSMCs showed good adsorption capacity for Hg²⁺ at pH values in the range of 2.0–7.0. This indicated that the organic thiol groups were integrated over a large surface area mesoporous silica network to enhance the adsorption capacity. It could also be clearly seen that the capacity of Hg²⁺ adsorption is reduced with decreasing pH because the affinity between Hg²⁺ and the sorbent is reduced. This also reflected the competitive adsorption process that was carried out between H⁺ and Hg²⁺. This result indicated that the loaded mercury on the adsorbents might be desorbed with acid, and favoured the regenerability of the adsorbents.

Fig. 6 (a) Hg²⁺ adsorption efficiency of SH-HMSMCs and HMSMCs as a function of pH. (b) Effect of adsorbent dose on Hg²⁺ removal at pH 6.5, Hg²⁺ concentration: 10 mg L⁻¹, adsorption time: 12 h.

Effect of adsorbent dose

To choose a moderate adsorbent concentration for the adsorption experiment, the effect of the adsorbent dose was measured in the removal of Hg²⁺. Different mass amounts of the adsorbents (10, 20, 35, 50, 65 and 80 mg L⁻¹) were used to treat a 100 mL bulk solution of Hg²⁺ (initial concentration was 10 mg L⁻¹) with other operational parameters kept constant, with continuous stirring for 12 h. As shown in Fig. 6b, the removal efficiency of the heavy metal Hg²⁺ was obviously raised along with the increase in the SH-HMSMCs addition amounts. When the concentration of SH-HMSMCs was greater than 50 mg L⁻¹, the removal efficiency of Hg²⁺ was over 95%, and further improved with the amount of absorbent dose; the Hg²⁺ removal efficiency did not change significantly. This result indicated that the best adsorbent dose was about 50 mg L⁻¹.

Adsorption kinetics

The adsorption kinetics of SH-HMSMCs for Hg²⁺ were measured using 1 mg L⁻¹ and 6 mg L⁻¹ of Hg²⁺ solution at pH 6.5 at room temperature, respectively. As shown in Fig. 7a, initially the adsorption capacity was rapidly enhanced, and then slowed down, finally reaching equilibrium. In the rapid process, within 20 min, approximately 90% adsorption was accomplished for the 1 mg L⁻¹ solution, and approximately 70% for the 1 mg L⁻¹ solution. These results showed that the adsorbent required a short time to achieve a high adsorption efficiency. A longer time was needed for the initial concentration 6 mg L⁻¹ solution to reach the equilibrium concentration than the 1 mg L⁻¹ solution, which indicated that adsorption of a low concentration of Hg²⁺ by the SH-HMSMCs was much easier than a high concentration of Hg²⁺. As shown in Fig. 7b, the adsorption of Hg²⁺ fitted well with a pseudo-second-order kinetic equation, and the constant K_ad was calculated to be 3.6 × 10⁻³ and 5.6 × 10⁻³ g mg⁻¹ min⁻¹ for the initial concentrations of 1 mg L⁻¹ and 6 mg L⁻¹, respectively (Table S2, ESI†).

Fig. 7 (a) Adsorption kinetics of SH-HMSMCs in an aqueous solution of Hg²⁺ at the initial concentrations: 1 mg L⁻¹ and 6 mg L⁻¹, and (b) the pseudo-second-order adsorption rates.

Selective Hg²⁺ adsorption behavior and trace Hg²⁺ removal by SH-HMSMCs

To confirm the selective function and efficiency of the thiol groups towards Hg²⁺ adsorption, the adsorption capacities of the SH-HMSMC samples were compared with those of the HMSMCs. Based on the observed concentration changes of the mixed metal ions in solutions before and after the treatment with the SH-HMSMCs as shown in Table 1, it could be clearly seen that the Hg²⁺ ions had been removed completely, while the concentrations of other metal ions (Pb²⁺, Zn²⁺, Cu²⁺) were slightly decreased. This suggested that the thiol functionalized HMSMCs exhibited a much higher complexation affinity for Hg²⁺ as compared with the other metal ions. While the mixed metal ion solution was directly treated with the HMSMCs under identical conditions, no significant changes in the concentrations of the metal ions were noted, indicating that the thiol group was vital for the removal of the metal ions. Moreover, the molar adsorption capacities of the SH-HMSMCs for the metal ions in the mixed solutions calculated from the data in Table 1 (seen in Table S3, ESI†) showed that the complexing capability between the metal ions and the grafted –SH groups was in the following sequence: Hg²⁺ > Cu²⁺ > Pb²⁺ > Zn²⁺. This is consistent with the concept of Lewis acid–base interactions and the corresponding enthalpy change values.²⁴

Table 1 The metal ion concentrations in the mixed solutions before and after the treatment with HMSMCs and SH-HMSMCs, analyzed by ICP. 50 mg of adsorbent was added into 100 mL of a mixed heavy metal ion solution at room temperature at pH 6.5 with agitation for 12 h

	Hg^2+ (mg L^−1)	Pb^2+ (mg L^−1)	Zn^2+ (mg L^−1)	Cu^2+ (mg L^−1)
Original solution	8.48	7.19	9.33	12.08
HMSMCs	8.41	7.08	9.41	12.07
SH-HMSMCs	0	6.07	9.38	11.06

---

To further verify the SH-HMSMC selective adsorption behavior and their higher complexation affinity for Hg²⁺, SH-HMSMC samples saturated with adsorbed Pb²⁺ or Cu²⁺ were immersed in a solution with or without Hg²⁺ for 10 h. The results (Table S4, ESI†) show that nearly 23.7% of Pb²⁺ was desorbed from the adsorbent by the Hg²⁺, while a comparative experiment only showed about 1.2% desorption in the water solution. These results illustrated that the Pb²⁺ loading could be reduced from the adsorbent by the presence of Hg²⁺, because the mercury–sulphur interaction is stronger than the lead–sulphur interaction. A similar result was found with the Cu²⁺-adsorbent. Therefore, the high selectivity of SH-HMSMCs for Hg²⁺ is owed to the fact that softer transition metal ions (Hg²⁺) are prone to forming stable complexes with ligands carrying softer donor atoms.⁸

At low concentrations of metal ions, the adsorption performance of the SH-HMSMCs was investigated using ICP-MS. It could be clearly seen that the low concentration of Hg²⁺ could be further reduced to a value below 1 μg L⁻¹ by the SH-HMSMCs (Table S5, ESI†), which reaches the drinking water standard allowed by worldwide legislation. At the same time, the concentrations of other trace metal ions (Pb²⁺, Zn²⁺, Cu²⁺) show a significant decrease but not as much as the Hg²⁺. These results further demonstrate that the adsorbents had a highly selective adsorption behavior for Hg²⁺. It was found that the lowest level of Hg²⁺ in the solution treated by the SH-HMSMCs was about 0.53 μg L⁻¹. Therefore, this adsorbent showed a promising application for the removal of trace mercury from aqueous media.

To illustrate the –SH complexation with Hg²⁺, the adsorbents were characterized using XPS. As can be seen in Fig. 8a, the binding energy of the narrow XPS spectrum for S2p of the RSH groups was 163.8 eV. After adsorption of Hg²⁺, the binding energy of S2p shifted to a slightly higher value, which might be due to the donation of electrons from the S atoms of the thiol groups to Hg.⁴⁰ The full-range XPS spectra for the SH-HMSMCs before and after Hg²⁺ adsorption are shown in Fig. S2 (ESI†). The mercury peaks were at binding energies of around 100.8 eV and 105.1 eV of the Hg4f_7/2 and Hg4f_5/2 peaks (Fig. 8b), indicating that Hg²⁺ was adsorbed through the chelating binding between mercury and sulfur.

Fig. 8 (a) S2p narrow XPS scan of SH-HMSMCs and SHg-HMSMCs. (b) Hg4f XPS scan of SHg-HMSMCs.

Adsorption isotherm and structure enhanced adsorption

The adsorption isotherm experiments were carried out at different initial concentrations of Hg²⁺ from 0.1 mg L⁻¹ to 50 mg L⁻¹ at a pH level of 6.5 ± 0.05 (adsorbent loading 5 mg per 100 mL). The curves of the isotherms demonstrated the adsorption as a function of the equilibrium concentration of the metal ions in solution. As shown in Fig. 9a, the thiol functionalized Fe₃O₄/C@SiO₂ (BET surface area: 102.76 m² g⁻¹ as shown in Table S1, ESI†) only showed about half the adsorption capacity than that of the SH-HMSMCs. The high surface area and the mesoporous channels of the magnetic mesoporous silica were responsible for this performance, which not only contributed to the modification of large amounts of thiol groups in the mesopore channels but also provided channels for the facile diffusion of metal ions towards the binding sites. Moreover, compared with the SH-HMSMCs and SH-Fe₃O₄/C@SiO₂ samples, the Fe₃O₄/C@SiO₂ and HMSMCs samples (Fig. S3, ESI†) only showed 21.7 and 15.1 mg g⁻¹ of Hg²⁺ adsorption, respectively. These results showed that the thiol groups were capable of forming a complex with Hg²⁺ ions selectively and removing them efficiently from the aqueous solutions. Therefore, the high surface area mesoporous silica was modified with thiol groups, and the adsorbed binding sites could be effectively increased, which resulted in more Hg²⁺ loading onto the adsorbent. The unique mesopore structure of these samples largely facilitated the access of the metal ions to the thiol binding sites, resulting in higher metal-loading capacities.

Fig. 9 (a) Adsorption isotherms of Hg²⁺ on SH-HMSMCs, SH-HMSMCs-A, SH-HMSMCs-B and SH-Fe₃O₄/C@SiO₂. (b) Removal efficiency of Hg²⁺ in different cycles by SH-HMSMCs.

In addition, the number of thiol functionalized groups was greatly ascribed to the population of the silanol groups on the mesoporous silica surface.²⁵ Thus, we could rehydroxylate the magnetic hollow mesoporous silica, and expand the amount of Si–OH groups by refluxing in different pH buffers. HMSMCs-A and HMSMCs-B were acquired by suspending magnetic mesoporous silica in 100 mL of buffer at pH = 7.5 and pH = 6.5, and heating to reflux for 4 h with the other operational parameters unchanged. As shown in the adsorption isotherms (Fig. 9a), it was indicated that the Hg²⁺ removal capacity of SH-HMSMCs-A was more effective than that of SH-HMSMCs-B. This also confirmed that the surface hydroxylation had facilitated the removal of Hg²⁺. This result could be explained by the rehydroxylation that could effectively expand the number of Si–OH groups in a weak alkaline environment, finally contributing to the grafting of more thiol groups.

Both Langmuir and Freundlich adsorption isotherms were used to normalize the adsorption data. As shown in Table S6 (ESI†), the correlation coefficients (R²) for the Langmuir model are larger than that from the Freundlich model. We also found that the Langmuir model fitted better than the Freundlich model, demonstrating that the Hg²⁺ adsorption of the SH-HMSMCs could be described as a monolayer adsorption process. The maximum adsorption capacity of the SH-HMSMCs for Hg²⁺ was determined to be 118.6 mg g⁻¹, which shows a greater adsorption capacity than lots of reported adsorbents (Table S7 ESI†).^41–44

Regenerability of the adsorbent

For practical application, the recycling and regeneration of the adsorbent is important. Therefore, the regeneration ability of the SH-HMSMCs was examined (the initial concentration of Hg²⁺ in the treatment cycles: 1 mg L⁻¹, adsorbent loading 5 mg per 100 mL). The pH effect on the adsorption of Hg²⁺ by the SH-HMSMCs was examined. The results showed that the adsorption of Hg²⁺ by the SH-HMSMCs decreased with a decrease in the pH values, that was to say the affinity between H⁺ and sorbent was enhanced in the competitive adsorption between H⁺ and Hg²⁺ at low pH value. This illustrated that the adsorbent could be refreshed easily by acid treatment. Thus, the SH-HMSMCs loaded with Hg²⁺ were treated in 1 M HCl under sonication for 30 min. The results showed that 89% of the adsorbed Hg²⁺ was released. The regenerated adsorbent could be re-used to adsorb Hg²⁺ in subsequent cycles. We repeated the above procedure for five cycles. As shown in Fig. 9b, the removal efficiency was slightly reduced in the subsequent cycles. However, the reusability was over 80% even after five cycles.

Conclusions

Novel hollow mesoporous spheres grafted with functional thiol groups, with a magnetic Fe₃O₄ core and mesoporous silica shell, were fabricated successfully. The products exhibited a unique selective adsorption for mercury and could lower the trace of mercury in drinking water to an acceptable limit. The adsorbent was recycled five times showing over 80% of the original performance. The approach in this work could be a promising technique for removing trace amounts of mercury from aqueous media.

Acknowledgements

This work was supported by the National Basic Research Program of China (Grant no. 2013CB934302), the Natural Science Foundation of China (Grant no. 21177132 and 51272255) and Strategic Priority Research Program of the Chinese Academy of Sciences (Grant no. XDA09030200).

Notes and references

1. G. M. Assessment, United Nations Environment Programme (UNEP) Chemicals: Geneva, Geneva, 2002.
2. D. Foucher and H. Hintelmann, Environ. Sci. Technol., 2008, 43, 33–39.
3. J. Wang, X. Feng, C. W. Anderson, Y. Xing and L. Shang, J. Hazard. Mater., 2012, 221, 1–18.
4. W. H. Organization, Guidelines for drinking-water quality: first addendum to volume 1, recommendations, World Health Organization, 2006.
5. F. Fu and Q. Wang, J. Environ. Manage., 2011, 92, 407–418.
6. I. Ali, Chem. Rev., 2012, 112, 5073–5091.
7. Z. Geng, Y. Lin, X. Yu, Q. Shen, L. Ma, Z. Li, N. Pan and X. Wang, J. Mater. Chem., 2012, 22, 3527–3535.
8. A. M. Liu, K. Hidajat, S. Kawi and D. Y. Zhao, Chem. Commun., 2000, 1145–1146.
9. A. C. Pradhan and K. M. Parida, J. Mater. Chem., 2012, 22, 7567–7579.
10. A. Walcarius, M. Etienne and B. Lebeau, Chem. Mater., 2003, 15, 2161–2173.
11. V. Antochshuk and M. Jaroniec, Chem. Commun., 2002, 258–259.
12. W. Huang, Y. Zhu, J. Tang, X. Yu, X. Wang, D. Li and Y. Zhang, J. Mater. Chem. A, 2014, 2, 8839–8848.
13. L. Mercier and T. J. Pinnavaia, Environ. Sci. Technol., 1998, 32, 2749–2754.
14. M. Lombardo, M. Videla, A. Calvo, F. Requejo and G. Soler-Illia, J. Hazard. Mater., 2012, 223, 53–62.
15. E. Da’na and A. Sayari, Desalination, 2012, 285, 62–67.
16. J. Aguado, J. M. Arsuaga and A. Arencibia, Microporous Mesoporous Mater., 2008, 109, 513–524.
17. A. Walcarius and L. Mercier, J. Mater. Chem., 2010, 20, 4478–4511.
18. M. Mureseanu, A. Reiss, N. Cioatera, I. Trandafir and V. Hulea, J. Hazard. Mater., 2010, 182, 197–203.
19. L. Zhang, W. Zhang, J. Shi, Z. Hua, Y. Li and J. Yan, Chem. Commun., 2003, 210–211.
20. S. A. Idris, S. R. Harvey and L. T. Gibson, J. Hazard. Mater., 2011, 193, 171–176.
21. K.-K. Yee, N. Reimer, J. Liu, S.-Y. Cheng, S.-M. Yiu, J. Weber, N. Stock and Z. Xu, J. Am. Chem. Soc., 2013, 135, 7795–7798.
22. A. Bibby and L. Mercier, Chem. Mater., 2002, 14, 1591–1597.
23. T. Sangvanich, J. Morry, C. Fox, W. Ngamcherdtrakul, S. Goodyear, D. Castro, G. E. Fryxell, R. S. Addleman, A. O. Summers and W. Yantasee, ACS Appl. Mater. Interfaces, 2014, 6, 5483–5493.
24. E. F. S. Vieira, J. D. Simoni and C. Airoldi, J. Mater. Chem., 1997, 7, 2249–2252.
25. X. Feng, G. E. Fryxell, L. Q. Wang, A. Y. Kim, J. Liu and K. M. Kemner, Science, 1997, 276, 923–926.
26. J. Liu, X. Feng, G. E. Fryxell, L. Q. Wang, A. Y. Kim and M. Gong, Adv. Mater., 1998, 10, 161–165.
27. J. Brown, L. Mercier and T. J. Pinnavaia, Chem. Commun., 1999, 69–70.
28. Y. X. Zhang, X. Y. Yu, Z. Jin, Y. Jia, W. H. Xu, T. Luo, B. J. Zhu, J. H. Liu and X. J. Huang, J. Mater. Chem., 2011, 21, 16550–16557.
29. Y. Deng, D. Qi, C. Deng, X. Zhang and D. Zhao, J. Am. Chem. Soc., 2008, 130, 28–29.
30. W. Zhao, H. Chen, Y. Li, L. Li, M. Lang and J. Shi, Adv. Funct. Mater., 2008, 18, 2780–2788.
31. J. Liu, S. Z. Qiao, Q. H. Hu and G. Q. Lu, Small, 2011, 7, 425–443.
32. M. S. Moorthy, D. J. Seo, H. J. Song, S. S. Park and C. S. Ha, J. Mater. Chem. A, 2013, 1, 12485–12496.
33. Y. Zhu, T. Ikoma, N. Hanagata and S. Kaskel, Small, 2010, 6, 471–478.
34. J. Zheng, Z. Q. Liu, X. S. Zhao, M. Liu, X. Liu and W. Chu, Nanotechnology, 2012, 23, 165601.
35. G. H. Zhao, J. Z. Wang, X. M. Peng, Y. F. Li, X. M. Yuan and Y. X. Ma, Chem.–Asian J., 2014, 9, 546–553.
36. Y. Zhu, E. Kockrick, T. Ikoma, N. Hanagata and S. Kaskel, Chem. Mater., 2009, 21, 2547–2553.
37. L. Guo, J. Li, L. Zhang, J. Li, Y. Li, C. Yu, J. Shi, M. Ruan and J. Feng, J. Mater. Chem., 2008, 18, 2733.
38. S. H. Xuan, L. Y. Hao, W. Q. Jiang, X. L. Gong, Y. Hu and Z. Y. Chen, Nanotechnology, 2007, 18, 035602.
39. W. Jiang, W. Wang, B. Pan, Q. Zhang, W. Zhang and L. Lv, ACS Appl. Mater. Interfaces, 2014, 6, 3421–3426.
40. S. X. Zhang, Y. Y. Zhang, J. S. Liu, Q. Xu, H. Q. Xiao, X. Y. Wang, H. Xu and J. Zhou, Chem. Eng. J., 2013, 226, 30–38.
41. C. B. Lopes, M. Otero, J. Coimbra, E. Pereira, J. Rocha, Z. Lin and A. Duarte, Microporous Mesoporous Mater., 2007, 103, 325–332.
42. C. Wang, S. Tao, W. Wei, C. Meng, F. Liu and M. Han, J. Mater. Chem., 2010, 20, 4635–4641.
43. Z. M. Chen, Z. R. Geng, Z. Y. Zhang, L. B. Ren, T. X. Tao, R. C. Yang and Z. X. Guo, Eur. J. Inorg. Chem., 2014, 20, 3172–3177.
44. J. Zhao, B. Zhu, H. Yu, L. Yan, Q. Wei and B. Dua, J. Colloid Interface Sci., 2013, 389, 46–52.

---

Footnote

† Electronic supplementary information (ESI) available: Additional figures including an EDS spectrum, elemental mapping, full-range XPS spectra and so on. See DOI: 10.1039/c5ra05184c

---