1 Introduction

With the rapid development of industry, water pollution caused by releasing of heavy metals has become a serious environmental problem. As a typical hypertoxic heavy metal, mercury (Hg(II)) is usually released directly to the environment through dissolving in liquids, and effectively removing aqueous Hg(II) has been a major challenge [1,2,3]. Various technologies have been developed to remove Hg from water, such as chemical precipitation, adsorption, and ion exchange etc [4,5,6]. Among these techniques, adsorptive removal is considered to be the most effective and economic one [7,8,9]. To date, a wide variety of adsorbents have been studied for Hg removal, however, conventional adsorbents often have several disadvantages such as low adsorption capacity and selectivity [10,11,12].

It has been recognized that incorporation of thiol groups onto porous materials could enhance the capture of heavy metals from aqueous solution, and the support materials often played a vital role [13, 14]. Although different porous supports such as SBA-15 [15], vermiculite [16], and Coal Fly Ash [17] have been investigated for the Hg2+ adsorption, these adsorbents suffer from some drawbacks such as low grafted amount of thiols and poor recyclability. Thus, development of new support materials is still urgent.

In this work, mesoporous silica nanoparticles with different morphologies were prepared and used as support materials for incorporation of thiol groups. In addition, factors affecting adsorption of Hg by the above materials were investigated, including adsorption kinetics, adsorption isotherms, and recyclability.

2 Experimental

2.1 Synthesis of mesoporous silica nanoparticles

The mesoporous silica nanoparticles were synthesized according to a reported method with minor changes [18]. In a typical procedure, 1.750 g of cetyltrimethylammonium bromide was dissolved into a mixture of 100 mL of deionized water and 25 mL of ethanol. Then, 2.5 mL of 1-Pentanol was added to the above solution. After stirring for 5 min, different amount of NH3∙H2O　(0.5, 1.5, or 2.5 mL) and 11 mL of tetraethoxysilane were added, resulting in the formation of white precipitate, followed by continuous stirring for 6 h at 40 °C. Subsequently, the precipitate was separated and washed with deionized water and ethanol for several times, and then dried at 50 °C in air. Finally, the as-prepared silica was calcined at 600 °C for 2 h in air. The obtained samples were labeled as SiO2-x, where x represents the amount of NH3∙H2O added during the synthesis.

2.2 Synthesis of thiol-functionalized mesoporous silica nanoparticles

Thiol-functionalized mesoporous silica nanoparticles were prepared by the following procedure. First, 5 mL of 3-mercaptopropyltrimethoxysilane (MPTS) was dropped into a suspension containing 50 mg of the as-prepared SiO2 nanoparticles dispersed in 30 mL of dried toluene. Then, the mixture was refluxed at 90 °C for 12 h. Finally, the white product was collected by centrifugation, washed several times with toluene and methanol, and dried at 80 °C for 12 h. The prepared sample was marked as SiO2-SH.

2.3 Characterization

The morphology of the samples was taken with a FEI Quanta 400 FEG scanning electron microscope (SEM). N2 adsorption–desorption isotherms were measured in a Quantachrome iQ2 porosimeter after sample degassing at 100 °C overnight. The pore size distribution was analyzed using Barrett–Joyner–Halenda (BJH) method based on the desorption branch. Average pore diameter was calculated as 4 V/A (V: pore volume; A: surface area). Thermogravimetric analysis (TGA, Netzsch STA 449 F5) measurement was performed under Ar gas at a flow rate of 100 cm3/min and a temperature ramp rate of 10 °C/min. Fourier‐Transform Infrared (FT-IR) spectra were measured with a Thermo Nicolet iS5 spectrometer.

2.4 Adsorption

Typically, 10 mg of adsorbent was put into 100 mL of Hg2+ solution with different concentrations. The mixture was drawn at regular intervals and filtered by a 0.45 µm filter membrane for analysis. The concentration of Hg2+ was measured with an elemental mercury analyzer (DMA-80).

3 Results and discussion

The morphologies of silica nanoparticles prepared with different amount of NH3∙H2O are shown in Figs. 1a−c. It can be clearly seen that the sample prepared with 0.5 mL of NH3∙H2O shows a flower-like nanospheres with wrinkles on its surface and has an average diameter of about 280 nm. With the amount of ammonia increased to 1.5 mL, the surface of nanoparticles became smooth with appearance of some concavities. There is also an increase of the diameter to around 400 nm. Further increasing the ammonia amount to 2.5 mL, sunken nanovesicles can be observed with diameter of 360 nm. The morphologies of nanoparticles prepared here are consistent with the reported results [18]. However, the particle size became larger, especially for sample SiO2-0.5 and SiO2-1.5, which may be due to the relatively large-scale preparation. With surface modification with MPTS, the morphologies of those silica nanoparticles show little change, as revealed in Figs. 1d−f. The MPTS on the surface of silica was investigated by Energy dispersive X-ray (EDX) elemental mapping, as shown in Figs. 1g−i, S element from MPTS could be detected and uniformly dispersed on the surface of silica nanoparticles, which suggests successful incorporation of mercaptopropyl group on the silica.

Fig. 1
figure 1

Scanning electron microscope (SEM) images of a SiO2-0.5, b SiO2-1.5, c SiO2-2.5, d SiO2-SH-0.5, e SiO2-SH-1.5, f SiO2-SH-2.5, and g elemental mappings (×7378 magnification) of h Si and i S of SiO2-SH-0.5

Full size image

The functional groups on the silica and MPTS modified silica was investigated by FT-IR (Fig. 2a). A broad peak at 3439 cm−1 and a strong peak around 1091 cm−1 are observed for all samples, which can be assigned to silanol -OH stretching and asymmetric Si-O-Si vibrations, respectively [19]. After modification with MPTS, a very weak peak appeared at 2560 cm−1 can be assigned to the -SH group [20], which further confirmed the successful grafting of S-H groups onto the silica. The surface area and porosity data of the materials are summarized in Table 1 and N2 adsorption–desorption isotherms of SiO2 and modified SiO2 are depicted in Fig. 2b and Fig. S1. It could be seen that SiO2-0.5 shows type IV isotherm with an obvious hysteresis loop at high relative pressure. The SiO2-0.5 shows the highest surface area and pore volume. There is a small decrease of surface area and pore volume with increasing ammonia amount for the SiO2 samples. However, a significant reduction of them is observed for MPTS modified SiO2, which should be attributed to occupation of the pores by organic groups. Interestingly, greater extent of loss in surface area was obtained for SiO2-1.5 and SiO2-2.5 ( > 66%) after surface functionalization, compared with that of SiO2-0.5 (with a loss of ~52). This phenomenon should be associated with relatively small pore size of the former, as partial pores in them may impede the entry of the organic groups during the functionalization.

Fig. 2
figure 2

a FT-IR of different SiO2 and SiO2-SH samples and b N2 adsorption-desorption isotherms of SiO2 -0.5 and SiO2-SH-0.5 (the insert is the corresponding pore size distributions)

Full size image
Table 1 The surface area and pore properties of SiO2 and SiO2-SH samples
Full size table

The TGA analysis of pure SiO2 and surface modified SiO2 samples were shown in Fig. S2. For pure SiO2 samples, an initial weight loss at < 100 °C was due to loss of adsorbed water molecule and little weight change can be seen with further increasing temperature. Whereas, for SiO2-SH samples, weight loss at temperature lower than 100 °C was smaller, and this was due to the presence of the hydrophobic mercaptopropyl group on the surface of silica [20]. Moreover, a weight loss at temperature > 300 °C could be assigned to the decomposition of mercaptopropyl group. The calculated mass loss between 300 and 800 °C was 17, 10, and 12% for SiO2-SH-0.5, SiO2-SH-1.5, and SiO2-SH-2.5, respectively, indicating that SiO2-SH-0.5 owns the most mercaptopropyl groups on its surface.

The adsorption capacities of different SiO2 and SiO2-SH for Hg2+ was comparatively studied (initial CHg(II) = 100 mg/L, T = 30 °C) and the results are shown in Fig. 3a. Pure SiO2 adsorbed small amount of Hg2+, after modification, the adsorption capacity improved dramatically. Among all the samples, SiO2-SH-0.5 exhibited the highest adsorption capacity of 87.3 mg/g, in virtue of the highest amount of mercaptopropyl group and surface area. Consequently, SiO2-SH-0.5 was selected for further study. The effects of initial Hg2+ concentration and contact time on the adsorption are shown in Figs. 3b, c, respectively. There is a rapid increase of adsorption capacity at initial stage and gradual stabilization as the initial Hg2+ concentration increased. The saturated adsorption capacity could be as high as 479 mg/g, which was much higher than other recently reported SH-functionalized sorbents, such as thiol modified Fe3O4@SiO2@C (184 mg/g) [21], Fe3O4@Cu3(btc)2 (348.43 mg/g) [22] and mercaptoamine-functionalized silica-coated magnetic nanoparticles (355 mg/g) [23]. It is also comparable with those of thiopyrene-containing porous carbon (518 mg/g) [24], MoS42− intercalated Layered double hydroxides (LDH) (500 mg/g) [25] and MOF (Metal-organic frameworks) FJI-H12 (439.8 mg/g) [26]. Figure 3c reveals that the adsorption process can be completed in 4 h, which demonstrated a rapid adsorption of Hg2+ with SiO2-SH-0.5.

Fig. 3
figure 3

a Adsorption capacity of different SiO2 and SiO2-SH samples for Hg2+, b adsorption isotherm of Hg2+ on SiO2-SH-0.5, and c effect of contact time on adsorption of Hg2+ on SiO2-SH-0.5 (the insert is the recyclability of SiO2-SH-0.5)

Full size image

Different typical isotherm and kinetic models have been used to further describe the adsorption behavior of SiO2-SH-0.5 for Hg removal. And their equations are shown in supporting information. Fig. S3 showed the fitting results of Langmuir and Freundlich isotherm models, and the parameters are presented in Table S1. It is found that Langmuir model provided a better fit to the adsorption data than Freundlich model, with higher correlation coefficient (R2 > 0.976). Moreover, the maximum adsorption capacity calculated by Langmuir equation was closer to the value obtained by experiment. This suggests that a monolayer adsorption occurred on the surface of SiO2-SH-0.5. The kinetic data are formulated by the pseudo-first-order model and pseudo-second-order model, respectively (Fig. S4). Compared with the pseudo-first-order model, the pseudo-second-order model gives better correlation coefficient ( > 0.998) for Hg2+ adsorption, and the calculated adsorption capacity is in good agreement with the experiment value (Table S2). This result indicates that the adsorption of Hg2+ by SiO2-SH-0.5 follows the pseudo-second-order model, and chemisorption may be involved in the adsorption process [27, 28].

After adsorption, SiO2-SH-0.5 was regenerated by washing with HCl solution (1 M) containing 1 wt. % of thiourea and subsequent separation and drying in air. Then, the recyclability was evaluated, and the results show that SiO2-SH-0.5 exhibit good reuse stability, for >94% of its original capacity could be retained after three adsorption–desorption cycles (Fig. 3c).

4 Conclusions

Three mesoporous silica nanoparticles with different morphologies were prepared by adjusting amount of ammonia during the synthesis. The sample with flower-like nanospheres morphology exhibited highest surface area and pore volume. Then the as-prepared silica nanoparticles were functionalized with MPTS and used as adsorbent for Hg2+ removal from aqueous solution. After modification, the sample with flower-like nanospheres morphology shows a highest Hg2+ adsorption capacity of 479 mg/g. Besides, the above adsorbent can adsorb Hg2+ quickly and be easily regenerated, which makes it potential to be an effective and promising adsorbent for Hg2+ removal.