In recent years, silica nanoparticles (SNPs) have received significant attention due to their chemical inertness, nontoxicity, optical transparency, and excellent thermal stability 1 2 3 , which can be widely used in catalysis 4 , chemical process industry 5 , removal of metal ions 6 , and metal ion preconcentration 7 8 9 through polymer coatings or other functional groups. For many applications, there are several chemical methods for controlling the nanoparticles’ surface functionality, such as chemisorptions 10 , sol–gel process, and immobilization of organic molecules by silane coupling reagents 11 , which can result in the immobility, mechanical stability, and water insolubility of functionalized SNPs 12 . Polymerization methods involving living free radical nitroxide-mediated polymerization, reversible addition-fragmentation chain transfer (RAFT), and atom transfer radical polymerization (ATRP) were commonly considered as effective techniques to functionalize materials 13 14 15 16 . Especially, the ATRP technique has been developed rapidly due to the advantages of being simple, inexpensive, and more general for controlled radical polymerization when it was firstly proposed by Wang and Matyjaszewsi 17 . There are a greater number of researchers using ATRP to functionalize SNPs 18 19 20 21 22 .

With the appearance of the click chemistry proposed by Sharpless 23 , the copper(I)-catalyzed azide-alkyne 1,3-dipolar cycloaddition reaction also becomes a particularly powerful approach to synthesize designed molecules or modify inorganic materials because only mild reaction conditions are required and the extreme selectivity toward molecules bearing azides and alkynes prevents unwanted side product 24 25 26 . Recently, some literatures reported that gold nanoparticles, silicon oxides, carbon nanotubes, and other materials were functionalized by click chemistry 27 28 29 30 .

Based on the advantages of ATRP and click chemistry, the combination of these two methods has attracted attention to functionalize nanomaterials. Ranjan and Brittain 31 successfully grafted polymer chains onto SNPs through a combination of RAFT polymerization and click chemistry. Wang and co-workers 32 utilized ATRP and click chemistry to modify particles. Polystyrene brushes were firstly introduced on the surface via ATRP, and then the click reaction was generated on the surface of polymerized SNPs. The combination of surface-initiated ATRP and Huisgen [3 + 2] cycloaddition 33 34 was also developed as a versatile method for the functionalizations of SNPs. Most of the literatures reported the combination of ATRP and click chemistry to modify nanomaterials in tandem.

Herein, we described the immobilization of the initiator on the SNP surface via 1,3-dipolar cycloaddition, which can simultaneously initiate surface-induced ATRP of poly(ethylene glycol) methacrylate on the SNP surface. The general steps for the preparation of functionalized SNPs are shown in Figure 1. Based on the special characteristics of the ethylene glycol end group, we investigated the enrichment properties of Pb(II) in aqueous solution.

We performed TEM to characterize the morphologies of SiO₂ and modified SiO₂ nanoparticles (Figure 2). From Figure 2A, we can see clearly that the mean size of the particles is about 300 nm and the shape shows regulated sphericity. The average diameter of amino group-modified SNPs is about 355 nm (Figure 2B). Azide-functionalized SNPs and SNPs modified via click chemistry and ATRP simultaneously are presented in Figure 2C,D, respectively. The shape of the modified SNPs maintained the original spherical morphology.

The FT-IR spectra of SiO₂, SiO₂-NH₂, and SiO₂-N₃ are shown in Figure 3. Strong adsorption peaks at about 1,113 cm⁻¹ were observed for all the samples, indicating the existence of Si-O-Si stretching vibration of silanol groups. The -CH₂ groups were confirmed by C-H stretching at 2,926 cm⁻¹ and C-H scissoring vibration at 1,456 cm⁻¹. O-H stretching and bending vibrations could be detected at 1,629 and 3,600 to 3,300 cm⁻¹. In Figure 3b, the adsorption peak at about 3,420 cm⁻¹ was assigned to N-H stretching vibration, suggesting that the amino organic groups were introduced to the surface of SNPs. A peak at 1,734 cm⁻¹ observed in Figure 3c corresponded to the C = O stretching. Furthermore, the azide group-modified SNPs are an essential intermediate for the click chemistry reaction, which was demonstrated by a new peak at 2,110 cm⁻¹ in the spectrum of Figure 3c.

In order to further confirm the SNPs modified via click chemistry and ATRP simultaneously, the products were characterized by XPS analysis. The introduction of the azide group onto SNPs was demonstrated by the XPS spectrum in Figure 4, which presented the wide-scan (A) and N 1s core-level (B) spectra. The wide-scan spectra of the SiO₂-N₃ surface was dominated clearly by signals attributable to Si 2p, Si 2s, C 1s, N 1s, and O 1s with binding energies at about 101.2, 154, 284.2, 401.2, and 535.2 eV, respectively. The N 1s core-level spectra of azide-modified SNPs can be curve-fitted into three peak components with binding energies at about 399.8, 400.7, and 403 eV, attributable to the O = C-N, C-N-H, and C-N = N = N species, respectively.

Figure 5A showed the wide-scan spectrum of the SiO₂-PEGMA surface, which was dominated by signals attributable to Si 2p, Si 2s, C 1s, N 1s, O 1s, and Br with binding energies at about 101.8, 153, 284.2, 399.4, 531.4, and 67.4 eV, respectively. The peak of Br was observed at 67.4 eV, which could have originated from 3-bromopropyl propiolate. Figure 5B and C present the XPS spectra of C 1s and N 1s narrow scan on the SiO₂-PEGMA nanoparticles. The C 1s core-level spectrum can be curve-fitted with four peak components having binding energies at about 284.4, 285.7, 286.2, and 287.2 eV attributable to the C-H, C-N, C-O, and O = C-O species, respectively. The N 1s core-level spectrum can be curve-fitted into three peak components with binding energies at about 397.9, 399.8, and 400.7 eV which are ascribed to the C-N-N = N, O = C-N, and C-N-H species, respectively (Figure 5C).

The TGA curves of the functionalized nanoparticles are shown in Figure 6, from which an approximate amount of functional groups on the silica surface could be confirmed. The weight loss of APTES-treated SNPs was about 13%, which is attributed to the loss of the APTES layer. We can see obviously from curve b of Figure 6 that the weight loss of azide-modified nanoparticles is about 18.9% for the whole temperature range. It was calculated that there was about 5.9% mass loss for azide molecule-functionalized SNPs. After click chemistry and ATRP simultaneously, the first stage shows a weight loss of about 21% at 400°C while the second stage accounts for 8.7% of another weight loss (Figure 6c). These two stages of weight loss may indicate the degradation of PEGMA and that 1,2,3-triazole ring derivates molecules on the surface of SNPs, respectively.

The mechanism of metal ion adsorption from aqueous solution is related with physical and chemical processes, and the chemical binding reactions which happened involve metal ions and surface functional groups 36 ; adsorption will be affected as the functional groups increase. After the successful preparation of PEGMA-modified SNPs, functionalized nanoparticles were utilized to absorb lead ions in the aqueous solution. From the adsorption experiment, the concentration of supernatant fluid is presented in Figure 7. We can find that polymerized SNPs were adsorbed faster than uncoated nanoparticles and the adsorption capacity of polymerized SNPs was also higher than that of uncoated nanoparticles, particularly within 1 h. It may be due to the polymerization process which can immobilize more functional groups on the surface of SNPs to chelate metal ions.

After introducing amino groups onto silica surfaces, the ATRP initiator was successfully immobilized via click chemistry while poly(ethylene glycol) methacrylate chain-modified SNPs were obtained. This method takes advantage of click chemistry to optimize ATRP, which we have offered as a versatile pathway to functionalize SNPs. It will develop a series of desired functional groups and also extend this current different functional polymer chains which have potential application in the fabrication of SNP-based nanocomposites. Aside from the applications that require the enrichment of lead, because of the diversity and controllable ability of functional molecules on the surface of SNPs, efforts in our group are being made to extend current work to applications in selective removal of heavy metal ions such as Hg(II), Cu(II), As(III), and Cd(II).

The authors declare that they have no competing interests.

WL carried out the preparation and modification of silica nanoparticles. YX participated in the analysis and the testing of the nanocomposite. YZ and WM supervised this work, helped in the analysis and interpretation of data, and, together with SW, worked on the drafting and revisions of the manuscript. WM and YD conceived of the study and participated in its design and coordination. SW participated in the design of the study and provided analysis instruments. All authors read and approved the final manuscript.

WL and YX are MD students. YZ is an associate professor and WM is a professor and the Dean of the Faculty of Metallurgical and Energy Engineering. SW is a postdoctoral fellow in Kunming University of Science and Technology. YD is an academician of the Chinese Academy of Engineering.