Synthesis of novel ion-imprinted polymers by two different RAFT polymerization strategies for the removal of Cs(I) from aqueous solutions

Abstract

In this work, two novel Cs(I) ion-imprinted polymers (Cs(I)-IIP1 and Cs(I)-IIP2) have been prepared by surface imprinting technique based on support matrix of SBA-15. The strategy was carried out by introducing two different reversible addition–fragmentation chain transfer (RAFT) polymerization including using the free RAFT agent in solution and surface-anchored RAFT agent with appropriate initiation method, which is expected to generate a well-defined surface ion-imprinted polymer with excellent adsorption capacity. The materials were verified by FT-IR, SEM, TEM, nitrogen adsorption–desorption and TGA. Owing to the intrinsic advantages of surface-anchored RAFT polymerization, the resultant surface ion-imprinted polymer (Cs(I)-IIP2) exhibited more homogeneous and thinner polymer layer (15 nm) with excellent macrostructure in comparison to Cs(I)-IIP1 (75 nm) prepared by using the free RAFT agent in solution. Furthermore, the adsorption capacity of both polymers are compared, indicating that Cs(I)-IIP2 displays higher adsorption property and excellent selectivity for Cs(I).

---

1 Introduction

Molecularly imprinted polymers (MIPs) are a new kind of smart materials, which have remarkable recognition properties for their template.¹ In recent decades, ion imprinted polymers (IIPs) similar to MIPs has attracted considerable attention due to its potential application.^2,3 However, this widely used technique could create some common disadvantages, such as incomplete template removal, slow mass transfer, high cross-linking density, irregular and heterogeneous binding site distribution.^4,5 Therefore, surface ion imprinted technology has been drawing more and more attention to solve these problems by generating binding cavities onto or near the surface of the imprinted polymers.^6–8 The IIPs prepared by this method have been confirmed to possess much more effective recognition sites, complete removal of templates, faster mass transfer, better mechanism intensity and excellent reusability compared with the conventional imprinted materials.^9,10 In this case, the shape of the surface ion imprinting polymer will be determined by the support. Recently, the silicon-based mesoporous materials selected as the support for its layered structure with chemical stability and favorable physical properties have attracted extensive research interest, which also achieve excellent combination of silicon-based mesoporous materials and surface imprinting technique.

As a member of the mesoporous silicon family, SBA-15 is an attractive candidate to be used as a support for surface ion imprinting polymer due to its large surface area, highly ordered pore, thick pore walls with high hydrothermal stability and available surface functionalization.^11–13 Therefore, several researchers have developed the surface imprinted polymer based on mesoporous silica to remove metal ions,^14–16 which was generally prepared by traditional free radical polymerization. However, the conventional radical polymerization processes are usually rather difficult to control with regard to chain propagation and termination, which normally lead to polymer networks with heterogeneous structures.^17,18 The presence of the heterogeneity within the network structures of the imprinted polymer could significantly exacerbate some of the inherent drawbacks of the imprinted polymer, such as the broad binding site heterogeneity and the relatively low affinity.¹⁹

To address the above problem, reversible addition–fragmentation chain transfer (RAFT) polymerization as one of the most successful controlled/living free radical polymerization methods, have drawn much attention and could be used to synthesize polymers with well defined and complex architectures.²⁰ Furthermore, most importantly, the RAFT technique could remarkably improve the match in chain growth and chain relaxation rates because of its thermodynamically controlled process, which leads to the homogeneous networks of the polymer.^21,22 In addition, another significant advantage of RAFT polymerization is that no expensive agent and no undesired metal species are introduced during the polymerization process.²³ Among various approaches based on RAFT technique, there are two strategies to conduct RAFT polymerization based on the species immobilized on the substrate. The first is using the free RAFT agent in solution and the second is a surface-anchored RAFT agent with appropriate initiation method.^24,25 For the former, the procedures proposed were simple and easy to operate.²⁶ In the latter case, RAFT agent can either be attached via the R-group or Z-group onto the surface of the substrate, which provides an effective control over the thicknesses and densities of the polymer grafts.²⁷ However, to the best of our knowledge, there are few reports on the removal of metal elements using two of RAFT strategies to synthesize surface ion-imprinted polymer, especially based on mesoporous SBA-15. Therefore, the strategies we adopted to investigate the differences between the two RAFT methods are interesting, which enable to produce well-defined surface ion-imprinted polymer and excellent adsorption capacity with different RAFT polymerized routes.

With the development of industrial and military activities, radioactive cesium (¹³⁷Cs), one of the main products of nuclear fission, is considered the most abundant and hazardous due to its high solubility, high transferability and relatively long half-life.^28,29 Furthermore, it can be easily incorporated into terrestrial and aquatic organisms because of its similar chemical characteristics with potassium.³⁰ Consequently, removing the radioactive ¹³⁷Cs is important in the environmental pollution monitoring. Therefore, the development of adsorbents with high selectivity and high adsorption capacity for removal of cesium is highly desired.

In the present study, for the first time, two novel Cs(I)-IIP (Cs(I)-IIP1 and Cs(I)-IIP2) have been successfully synthesized by different RAFT polymerization strategies via surface imprinting technique based on mesoporous SBA-15. The obtained adsorbent (Cs(I)-IIP2) was applied to remove Cs(I) from aqueous solution by static and dynamic adsorption tests. The effects on the static adsorption behavior of the removal process parameters were studied and the equilibrium data were fitted by the Langmuir, Freundlich and Redlich–Peterson models. Furthermore, the removal of Cs(I) from aqueous solution by Cs(I)-IIP2 as adsorbent in the fixed-bed was investigated and the parameters on the column dynamics were also discussed and optimized. The breakthrough curves for the adsorption of Cs(I) were analyzed using Adams–Bohart and Thomas models.

2 Experimental

2.1. Materials and chemicals

Pluronic P123, triblock poly-(ethyleneoxide)-poly(propyleneoxide)-poly(ethyleneoxide), (EO₂₀PO₇₀EO₂₀, molecular weight 5800) (P123, Sigma, USA), tetraethyl orthosilicate (TEOS, Guoyao Chemical Reagents Corp, Shanghai, China), CsCl, N,N′-dicyclohexylcarbodiimide (DCC), methyl acrylic acid (MAA), (3-aminopropyl)triethoxysilane (APTES), 3-(methacryloxyl)propyltrimethoxysilane (MPS), ethylene glycol dimethacrylate (EGDMA), 2,2′-azobisisobutyronitrile (AIBN) and acetonitrile (ACN) were purchased from Aladdin Reagent Co. Ltd. (Shanghai, China). S-1-Dodecyl-S-(α,α′-dimethyl-α′′-acetic acid)trithiocarbonate (TTCA) was prepared following literature procedure.³¹

Cs(I) stock solution (1 g L⁻¹) was prepared from CsCl, using doubly distilled water to give a final volume. The working solutions of the metals were obtained by appropriate dilution of the stock solutions. The stock solutions (1 g L⁻¹) of interfering ions were prepared using their corresponding compounds. All the chemicals used were of analytical grade.

2.2. Apparatus and measurements

VISTA-MPX Inductive Coupled Plasma-Atomic Emission Spectrometer (ICP-AES, Varian, USA) was used to determine all metal ions in aqueous solution. Fourier transmission infrared spectra (FT-IR, 4000–400 cm⁻¹) in KBr were recorded on a NICOLET NEXUS 4700 FT-IR spectrometer (Nicolet, USA). NOVA2000 surface area and pore size analyzer (Quntachrome, USA) was used to measure the surface area and pore size distribution. Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods were employed to calculate the surface area and pore diameter. Scanning electron microscopy (SEM) was utilized for assessing the morphology of the obtained adsorbents by using JEOL JSM-6480 at 25 kV (JEOL, Japan). Transmission electron microscopy (TEM) analysis was performed by using JEM-2010HR at 200 kV (JEOL, Japan). Thermogravimetric analysis (TGA) was carried out using a DSC/DTA-TG (STA 449C Jupiter Netzsch, Germany).

2.3. Preparation of Cs(I)-IIP

2.3.1. Preparation and activation of SBA-15. The mesoporous silica SBA-15 was synthesized according to the procedure described by Zhao et al., using P123 as the structure-directing agent and TEOS as the silica source.³² It was then activated with HCl (3 mol L⁻¹) by refluxing for 24 h, followed by washing with doubly distilled water (DDW) to neutral pH. Finally, the product (SBA-15) was dried at 110 °C. The activated SBA-15 was thus obtained.

2.3.2. Preparation of SBA-15-MPS and SBA-15-NH₂. (1) To prepare SBA-15-MPS, 500 mg of activated SBA-15 and 10 mL MPS were dispersed in 50 mL dry toluene and stirred under nitrogen at 50 °C for 24 h. The mixture was centrifuged and the modified SBA-15 (SBA-15-MPS) was rinsed with toluene several times and dried under vacuum at room temperature.

(2) To prepare SBA-15-NH₂, 500 mg of activated SBA-15 was suspended in 40 mL of anhydrous toluene. Next, 0.94 mL of APTES was added drop wise over 15 min and the reaction mixture was refluxed for 12 h. The particles were then separated from the mixture by filtration. In order to remove the unreacted APTES, the product was washed with toluene and ethanol several times, respectively. Finally, the obtained amino-functionalized SBA-15 (SBA-15-NH₂) was dried under vacuum at 50 °C.

2.3.3. Preparation of SBA-15-TTCA. SBA-15-TTCA was synthesized as follows. 150 mg SBA-15-NH₂ was dispersed in 2 mL ACN under ultrasound. Then, 22 mL ACN and 0.43 mmol TTCA were added to the above mixture solution. The resulting system was sonicated for 10 min and ice-cooled before the addition of 0.474 mmol DCC in ice-cooled 8 mL ACN. The dispersion was mechanically stirred overnight at room temperature. Particles were collected using filtration and washed 4 times with ACN, then dried under vacuum and stored at 4 °C.

2.3.4. Preparation of Cs(I)-IIPs. (1) Preparation of Cs(I)-IIP1: in a 100 mL round-bottomed flask, 0.1 mmol of CsCl, 0.4 mmol of MAA, 2.0 mmol of EGDMA, 62.7 mg TTCA and 60 mg of SBA-15-MPS were dispersed into 5 mL of methanol and 20 mL of acetonitrile under heating and magnetic stirring. After sealing, mixing, and purging the mixture with nitrogen, 6 mg of AIBN (initiator) was added into this suspension. The resultant mixture was stirred in a thermostated oil bath at 60 °C for 10 h under nitrogen protection. After polymerization, the contents were recovered by centrifugation (named A). As a control, non-imprinted polymer (NIP1) was also prepared as a blank in parallel but omitting the template of CsCl.

(2) Preparation of Cs(I)-IIP2: in a 100 mL round-bottomed flask, 0.1 mmol of CsCl, 0.4 mmol of MAA, 2.0 mmol of EGDMA, and 75 mg of SBA-15-TTCA were dispersed into 5 mL of methanol and 20 mL of acetonitrile under heating and magnetic stirring. After sealing, mixing, and purging the mixture with nitrogen, 6 mg of AIBN (initiator) was added into this suspension. The resultant mixture was stirred in a thermostated oil bath at 60 °C for 10 h under nitrogen protection. After polymerization, the contents were recovered by centrifugation (named B). By comparison, NIP2 was also prepared as a blank in parallel but without the addition of CsCl.

(3) Purification of Cs(I)-IIPs: in order to remove unreacted ingredients, the imprinted particles of A and B were extensively washed with methanol–acetonitrile (1 : 4, v/v) for several times, accordingly. Then, the solids were treated with 2 mol L⁻¹ HCl to completely leach the coordinated and non-coordinated Cs(I). At last, the polymers were filtered with DDW to neutralization and dried overnight at 50 °C under vacuum. At last, Cs(I)-IIP1 and Cs(I)-IIP2 were obtained.

2.4. Static adsorption of metal ions

Adsorption of Cs(I) from aqueous solutions was studied in batch experiments. 0.01 g of the prepared adsorbents was added into 25 mL colorimetric tube containing certain amount of Cs(I) in the pH range of 2–9 (which was adjusted with 0.1 mol L⁻¹ HNO₃ and 0.1 mol L⁻¹ NH₃·H₂O) at 25 °C. Then, the mixture was shaken for 10 min and the adsorption time was maintained for 12 h. Afterwards, the mixture was centrifuged and the concentration of the residual Cs(I) was measured by ICP-AES. The absorption capacity Q_e (mg g⁻¹) at equilibrium was calculated as follows:where C₀ (mg L⁻¹) and C_e (mg L⁻¹) are concentrations of Cs(I) in the initial and equilibrium concentrations in aqueous phase, respectively. V (L) and W (g) are the volume of solution and the mass of the sorbent, respectively.

2.5. Dynamic adsorption and desorption experiments

The fixed-bed column adsorption tests were conducted in a glass column of 1 cm inner diameter and a length of 15 cm. An appropriate quantity of the adsorbent (0.1 g) of Cs(I)-IIP2 (NIP2) was packed into the column, achieving a bed height of 0.5 cm. Then, prior to the adsorption process for protonation and washing the composite particles, a stream of distillated water with pH 6 and flow rate of 1.5 mL min⁻¹ was applied to the column for about 2 h. Afterwards, the effects of adsorption flow rates, initial Cs(I) concentrations and temperatures on breakthrough curves were evaluated. The concentration of effluent samples were collected at defined time intervals and analyzed by ICP-AES. A parallel experiment was conducted without the adsorbent to determine metal ion loss during the adsorption.

The equilibrium absorption capacity for the Cs(I)-IIP2 in the fixed bed studies was calculated by the following equation:

where q_e(exp) (mg g⁻¹) is the equilibrium metal ions uptake of the column, C₀ (mg L⁻¹) is the initial metal ions concentration of feed solution, C_t (mg L⁻¹) is the metal ions concentration of the effluent, V_E (L) is the volume of the solution required to reach the exhaustion point (C_t/C₀ = 90%) and m (g) is the mass of adsorbent. The effluent concentration (C_t) from the column that reaches 10% of the influent concentration (C₀) is the breakthrough point and t_b (min) is the time to the breakthrough point.

V_E can be obtained from the following equation:

V_E (mL) = Q_te (3)

where t_e (min) is the time to the exhaustion point and Q (mL min⁻¹) is the flow rate.

The amount adsorbed in the fixed-bed column [q_total (mg)] was evaluated by the following equation:

q_total = q_e(exp)m (4)

The flow rate represents the empty bed contact time (EBCT) in the column, as described in the following equation:

ECBT (min) = bed volume (mL)/Q (5)

2.6. Selectivity study

In order to measure the selectivity of the imprinted polymer, the binary binding ability was investigated using Cs(I)-IIP2 and NIP2. 0.01 g of RAFT-IIP or NIP was added in 25 mL of 10.0 mg L⁻¹ binary metal (Cs(I)/Co(II), Cs(I)/Sr(II), Cs(I)/Ce(III), Cs(I)/Pb(II), Cs(I)/Ba(II) and Cs(I)/Cd(II)) mixed solutions at pH 6.0. Finally, the concentration of the metal ions was determined by ICP-AES. The distribution coefficient K_d (L g⁻¹), selectivity coefficient k, and the relative selectivity coefficient k′ were given as follows:where C_i and C_f represent the initial and equilibrated concentrations of the given metal ions in solution, respectively. K_d(Cs(I)) and K_d(M) represent the distribution coefficient of Cs(I) and M ions respectively. k_IIP2 and k_NIP2 represent the selectivity coefficient of Cs(I)-IIP2 and NIP2, respectively.

3 Results and discussion

3.1. Preparation

Scheme 1 illustrates the main steps of the two kinds of strategies involved in surface imprinting procedure with RAFT polymerization. As for Scheme 1(A), in the first step, vinyl groups were introduced to the surface of SBA-15 through chemical modification with MPS. Afterwards, the RAFT grafting polymerization of functional monomer (MAA) can be processed on the surface of SBA-15-MPS in methanol–acetonitrile solution using AIBN as initiator, TTCA as free chain transfer agent, EGDMA as crosslinker and Cs(I) as template, respectively. After polymerization, recognition sites located at the surface of the obtained Cs(I)-IIP1 were shaped after the removal of the template ions. Finally, a thin homogeneous Cs(I) ion-imprinted polymer layer (Cs(I)-IIP1) was successfully prepared by the chain propagation of the free RAFT chain transfer agent under controlled polymerization. Meanwhile, the strategy of preparing Cs(I)-IIP2 consisted two steps (Scheme 1(B)): the synthesis involved grafting of silane coupling agent APTES to the surface of SBA-15, followed by the condensation reaction of the chain transfer agent TTCA on the surface. Then, the corresponding amounts of SBA-15-supported chain transfer agent (SBA-15-TTCA), AIBN, MAA, Cs(I) and EGDMA were calculated and dissolved in the polymerization solvent. After polymerization, the chelated Cs(I) was removed by eluent. Finally, a thin homogeneous Cs(I) ion-imprinted polymer layer (Cs(I)-IIP2) was formed on the surface of SBA-15 through the chain propagation of SBA-15-supported chain transfer agent under controlled polymerization with surface imprinting technique.

Scheme 1 Schematic representation of synthesis procedure of Cs(I)-IIP1 (A) and Cs(I)-IIP2 (B).

3.2. Characterization

3.2.1. Characteristic of the FT-IR spectra. FT-IR spectral analysis was preliminarily performed to illustrate the different samples, which are shown in Fig. 1. As shown in the figure, all of the samples displayed the stretching and bending vibration of Si–O–Si bond around 1090 cm⁻¹, 806 cm⁻¹ and 462 cm⁻¹, suggesting that the structure of SBA-15 was well preserved after modification and polymerization. The peaks near 1630 cm⁻¹ were attributed to stretching vibration of the surface silanol groups of SBA-15. The observed features around 3440 cm⁻¹ indicate the stretching vibrations of –OH. After grafting of the amino functionality, SBA-15-NH₂ exhibited the characteristic peaks of N–H vibration band at 1573 cm⁻¹. In addition, the bands at 2925 and 688 cm⁻¹ were attributed to the symmetric vibrations of –CH₂ in the propyl chain of the APTES and the –CH₂ rocking vibration of Si–CH₂R, respectively.³³ The result indicated that APTES was successfully immobilized on SBA-15 surface. Compared with SBA-15-NH₂, the new peaks near 1700 cm⁻¹ and 1405 cm⁻¹ appeared after TTCA grafted onto SBA-15-NH₂, ascribing to the stretching band of C=O and C=S, respectively. Furthermore, the peaks at 1573 cm⁻¹ of N–H were also well maintained. All of the facts suggested that SBA-15-TTCA was successfully prepared. Additionally, as for SBA-15-MPS, the expected C=O bands in MPS (1719 cm⁻¹) appeared. Moreover, the vibration absorption bands of C–H bond near 2925 cm⁻¹ also appeared. The results indicated that MPS was successfully introduced onto SBA-15. From Fig. 1b, it can be clearly seen that after ion imprinting by RAFT polymerization the vibration of C=O (1732 cm⁻¹) and C–H (2975 cm⁻¹) became stronger after imprinting, which was attributed to the polymerization of MAA on the surface of SBA-15. The characteristic adsorption peaks at about 1150, 1255, 1400 and 1458 cm⁻¹ were ascribed to the stretching vibrations of C–O, the symmetric distortion vibrations of Si–C bonds, the stretching vibration of C=S and the scissor bending vibrations of the –CH₂ groups, respectively.^34,35 Moreover, the characteristic peak at 1573 cm⁻¹ of N–H bending vibration for Cs(I)-IIP2 also can be obvious after polymerization. All of the above results confirmed that Cs(I)-IIP1 and Cs(I)-IIP2 were successfully synthesized.

Fig. 1 FT-IR spectra of mesoporous supports (a) and imprinted polymers (b).

3.2.2. Characteristic of EDS. The existence of RAFT agent and imprinted polymer was further confirmed by the EDS analysis. In Fig. 2a, the signal of oxygen and silica appeared for the pure SBA-15. As shown in Fig. 2b, it was observed that the new signal of carbon and sulfur appeared due to the modification of TTCA onto the surface of SBA-15. Furthermore, the intensity of carbon and oxygen increased and the intensity of silica decreased owing to the imprinted polymer onto the surface of SBA-15 (Fig. 2c). As for Fig. 2d, the signal of sulfur was present for Cs(I)-IIP2, which also demonstrated the graft of imprinted polymer onto the surface of SBA-15. In addition, the emergence of Cs signal in Fig. 2c and d also provided additional evidence about the successful synthesis of the surface ion polymer by RAFT.

Fig. 2 EDS spectra of SBA-15 (a), SBA-15-TTCA (b), Cs(I)-IIP1 (c) and Cs(I)-IIP2 (d).

3.2.3. SEM and TEM characterization. SEM and TEM were used to capture the morphology and microstructure of the prepared materials. The results of SEM images are presented in Fig. 3a, c and e. It can be clearly observed that there were many short rod-like structures with relatively uniform size of approximately 1 μm of the pure SBA-15 (Fig. 3a). Additionally, as shown in Fig. 3c and e, the surface of SBA-15 became rough and the ordered morphology was destroyed to a certain degree, which was ascribed to a series of functionalization and ion-imprinted polymer thin layer formed, which also suggested that Cs(I)-IIP1 and Cs(I)-IIP2 were successfully prepared by RAFT polymerization. Meanwhile, the TEM images are displayed in Fig. 3b, d and f, which provide the evidence that all of the prepared materials exhibited well-ordered hexagonal mesoporous structures. The results indicated that the highly ordered mesoporous structures of the SBA-15 were still preserved after polymerization. Moreover, as can be seen from the figure, there were homogeneous thin polymer layers formed onto the surface of the pure SBA-15 for Cs(I)-IIP1 (∼75 nm) and Cs(I)-IIP2 (∼15 nm), indicating that the method of RAFT agent attached onto the surface of SBA-15 displayed more effective control over the thicknesses of the polymer grafts. Furthermore, compared with Cs(I)-IIP1, Cs(I)-IIP2 exhibited better regularity in macrostructure (SEM) and thinner polymer layer in microstructure (TEM). The facts implied that Cs(I)-IIP2 retained better uniformity of SBA-15 and the polymer layer was more homogeneous and thin. The phenomenon may be explained by the theory that a method for the immobilization of RAFT agents on surfaces were fixed on the solid surface so that propagating radicals have to get close to the substrate materials surface across the barrier of polymer layer in order to undergo the RAFT reaction during the course of polymerization, typically providing an effective control over the thickness and densities of the polymer grafts.³⁶ Moreover, the disadvantage of the approach for preparing Cs(I)-IIP1 was that for a well defined RAFT polymerization. All chains should be initiated at the same time via the free RAFT agent, which may make the polymer layer nonuniform.³⁷

Fig. 3 SEM images of SBA-15 (a), Cs(I)-IIP1 (c) and Cs(I)-IIP2 (e); TEM images of SBA-15 (b), Cs(I)-IIP1 (d) and Cs(I)-IIP2 (f).

3.2.4. Nitrogen adsorption–desorption characterization. The nitrogen adsorption–desorption isotherms of SBA-15, Cs(I)-IIP1 and Cs(I)-IIP2 are shown in Fig. 4. As can be seen from the figure, all samples exhibited typical type IV isotherms with clear hysteresis loops of H1 type associated with capillary condensation at P/P₀ from 0.4 to 0.8.³⁸ Despite the decrease in the adsorbed amount of N₂ after polymerization in the surface of SBA-15, the shape of the hysteresis loops of all isotherms remained the same. This also indicated that the ordered hexagonal structure of SBA-15 was well preserved after modification. Moreover, compared with SBA-15, the volumes of Cs(I)-IIP1 and Cs(I)-IIP2 decreased after polymerization, which suggested that pore sizes of SBA-15 had decreased due to polymer grafted on the channel. Additionally, Fig. 4 displays the pore size distribution curves (insert), which further indicated the decrease of the pore size and provided evidence about their uniform framework mesoporosity after polymerization.

Fig. 4 N₂ adsorption–desorption isotherms and pore size distribution (inset) of SBA-15, Cs(I)-IIP1 and Cs(I)-IIP2.

The physical properties of the samples are given in Table 1. As expected, BET surface area, pore volume and average pore diameter of Cs(I)-IIP1 and Cs(I)-IIP2 displayed a considerable decrease, which suggested the polymer was grafted onto the channel. Comparison of Cs(I)-IIP1 and Cs(I)-IIP2 showed that the surface area, pore volume and average pore diameter of the latter were little increased. This phenomenon confirmed that the former formed a thick polymer layer on the surface of SBA-15, which also corresponded to the characterization of TEM.

Table 1 Physical parameters of SBA-15, Cs(I)-IIP1 and Cs(I)-IIP2 measured by N₂ adsorption–desorption isotherms

Samples	BET surface area (m^2 g^−1)	Pore size (nm)	Pore volume (cm^3 g^−1)
SBA-15	456.07	4.92	0.70
Cs(I)-IIP2	245.61	3.83	0.46
Cs(I)-IIP1	221.97	3.80	0.41

---

3.2.5. Thermogravimetric analysis (TGA). TGA analysis was employed to further estimate the content of imprinting layer grafted on the surface of SBA-15 according to different thermal stability between mesoporous silica materials and polymers (Fig. 5). As seen from Fig. 5, all of the samples displayed a slight weight loss at temperatures from 25 °C to 100 °C, which was mainly due to the loss of absorbed water. As for Cs(I)-IIP2, it can be obviously seen that the weight loss increased rapidly in the temperature ranging from 200 °C to 800 °C because of the thermal decomposition of the polymer (approximately 32.11%). Moreover, for Cs(I)-IIP1, there were about 59.40% weight loss from 225 °C to 650 °C. All of the results demonstrated that the polymer was successfully grafted onto the surface of SBA-15 by RAFT.

Fig. 5 The thermogravimetric analysis curves of SBA-15, Cs(I)-IIP1 and Cs(I)-IIP2.

3.3. Adsorption experiments

3.3.1. Effect of pH. Among all parameters affecting the adsorption, pH is one of the most effective parameter for adsorption of metals on Cs(I)-IIP1 and Cs(I)-IIP2. Thus, the effect of variety of pH values on Cs(I) ions adsorption was investigated at the range of 2.0–9.0, which is depicted in Fig. 6. From the figure, it is obvious that the adsorption capacity increased significantly with the values of pH increasing from 2.0 to 6.0. This pH dependency could be attributed to these facts that in acidic pH, the carboxyl groups are protonated and could not coordinate to Cs(I) and the proton competitiveness for binding sites becomes weakened with increase in the pH. However, when pH increased from 6.0 to 9.0, the adsorption capacity of Cs(I) onto Cs(I)-IIP1 and Cs(I)-IIP2 decreased, which was consistent with the reported literature.^39,40 This decrease may be explained taking into consideration that the pH adjustment for alkaline solutions was made with HNO₃ and NH₃·H₂O, which could introduce NH₄⁺ as a competing ion into the solution with increase of the value of pH. The fact might decrease Cs(I) adsorption when present in excess amount. Furthermore, comparing the adsorption capacity of the two adsorbents, Cs(I)-IIP2 exhibited a better adsorption property than that of Cs(I)-IIP1. The phenomenon might be explained by the fact that Cs(I)-IIP2 displayed more homogeneous and thinner polymer layer than that of Cs(I)-IIP1 due to the different RAFT strategies, which may make the binding sites of Cs(I)-IIP2 exhibit better advantages in accessibility to the target species and removal of templates. The facts can cause Cs(I)-IIP2 show more excellent adsorption capacity and relatively larger surface area than that of Cs(I)-IIP1. Therefore, pH 6.0 and Cs(I)-IIP2 were selected as the optimum for further experiments.

Fig. 6 Effect of pH on adsorption of Cs(I) adsorption on Cs(I)-IIP1 and Cs(I)-IIP2.

3.3.2. Adsorption isotherms. The adsorption isotherms were used to assess the binding properties in the batch rebinding method, which was a key for understanding the adsorption mechanism. Therefore, the adsorption properties of Cs(I)-IIP2 and NIP2 were evaluated by adsorption isotherms in the batch experiments, which is shown in Fig. 7. With increase of initial Cs(I) concentration, first there was rapid increase in adsorption capacity, then the adsorption becomes slow and finally reaches saturation. It is observed that the adsorption capacity of Cs(I) on Cs(I)-IIP2 was near two times that of NIP2, suggesting that the former has stronger affinity for Cs(I) and has more adsorption sites than the latter. This can be explained that there are many specific binding sites on the surface of Cs(I)-IIP2. Additionally, an increase in temperature resulted in an increase in the adsorption capacity of Cs(I) demonstrating that the process was endothermic, which was in accordance with the reported results.^41,42 Furthermore, the enthalpy of Cs(I) onto Cs(I)-IIP2 and NIP2 were 32.749 and 19.873 KJ mol⁻¹, respectively, which also indicated the adsorption process was endothermic. Moreover, in order to further investigate the adsorption isotherms, Langmuir,⁴³ Freundlich⁴⁴ and Redlich–Peterson⁴⁵ models were applied to describe the adsorption isotherms. The models can be expressed as follows.

Freundlich: Q_e = K_FC_e^1/n (11)

where Q_e (mg g⁻¹) is the equilibrium adsorption capacity, C_e (mg L⁻¹) is the equilibrium concentration of Cs(I) in solution, K_D is the equilibrium constant (K_D = Q_e/C_e), Q_m (mg g⁻¹) is the maximum adsorption capacity, K_L (L mg⁻¹) is the affinity constant. K_F (mg g⁻¹) is an indicative constant for the adsorption capacity and 1/n is an empirical parameter related to the adsorption intensity, which is affected by the heterogeneity of the material. K_R (L g⁻¹) and α_R (L mg⁻¹) are Redlich–Peterson isotherm constant, β Redlich–Peterson isotherm exponent, which lies between 0 and 1, has two limiting behaviors: Langmuir form for β = 1 and Henry's law form for β = 0.

Fig. 7 Adsorption isotherms of Cs(I) onto Cs(I)-IIP2 (A) and NIP2 (B).

The results of the three models' nonlinear fitting are also shown in Fig. 7 and their adsorption constants calculated from the corresponding isotherms with the correlation coefficients are presented in Table 2. As seen from the table, Redlich–Peterson model exhibited the best fitting in the case. Redlich–Peterson equation incorporates the advantages of both Langmuir and Freundlich equations, which can be applied either in homogeneous or heterogeneous system.⁴⁶ In the present study, the β values were close to 1, which means that the isotherms conform to Langmuir model better than Freundlich model and the adsorption of Cs(I) onto Cs(I)-IIP2 and NIP2 were well fitted by Langmuir model with R² in the range of 0.96–0.99, whereas the Freundlich model with lower correlation coefficients was not suitable. Furthermore, as can be seen from Table 2, the maximum adsorption amounts for Cs(I) onto Cs(I)-IIP2 was found to be 54.54 mg g⁻¹ at 25 °C, which was nearly two times larger than that of NIP2 (29.62 mg g⁻¹). Moreover, it was also clearly observed that the ratio for the other sets of temperatures was similar to 25 °C. The results concluded that Cs(I)-IIP2 prepared by RAFT was a thin homogeneous layer and the adsorption capacity was satisfactory.

Table 2 The Langmuir, Freundlich and Redlich–Peterson isotherm parameters for Cs(I) adsorption onto Cs(I)-IIP2 and NIP2

Adsorbent	Tem. (°C)	Langmuir model	Freundlich model					Redlich–Peterson model
		Qm (mg g^−1)	KL (mL mg^−1)	R^2	KF (mg g^−1)	1/n	R^2	KR (L g^−1)	αR (L mg^−1)	β	R^2
Cs(I)-IIP2	25	54.54	0.0300	0.9876	7.675	0.3304	0.9493	2.192	0.0696	0.9072	0.9896
	35	74.64	0.0355	0.9780	10.16	0.3353	0.9667	4.137	0.1563	0.8253	0.9895
	45	97.13	0.0390	0.9694	16.06	0.3098	0.9697	10.67	0.3414	0.8054	0.9927
NIP2	25	29.62	0.0279	0.9851	4.101	0.3297	0.9345	0.8797	0.0345	0.9746	0.9868
	35	37.54	0.0293	0.9824	5.281	0.3289	0.9487	1.500	0.0431	0.9680	0.9851
	45	44.67	0.0313	0.9902	6.845	0.3156	0.9398	1.673	0.0531	0.9411	0.9905

---

3.3.3. Effect of flow rates on breakthrough curve. The effect of flow rate on Cs(I)-IIP2 of Cs(I) in the fixed-bed with a bed depth of 0.5 cm and initial Cs(I) concentration of 10 mg L⁻¹ at 298 K was investigated. The flow rate was changed in the range of 1.0–3.0 mL min⁻¹. The breakthrough curves are shown in Fig. 8A. It is clear from the figure that the breakthrough occurred significantly faster as the flow rate increases. The results may be explained on the basis of mass transfer fundamentals.⁴⁷ At low flow rate, the metal ions have more time to be in contact with adsorbent than at higher flow rate, which resulted in a better adsorption capacity and higher removal efficiency from solution. At a higher flow rate, there was not sufficient time for diffusion of Cs(I) into the pores of Cs(I)-IIP2 through intra-particle diffusion.⁴⁸ Moreover, as seen from Table 3, the exhaust time decreased from 66 min to 33 min and EBCT decreased from 0.3925 min to 0.1308 min with the flow rate increasing and the adsorption capacity also decreased with increase in flow rate. Therefore, 1.0 mL min⁻¹ was selected as the optimal flow rate in subsequent experiments.

Fig. 8 Experimental and predicted breakthrough curves of Cs(I) adsorption by Cs(I)-IIP2 fixed-bed columns predicted by the Thomas and Adams–Bohart model: (A) different flow rates; (B) different initial Cs(I) concentrations; (C) different temperatures.

Table 3 Parameters in fixed-bed column for Cs(I) adsorption by Cs(I)-IIP2

C0 (mg L^−1)	m (g)	Q (mL min^−1)	EBCT (min)	T (K)	tb (min)	te (min)	qtotal (mg)	qe(exp) (mg g^−1)	VE (mL)
10.0	0.1	1.0	0.3925	298	39	66	0.6442	6.442	66
10.0	0.1	2.0	0.1963	298	21	48	0.4638	4.638	96
10.0	0.1	3.0	0.1308	298	8	33	0.3140	3.140	99
15.0	0.1	1.0	0.3925	298	22	48	0.7051	7.051	48
20.0	0.1	1.0	0.3925	298	7	31	0.7596	7.596	31
10.0	0.1	1.0	0.3925	308	20	51	0.5052	5.052	51
10.0	0.1	1.0	0.3925	318	5	36	0.3533	3.533	36

---

3.3.4. Effect of initial ion concentration on breakthrough curve. The effect of initial ion concentration on the breakthrough time and the shape of breakthrough curves were investigated at 10.0, 15.0 and 20.0 mg L⁻¹, respectively. The adsorption breakthrough curves obtained at different initial ion concentrations, with 0.5 cm of bed depth and a flow rate of 1.0 mL min⁻¹ at 298 K, are shown in Fig. 8B. As can be seen from the plots and Table 3, the exhaust time decreased from 66 min to 31 min and the breakthrough time decreased from 39 min to 7 min with the initial ion concentration increasing. In addition, sharper breakthrough curves were obtained as influent concentration increased. The phenomenon can be explained by the fact that more adsorption sites were being covered as Cs(I) concentration increases.⁴⁹ Moreover, it also might be explained by the theory that lowers concentration gradient caused slow transport due to decreased diffusion coefficient.⁵⁰ However, as presented in Table 3, the adsorption capacity increased from 6.442 mg g⁻¹ to 7.596 mg g⁻¹ with increasing the initial ion concentration, which may be ascribed to the high initial ion concentration providing the bigger driving force for the adsorption process.⁵¹

3.3.5. Effect of temperature on breakthrough curve. In order to investigate the effect of temperature on the breakthrough curve, the adsorbate solution having Cs(I) concentration of 10.0 mg L⁻¹, bed depth 0.5 cm and flow rate of 1.0 mL min⁻¹ was passed through the adsorption column by varying the temperature. Fig. 8C and Table 3 present the performance of breakthrough curves at temperatures of 298, 308 and 318 K. When the temperature was increased from 298 K to 318 K, the exhaust time and the breakthrough time decreased from 66 min to 36 min and 39 min to 5 min, respectively. This might be for the reason that the higher operating temperature was conducive to diffuse Cs(I) faster into the adsorbent, which favored a low breakthrough time and fast saturation.

3.4. Dynamic models

It is important to understand the breakthrough performance of the fixed-bed system for the successful design and optimization of the column for the adsorption process in packed columns. Several mathematical models can be used to describe fixed-bed adsorption. In this study, the Thomas model and Adams–Bohart model were employed to describe the adsorption performance.

3.4.1. Thomas model fit. The Thomas model is one of the most widely used models to describe the performance theory of the sorption process in a fixed-bed column. This model assumes that Langmuir kinetics of adsorption–desorption and no axial dispersion are derived with the assumption that the rate driving force obeys second-order reversible reaction kinetics.^52,53 The nonlinear expression by Thomas for an adsorption column is given as follows:where C₀ (mg L⁻¹) is the influent metal ion concentration, C_t (mg L⁻¹) is the effluent concentration at time t, K_T (mL min⁻¹ mg⁻¹) is the Thomas model constant, q₀ (mg g⁻¹) is the adsorption capacity of the adsorbent, Q (mL min⁻¹) is the flow rate and m (g) is the sorbent mass.

The fit of the Thomas model allows estimating the capacity of the column (q₀) and the Thomas rate constant (K_T). The predicted curves for experimental data using Thomas model are shown in Fig. 8A–C. It can be seen that there is a good agreement between the experimental and the predicted breakthrough curves for Cs(I) onto Cs(I)-IIP2. Moreover, Cs(I)-IIP2 column adsorption parameters derived from the nonlinear Thomas are listed in Table 4. As seen from the table, the values of q₀ estimated by Thomas model were close to q_e(exp) calculated from the experiments. In addition, the values of R² ranged from 0.9974 to 0.9997. The well fit of the experimental data on to the Thomas model implied that the external and internal diffusion would not be the limiting step.⁵⁴

Table 4 Parameters of Adams–Bohart and Thomas model analyzed for Cs(I) adsorption by Cs(I)-IIP2 in a fixed-bed column

Parameter	Thomas model			Adams–Bohart model
	KT (mL min^−1 mg^−1)	q0 (mg g^−1)	R^2	KAB (mL min^−1 mg^−1)	N0 (mg mL^−1)	R^2
Flow rate (mL min^−1)
1.0	14.84	6.691	0.9997	3.031	1.278	0.8445
2.0	17.06	5.511	0.9989	2.873	3.209	0.7873
3.0	17.24	3.234	0.9974	2.797	3.586	0.7618
Influent concentration (mg L^−1)
10.0	14.84	6.691	0.9997	3.031	1.278	0.8445
15.0	11.28	7.244	0.9997	2.373	1.397	0.8349
20.0	9.008	8.092	0.9975	1.620	1.394	0.7860

---

From Table 4, it can also be seen that with the flow rate increased, the value of q₀ decreased but the value of K_T increased. The behaviors can be ascribed to the unavailability of reaction sites and the decrease of the mass-transport resistance.⁵⁵

As the initial ion concentration increased, the values of q₀ increased and K_T decreased, respectively. It was attributed to the driving force for adsorption in the concentration difference.⁵⁶ With temperature increasing, the value of q₀ increased while the value of K_T decreased, suggesting that the adsorption was favored by lowering the temperature. Therefore, the lower flow rate, higher initial ion concentration and lower temperature would increase the adsorption of Cs(I) onto Cs(I)-IIP2 column.

3.4.2. Adams–Bohart fit. Adams–Bohart model was based on the surface reaction theory, and it is assumed that the equilibrium is not instantaneous and the adsorption rate is proportional to both the residual capacity used to describe the initial part of the breakthrough curve.^57,58 The expression is the following:where K_AB (mL min⁻¹ mg⁻¹) is the kinetic constant, U₀ (cm min⁻¹) is the linear velocity calculated by dividing the flow rate by the column section area, Z (cm) is the bed depth of column and N₀ (mg mL⁻¹) is the saturation concentration.

The Adams–Bohart adsorption model was applied to experimental data for the description of the initial part of the breakthrough curves, the regression results are presented in Fig. 8A–C and Table 4. It can be clearly observed that the data fits were relatively good with the present model (R²). But, it was poorer than that of Thomas model. On the other hand, as seen from Fig. 8A–C, there was a good agreement between the predicted values and the experimental data especially for the initial portion of the breakthrough curve. The results indicated that the Adams–Bohart model was valid for the relative concentration region up to the C_t/C₀ range of 0.1. Whereas large discrepancies were found between the experimental and the predicted curves above the specified range. Therefore, only the initial part of the breakthrough curve can be well fitted by Adams–Bohart because of its limitation in the range of conditions used.⁵⁹

3.5. Selectivity study

Competitive adsorption of Cs(I)/Ba(II), Cs(I)/Co(II), Cs(I)/Cd(II), Cs(I)/Sr(II), Cs(I)/Pb(II) and Cs(I)/Ce(III) from their binary mixture was investigated by using self-prepared Cs(I)-IIP2 and NIP2 as adsorbents and ICP-AES as detection technique. As can be seen from Table 5, Cs(I)-IIP2 exhibited an excellent adsorption selectivity for Cs(I) in presence of competitive metal ions. However, the NIP2 did not display the obvious difference in the rebinding capacities of the other ions. The results might be explained by the theory that the cavities created after removal of the template for Cs(I)-IIP2 were complementary to the imprinted ion in shape and coordination geometries. Although some ions had similar size with Cs(I), Cs(I)-IIP2 still showed higher selectivity for Cs(I) than NIP2 and also further illustrated that the imprinting process was effective. Therefore, Cs(I)-IIP2 could be applied to selectively remove Cs(I) from the solution.

Table 5 Competitive sorption of different ions by Cs(I)-IIP2 and NIP2 sorbent

Metal type	Cs(I)-IIP2			NIP2			k′
	Qe (mg g^−1)	Kd (L g^−1)	kIIP	Qe (mg g^−1)	Kd (L g^−1)	kNIP
Cs(I)	12.08	2.337	—	4.64	0.5697	—	—
Co(II)	4.76	0.5879	3.975	3.96	0.4705	1.211	3.282
Sr(II)	3.87	0.4579	5.104	4.34	0.5252	1.085	4.704
Ce(III)	3.98	0.4734	4.937	5.92	0.7757	0.7344	6.722
Pb(II)	4.75	0.5864	3.985	7.84	1.142	0.4989	7.988
Ba(II)	2.43	0.2692	8.681	5.18	0.6534	0.8719	9.956
Cd(II)	5.24	0.6630	3.525	6.48	0.8747	0.6513	5.412

---

3.6. Dynamic desorption

The regeneration of a bed is a key factor for improving wastewater process economics and evaluating its feasibility for practical use. To investigate the regeneration and reuse capability of exhausted Cs(I)-IIP2, the desorption process was carried out using 2 mol L⁻¹ HCl in a fixed-bed column system. The dynamic desorption curve is shown in Fig. 9. It can be seen clearly that the Cs(I) concentration of the eluent reached a maximum value at V 17.0 mL and then decreased with the volume of eluent. By calculating, more than 92% of Cs(I) was desorbed from the adsorbents at 40.0 mL, which suggested that the surface imprinted material Cs(I)-IIP2 prepared by RAFT present an excellent eluting performance. Additionally, in order to display the reusability of the adsorbents, the regeneration experiment was repeated three times by using the same sorbent. The results implied that Cs(I)-IIP2 could be successfully reused at least three times and displayed an excellent regeneration ability.

Fig. 9 Dynamic desorption curve of Cs(I)-IIP2.

3.7. Comparison with other type of adsorbents

The adsorption behaviors of Cs(I)-IIP2 for the removal of Cs(I), such as the Q_e, and selectivity, were compared with those of the other type of adsorbents. The results are displayed in Table 6. When it comes to the common functionalized raw materials, higher Q_e and good adsorption selectivity made Cs(I)-IIP2 outstanding.^{30,41,59–63} Moreover, compared with other Cs(I)-IIP,⁴⁰ Cs(I)-IIP2 prepared in this study also presented excellent Q_e and adsorption selectivity.

Table 6 Comparison with other type of adsorbents for Cs(I)

Adsorbents	Qe (mg g^−1)	Selectivity	Reference
Copper hexacyanoferrate–PAN composite	25.52	—	30
Surface whisker-supported ion-imprinted polymer	32.90	Pb, Co, Ce, Sr, Zn, Ba, Ni	40
Nickel hexacyanoferrate–walnut shell	4.94	—	41
Copper ferrocyanide functionalized mesoporous silica	17.10	—	60
Oxidized multiwall carbon nanotubes	12.75	—	61
Transition metal modified akadama clay	16.19	—	62
Florisil impregnated with trihexyl(tetradecyl)phosphonium chloride	3.086	—	63
Cs(I)-IIP2	54.54	Pb, Co, Ce, Sr, Cd, Ba	This work

---

4 Conclusions

In this work, a novel Cs(I) ion-imprinted polymer (Cs(I)-IIP1 and Cs(I)-IIP2) were successfully prepared by different reversible addition–fragmentation chain transfer (RAFT) polymerization routes with surface imprinting technique based on support matrix of SBA-15. The formation of these hybrid materials were verified by FT-IR spectroscopy, SEM, TEM, nitrogen adsorption–desorption and TGA. Compared with the two adsorbents, Cs(I)-IIP2 showed more excellent morphology and adsorption capacity. The adsorption isotherm was well fitted by Langmuir model and Cs(I)-IIP2 had higher selectivity and nearly two times larger Langmuir adsorption capacity (54.54 mg g⁻¹) than that of NIP2 at room temperature. Furthermore, Cs(I)-IIP2 also proved to be an effective and promising adsorbent of Cs(I) from aqueous solution by fixed-bed column study. The adsorption of Cs(I) on Cs(I)-IIP2 bed was found to depend on the flow rate, influent concentration and temperature. Moreover, Cs(I)-IIP2 could be easily regenerated by desorption process with dynamic desorption ratio of 92%, which is in conformity with the green chemistry principle. All the results suggest that Cs(I)-IIP2 could be used as an excellent adsorbent for efficient removal of Cs(I) from aqueous solution.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (nos 21207051), Ph.D. Programs Foundation of Ministry of Education of China (no. 20123227120015), China Postdoctoral Science Foundation funded project (no. 2012M511220), Special Financial Grant from the China Postdoctoral Science Foundation (2014T70488), Society Development Fund of Zhenjiang (nos SH2012021, SH2013110), Programs of Senior Talent Foundation of Jiangsu University (no. 11JDG125) and Programs of innovation practical training of students of Jiangsu University (nos 201410299153W, 201410299154W).

References

1. S. A. Piletsky, S. Alcock and A. P. Turner, Trends Biotechnol., 2001, 19, 9.
2. C. H. Hu, J. Deng, Y. B. Zhao, L. S. Xia, K. H. Huang, S. Q. Ju and N. Xiao, Food Chem., 2014, 158, 366.
3. Y. K. Tsoi, Y. M. Ho and K. S. Y. Leung, Talanta, 2012, 89, 162.
4. C. Branger, W. Meouche and A. Margaillan, React. Funct. Polym., 2013, 73, 859.
5. L. Ozcan and Y. Sahin, Sens. Actuators, B, 2007, 127, 362.
6. C. J. Tan, S. Wangrangsimakul, R. B. Bai and Y. W. Tong, Chem. Mater., 2008, 20, 118.
7. B. Guo, F. Deng, Y. Zhao, X. B. Luo, S. L. Luo and C. Au, Appl. Surf. Sci., 2014, 292, 438.
8. S. F. Xu, L. X. Chen, Y. F. Guan and H. Z. Lu, J. Hazard. Mater., 2012, 237, 347.
9. F. F. Duan, C. Q. Chen, G. Z. Wang, Y. Z. Yang, X. G. Liu and Y. Qin, RSC Adv., 2014, 4, 1469.
10. H. T. Fan, J. X. Liu, H. Yao, Z. G. Zhang, F. Yan and W. X. Li, Ind. Eng. Chem. Res., 2014, 53, 369.
11. J. E. Lofgreen and G. A. Ozin, Chem. Soc. Rev., 2014, 43, 911.
12. E. Magner, Chem. Soc. Rev., 2013, 42, 6213.
13. J. Morell and M. Fröba, Angew. Chem., Int. Ed., 2006, 45, 3216.
14. C. C. Kang, W. M. Li, L. Tan, H. Li, C. H. Wei and Y. W. Tang, J. Mater. Chem. A, 2013, 1, 7147.
15. H. Ebrahimzadeh, E. Moazzen, M. M. Amini and O. Sadeghi, Microchim. Acta, 2013, 180, 445.
16. Y. K. Lu and X. P. Yan, Anal. Chem., 2004, 76, 453.
17. A. R. Wang and S. P. Zhu, Polym. Eng. Sci., 2005, 45, 720.
18. Y. Li, X. Li, C. K. Dong, Y. Q. Li, P. F. Jin and J. Y. Qi, Biosens. Bioelectron., 2009, 25, 306.
19. H. Q. Zhang, Eur. Polym. J., 2013, 49, 579.
20. A. W. York, S. E. Kirkland and C. L. Cormick, Adv. Drug Delivery Rev., 2008, 60, 1018.
21. S. Perrier, C. Barner-Kowollik, J. F. Quinn, P. Vana and T. P. Davis, Macromolecules, 2002, 35, 8300.
22. X. G. Hu, Y. N. Fan, Y. Zhang, G. M. Dai, Q. L. Cai, Y. J. Cao and C. J. Guo, Anal. Chim. Acta, 2012, 731, 40.
23. G. Moad, M. Chen, M. H€aussler, A. Postma, E. Rizzardo and S. H. Thang, Polym. Chem., 2011, 2, 492.
24. Y. Li and B. C. Benicewicz, Macromolecules, 2008, 41, 7986.
25. D. P. Zhou, E. Mastan and S. P. Zhu, Macromol. Theory Simul., 2012, 21, 602.
26. S. F. Xu, J. H. Li and L. X. Chen, J. Mater. Chem., 2011, 21, 4346.
27. M. H. Stenzel, L. Zhang and S. W. S. Huck, Macromol. Rapid Commun., 2006, 27, 1121.
28. X. Liu, G. R. Chen, D. J. Lee, T. Kawamoto, H. Tanaka, M. L. Chen and Y. K. Luo, Bioresour. Technol., 2014, 160, 142.
29. A. Ararem, O. Bouras and F. Arbaoui, Chem. Eng. J., 2011, 172, 230.
30. A. Nilchia, R. Saberi, M. Moradib, H. Azizpour and R. Zarghami, Chem. Eng. J., 2011, 172, 572.
31. J. T. Lai, D. Filla and R. Shea, Macromolecules, 2002, 35, 6754.
32. D. Y. Zhao, J. L. Feng, Q. S. Huo, N. Melosh, G. H. Fredrickson and B. F. Chmelka, Science, 1998, 279, 548.
33. L. L. Lou, S. Jiang, K. Yu, Z. C. Gu, R. N. Ji, Y. L. Dong and S. X. Liu, Microporous Mesoporous Mater., 2011, 142, 214.
34. S. Y. Bai, H. Zhang, J. H. Sun, J. Han and Y. Y. Guo, J. Appl. Polym. Sci., 2014, 131, 40403.
35. Y. L. Luo, F. Xu, Q. S. Feng, Y. S. Chen and C. Ma, J. Biomed. Mater., 2010, 92, 243.
36. K. Ohno, Y. Ma, Y. Huang, C. Mori, Y. Yahata, Y. Tsujii, T. Maschmeyer, J. Moraes and S. Perrier, Macromolecules, 2011, 44, 8944.
37. M. D. Rowe-Konpacki and S. G. Boyes, Macromolecules, 2007, 40, 879.
38. G. Leofantia, M. Padovan, G. Tozzola and B. Venturelli, Catal. Today, 1998, 41, 207.
39. M. R. Awual, T. Yaita, T. Taguchi, H. Shiwaku, S. Suzuki and Y. Okamoto, J. Hazard. Mater., 2014, 278, 227.
40. Z. L. Zhang, X. H. Xu and Y. S. Yan, Desalination, 2010, 263, 97.
41. D. Ding, Y. N. Zhao, S. Yang, W. Shi, Z. Zhang, Z. F. Lei and Y. Yang, Water Res., 2013, 47, 2563.
42. H. Faghihian, M. Moayeda, A. Firooz and M. Iravani, C. R. Chimie, 2014, 17, 108.
43. Q. Manzoor, R. Nadeem, M. Iqbal, R. Saeed and T. M. Ansari, Bioresour. Technol., 2013, 132, 446.
44. X. D. Dong, C. Wang, H. Li, M. Wu, S. H. Liao, D. Zhang and B. Pan, Bioresour. Technol., 2014, 160, 123.
45. M. Xu, P. Hadi, G. H. Chen and G. McKay, J. Hazard. Mater., 2014, 273, 118.
46. K. Y. Foo and B. H. Hameed, Chem. Eng. J., 2010, 156, 2.
47. B. Y. Chen, C. Y. Chen, W. Q. Guo, H. W. Chang, W.-M. Chen, D. J. Lee, C. C. Huang, N. Q. Ren and J. S. Chang, Bioresour. Technol., 2014, 160, 175.
48. E. Sabio, F. Zamora, J. Gañan, C. M. González-García and J. F. González, Water Res., 2006, 40, 3053.
49. R. P. Han, Y. Wang, X. Zhao, Y. F. Wang, F. L. Xie, J. M. Cheng and M. S. Tang, Desalination, 2009, 245, 284.
50. M. Jaina, V. K. Garg and K. Kadirvelu, Bioresour. Technol., 2013, 129, 242.
51. W. H. Zou, L. Zhao and L. Zhu, J. Radioanal. Nucl. Chem., 2013, 295, 717.
52. Z. Aksu and F. Gönen, Process Biochem., 2004, 39, 599.
53. H. C. Thomas, J. Am. Chem. Soc., 1944, 66, 1664.
54. S. S. Baral, N. Dasb, T. S. Ramulub, S. K. Sahoob, S. N. Dasb and G. R. Chaudhury, J. Hazard. Mater., 2009, 161, 1427.
55. Z. Saadi, R. Saadi and R. Fazaeli, J. Nanostruct. Chem., 2013, 3, 2.
56. A. A. Ahmad and B. H. Hameed, J. Hazard. Mater., 2010, 175, 298.
57. A. Ghosh, S. Chakrabarti and U. C. Ghosh, Chem. Eng. J., 2014, 248, 18.
58. E. I. Unuabonaha, M. I. El-Khaiary, B. I. Olu-Owolabi and K. O. Adebowale, Chem. Eng. Res. Des., 2012, 90, 1105.
59. S. Kundu and A. Gupta, J. Colloid Interface Sci., 2005, 290, 52.
60. T. Sangvanich, V. Sukwarotwat, R. J. Wiacek, R. M. Grudzien, G. E. Fryxell, R. S. Addleman, C. Timchalk and W. Yantasee, J. Hazard. Mater., 2010, 182, 225.
61. R. Yavari, Y. D. Huang and S. J. Ahmadi, J. Radioanal. Nucl. Chem., 2011, 287, 393.
62. D. H. Ding, Z. F. Lei and Y. B. Yang, Chem. Eng. J., 2014, 236, 17.
63. L. Lupa, A. Negrea, M. Ciopec and P. Negrea, Molecules, 2013, 18, 12845.

---