当前位置:首页 >> 化学 >>

Grafting of Ion-Imprinted Polymers on the Surface of Silica Gel Particles

Anal. Chem. 2007, 79, 7116-7123

Grafting of Ion-Imprinted Polymers on the Surface of Silica Gel Particles through Covalently Surface-Bound Initiators: A Selective Sorbent for Uranyl Ion
Mojtaba Shamsipur,*,? Javad Fasihi,? and Khadijeh Ashtari?

Department of Chemistry, Razi University, Kermanshah, Iran, and Department of Chemistry, Tarbiat Modarres University, Tehran, Iran

A new ion imprinted polymer coated silica gel sorbent has been prepared using the radical “grafting from” polymerization method through surface-bound azo initiators for selective uranyl uptake. The introduction of azo initiator onto the silica surface was achieved by the reaction of surface amino groups with 4,4′-azobis(4-cyanopentanoic acid chloride). The grafting step was then carried out in a stirred solution of initiator-modified silica particles in the presence of uranyl ion and functional and cross-linking monomers. The prepared sorbent was characterized using FT-IR spectroscopy, scanning electron microscopy (SEM), elemental analysis (EA), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and BET adsorption isotherm analysis. The influence of the uranyl concentration, pH, and flow rate of solution on the grafted polymer affinity has been investigated. Maximum uptake of uranyl ion was observed at a pH 3.0. The rebinding behavior of the sorbent has been successfully described by the Langmuir-Freundlich isotherm. The dynamic column capacity of sorbent and enrichment factor for uranyl ion were 52.9 ( 3.4 ?mol g-1 and 52, respectively. It was found that imprinting results in increased affinity of the sorbent toward uranyl ion over strong competitor metal ions such as Fe(III) and Th(IV). The sorbent was repeatedly used and regenerated for 3 months without any significant decrease in polymer binding affinities. Finally the sorbent was applied to the preconcentration and determination of uranyl ion in real water samples.
During the past decade, molecular imprinting technology has become a well-established analytical tool.1-3 Imprinted polymers were originally prepared by interactive preorganization of a functionalized monomer with a template, followed by polymerization with or without an excess of cross-linking agent. Removal
* Towhomcorrespondenceshouldbeaddressed.E-mail: mshamsipur@yahoo.com. Fax: +98-21-66908030. ? Razi University. ? Tarbiat Modarres University. (1) Andersson, L. I.; Nicolla, I. A. J. Chromatogr., B 2004, 804, 1. (2) Alexander, C.; Andersson, H. S.; Andersson, L. I.; Ansell, R. J.; Krisch, N.; Nicholls, I. A.; O’Mahony, J.; Whitcombe, M. J. J. Mol. Recognit. 2006, 19, 106. (3) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495.

of the template yields an insoluble polymeric network material with cavities that are complementary in size, shape, and functional group orientation to those of the template.4 Under appropriate conditions, these cavities bind the template, or its analogues, efficiently and selectively. Many researchers have reported bulk polymerization as a successful method for the imprinting process.2 The resulting molecularly imprinted polymers (MIPs) have been encountered with various limitations such as nonuniform distribution of particle size,2,6-8 high consumption of template and material,6,8,16 thermal instability, and swelling and shrinkage effects.7 Recently, these problems have been overcome by grafting techniques where the MIPs layers are coated onto substrates with known surface morphologies.8,9 The MIPs have been prepared as a grafted coating on silica particles,8-12 silica capillary columns,15,16 alumina oxide membrane,17 and polymeric supports.18-22 There are two methods for covalent bonding of a polymer chain onto inorganic support surfaces called “grafting from” and “grafting
(4) Cui, A.; Singh, A.; Kaplan, D. L. Biomacromolecules 2002, 3, 1353. (5) Rao, T. P.; Kala, R.; Daniel, S. Anal. Chim. Acta 2006, 578, 105. (6) Bru ¨ ggemann, O.; Haupt, K.; Ye, L.; Yilmaz, E.; Mosbach, K. J. Chromatogr., A 2000, 889, 15. (7) Prasad, B. B.; Banerjee, S. React. Funct. Polym. 2003, 55, 159. (8) Sulitzky, C.; Ruckert, B.; Hall, A. J.; Lanza, F.; Unger, K.; Sellergren, B. Macromolecules 2002, 35, 79. (9) Sellergren, B.; Sulitzky, C.; Ruckert, B. U.S. Patent 6,759,488 B1, 2004. (10) Tamayo, F. G.; Titirici, M. M.; Esteban, A. M.; Sellergren, B. Anal. Chim. Acta 2005, 542, 38. (11) Quaglia, M.; Lorenzi, E. D.; Sulitzky, C.; Massoloni, G.; Sellergren, B. Analyst 2001, 126, 1495. (12) Quaglia, M.; Lorenzi, E. D.; Sulitzky, C.; Caccialanza, G.; Sellergren, B. Electrophoresis 2003, 24, 952. (13) Prucker, O.; Ru ¨ he, J. Macromolecules 1998, 31, 592. (14) Meyer, T.; Spange, S.; Hesse, S.; Ja ¨ger, C.; Bellmann, C. Macromol. Chem. Phys. 2003, 204, 725. (15) Schweitz, L. Anal. Chem. 2002, 74, 1192. (16) Ou, J.; Li, X.; Feng, Sh.; Dong, J.; Dong, X.; Kong, L.; Ye, M.; Zou, H. Anal. Chem. 2007, 79, 639. (17) Wang, H.-J.; Zhou, W.-H.; Yin, X.-F.; Zhuang, Z.-X.; Yang, H.-H.; Wang, X.R. J. Am. Chem. Soc. 2006, 128, 15954. (18) Ulbricht, M. J. Chromatogr., B 2004, 804, 113. (19) Sergeyeva, T. A.; Matuschewski, H.; Piletsky, S. A.; Bendig, J.; Scheler, U.; Ulbricht, M. J. Chromatogr., A 2001, 907, 89. (20) Piletsky, S. A.; Matuschewski, H.; Scheler, U.; Wilpert, A.; Piltska, E. V.; Thiele, T. A.; Ulbricht, M. Macromolecules 2000, 33, 3092. (21) Piletsky, S. A.; Piletska, E. V.; Chen, B.; Karim, K.; Weston, D.; Barrett, G.; Lowe, P.; Turner, A. P. F. Anal. Chem. 2000, 72, 4381. (22) Bossi, A.; Piletsky, S. A.; Piletska, E. V.; Righetti, P. G.; Turner, A. P. F. Anal. Chem. 2001, 73, 5281. 10.1021/ac070968e CCC: $37.00 ? 2007 American Chemical Society Published on Web 08/21/2007

7116 Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

to” techniques. Often in the “grafting to” technique, the living polymer chains are reacted with suitable termination groups on the surface.13,14 Due to kinetic hindrance, formation of polymer layers on the surface by this technique is intrinsically limited, and this leads to low graft density.8,13 To obtain MIP-grafted inorganic particles (such as silica), it is preferred to initiate the graft polymerization from an initiating group immobilized onto the particle surface, the technique being called “grafting from” polymerization. Here, in the course of polymerization, the grafted polymer chains are propagated from the surface radicals formed by the thermal or photochemical decomposition of azo groups introduced onto the surface. In this method, the generated free radicals remain confined to the surface or in the vicinity of the surface.8 In this view, the polymer propagates mainly on the surface and chain growth in solution is limited by diffusion of free radicals to the bulk.13,23 This would lead to a high polymer grafting yield in comparison with the “grafting to” technique. The use of ion-imprinted polymer (IIP) sorbents to separate and preconcentrate metal ions is one of the new developments in the area.2,5,24-27 In order to prepare effective IIPs, it is necessary to immobilize the ligands with proper functionalities inside the rigid polymeric matrix. In this method, vinyl ligand-metal ion (host-guest) complex is copolymerized with a cross-linking agent, which provides spatially regulated sites in the polymer matrix. The extraction of uranium from dissolved sources such as seawater and leach liquors is desirable for subsequent uses, for instance in the nuclear fuel cycle.24 Meanwhile, because of high chemical toxicities and radiological hazards of uranium and its salts, preconcentration and determination of this element in environmental samples (such as drinking and wastewaters) is an important task. Ion-imprinted polymers have been used to separate uranyl and other rare earth cations.24,28-36 The unique shape of the uranyl ion may be expected to lead to much greater selectivity for the uranyl ion by ion imprinting than other metal ions.27 Due to the strong tendency of carboxylic acid functional groups toward uranyl ion, recently some carboxylic acid functional monomers have been successfully used as polymerizable ligands in the preparation of uranyl-selective IIPs.24,28-30 We were interested in the development and evaluation of the grafting of IIPs onto silica particle surfaces using the “grafting
(23) Gaspirrini, F.; Misiti, D.; Rompietti, R.; Villani, C. J. Chromatogr., A 2005, 1064, 25. (24) Port, S. N.; Joyce, M. J.; Walton, P. H.; Saunders, G. D. U.S. Patent 6,372,872 B1, 2002. (25) Uezu, K.; Nakamura, H.; Kanno, J.; Sugo, T.; Goto, M.; Nakashio, F. Macromolecules 1997, 30, 3888. (26) Fang, G.-Z.; Tan, J.; Yan, X.-P. Anal. Chem. 2005, 77, 1734. (27) Zheng, H.; Zhang, D.; Wang, W. Y.; Fan, Y. Q.; Li, J.; Han, H. P. Microchim. Acta 2007, 157, 7. (28) Bae, S. Y.; Southard, G. L.; Murray, G. M. Anal. Chim. Acta 1999, 397, 173. (29) Saunders, G. D.; Foxon, S. P.; Walton, P. H.; Joyce, M. J.; Port, S. N. Chem. Commun. 2000, 273. (30) Kimaro, A.; Kelly, L. A.; Murray, G. M. Chem. Commun. 2001, 1282. (31) Dai, Sh.; Shin, Y. S.; Barnes, C. E.; Toth, L. M. Chem. Mater. 1997, 9, 2521-2525. (32) Metilda, P.; Gladis, J. M.; Rao, T. P. Anal. Chim. Acta 2004, 512, 63. (33) Bu ¨ yu ¨ ktiryaki, S.; Say, R.; Ero ¨z, A.; Birlik, E.; Denizli, A. Talanta 2005, 67, 640. (34) Preetha, C. R.; Gladis, J. M.; Rao, T. P. Environ. Sci. Technol. 2006, 40, 3070. (35) Vigneau, O.; Pinel, C.; Lemaire, M. Anal. Chim. Acta 2001, 435, 75. (36) Chauvin, A. S.; Bu ¨ nzli, J. C. G.; Bochud, F.; Scopelliti, R.; Froidevaux, P. Chem. Eur. J. 2006, 12, 6852.

from” technique. To the best of our knowledge, there is no previous literature report on the use of this technique in combination with an ion-imprinting process. In this study, we report on the first photograft copolymerization of methacrylic acid-uranyl ion complex with ethyleneglycoldimethacrylate onto an aminopropyl silica support through covalently surface-bound azo initiator for the preparation of a new uranyl-selective ion imprinted polymer coated sorbent. The characterization of the new synthesized sorbent and its rebinding abilities after removal of uranyl ion are described and discussed. EXPERIMENTAL SECTION Apparatus. A Varian Vista Pro (CCD simultaneous) model inductively coupled plasma (ICP) (Varian, U.S.A.) was used for the determination of all metal contents. The operational conditions for the ICP were in accordance with the manufacture’s instructions. A 713 model digital pH meter (Metrohm, Germany) was used for pH adjustments. The FT-IR (ATR) spectra (4000-500 cm-1) were recorded on a Vertex 70 spectrometer (Bruker, Germany). Thermal analyses were performed using an STD1500 model thermal analyzer (Rehometric Science, England). The grafting yields were obtained from elemental analysis (EA) using a Vario El III CHN analyzer (Elementar, Germany). Nitrogen sorption measurements were performed using a NOVA 2200 highspeed gas sorption analyzer (Quantachrome, U.S.A.). Reagents. Aminopropyl silica gel (Fluka, particle size 42-60 ?m, 0.9 mmol g-1 amino groups) with a specific surface area of 239 m2 g-1 (obtained from nitrogen adsorption measurements via the BET technique) was dried overnight at 105 °C. Methacrylic acid (MAA) and ethyleneglycoldimethacrylate (EDMA) were supplied by Merck. 4,4′-Azobis(4-cyanopentanoic acid) (ACPA) was obtained from Aldrich. Analytical grade uranium nitrate hexahydrate, lanthanum nitrate hexahydrate, yttrium nitrate hexahydrate, iron nitrate nonahydrate, zirconium oxide chloride octahydrate, cerium nitrate hexahydrate, samarium oxide, and thorium nitrate pentahydrate were purchased from Merck and used as received. Solutions of metal ions were prepared in doubly distilled water. All solvents used were at least reagent grade from Merck. Synthesis of 4,4′-Azobis(4-cyanopentanoic Acid Chloride) (ACPC).14 A slurry of 5 g of ACPA in 50 mL of anhydrous dichloromethane (DCM) was added in small portions to a suspension of 40 g of PCl5 in 100 mL of DCM. During the process, the mixture was stirred vigorously under an argon atmosphere and the temperature was kept under 5 °C. After overnight stirring at room temperature, the unreacted PCl5 was removed from the mixture. The resulting solution was mixed with 300 mL of precooled dried hexane, and the mixture was stored at 0 °C for 48 h. ACPC was crystallized as white needles. The product was filtered, washed with precooled hexane, and dried under vacuum at room temperature. Yield: 4.3 g (76%); mp 79-80 °C; FT-IR: 1401, 1436, 1785, 2243 cm-1. Introduction of Azo Groups onto Aminopropyl Silica Particle Surface. The introduction of azo groups onto the silica surface was achieved by the reaction of surface amino groups with ACPC in the presence of pyridine as already described by Ueda
Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Chart 1. Schematic Representation of the Syntheses of the Azo Initiator Modified Aminopropyl Silica Gel (Azo-APSG) (Top) and Ion Imprinted Polymer Coated Silica (Bottom)

et al.37 The initiator-modified particles were stored in the dark below 0 °C. Quantitative elemental analysis: C, 9.57%; N, 3.46%; δazo ) 425 ?mol g-1. Differential scanning calorimetry (DSC): Tdec ) 120 °C. FT-IR: 2896, 2943, 2840, 2246, 1652, 1541, 1049, 794 cm-1. “Grafting from” Polymerization. The radical graft polymerization initiated by surface-bound azo groups can be carried out in a stirred suspension of modified particles (Chart 1). For preparation of IIP-grafted silica, the stirred solution should contain the functional host monomer, metal ion, and cross-linker monomers. A typical procedure is as follows. The prepared initiatormodified silica particles (0.1 g) were added to 6 mL of DCM/ MeOH (80/20 v/v) solution containing 0.2 mmol of uranyl nitrate hexahydrate and 0.38 mL (4 mmol) of EDMA in the absence or presence of 0.051 mL (0.6 mmol) of MAA. The temperature was fixed at 10 °C by immersing the reaction vessel in a thermostated water bath. After purging with Ar and sealing, the suspension was stirred and irradiated with a UV light source (1000 W mercury lamp; Beltron GmbH, Germany), for selected periods of time. In order to remove the ungrafted polymers, after centrifugation, the grafted particles were worked up by successive washing with DCM until no free polymer precipitated when the supernatant solution was added dropwise to an excess of methanol. The entrapped uranyl ions were then leached using concentrated hydrochloric acid as follows. The resulting materials were treated with (10 mL) aliquots of 2 M HCl in batch experiments. The suspensions were stirred for about 5 h, and after centrifugation, the uranyl contents of the supernatant solutions were determined by ICP. This step was carried out for several times until the supernatant solution was free from uranyl ions, and then the sorbent was repeatedly washed with double-distilled water until neutral pH. Then the particles were washed with methanol and dried under vacuum. The control nonimprinted polymer materials were
(37) Ueda, J.; Sato, S.; Tsunokawa, A.; Yamauchi, T.; Tsubokawa, N. Eur. Polym. J. 2005, 41, 193.

similarly prepared using an identical procedure but without the addition of UO2(NO3)2?6H2O. The suitability of polar organic solvents such as DCM has been already reported by Sulitzky et al.8 However, because of the fast reaction kinetics in DCM, some particle agglomerations and mixture gelation were observed after an irradiation time of 15 min. Thus, in this work, we found that the addition of 20% (by volume) of MeOH to DCM results in controllable kinetics of polymerization without any particle agglomeration and gelation at even longer irradiation times (30 min). Sorption Study. The sorption studies of the uranyl ion by the IIP-grafted silica particles were carried out in the batch and column procedures as follows. In the batch procedure, 100 mg of IIP material was added to a 10 mL buffered solution of 5 ?g mL-1 uranyl ion. The pH of the solution was adjusted with nitric acid for a pH of 1.5 and with sodium acetate/nitric acid for the pH range of 2-4. The suspension was stirred for preselected periods of time using a magnetic stirrer. After centrifugation, the supernatant solution was removed and the uranyl ions preconcentrated onto IIP particles were eluted for 5 min using 5 mL of 1 M HCl and filtered through a Millipore filter (0.45 ?m). Uranium contents of the resulting solutions were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES). The percentage of metal ion adsorbed on the sorbent was determined by comparing its concentrations before, Ci (?g mL-1), and after extraction, Cf (?g mL-1), as

% uptake ) (Ci - Cf) × 100/Ci


The distribution ratio (mL g-1) of uranyl ion between the IIP particles and aqueous solution was also determined by the following equation:

Kd ) (Ci - Cf)V/Cfm


where V is the volume of the initial solution and m is the mass of


Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

IIP-grafted materials. Selectivity coefficients for uranyl ion relative to foreign ions are defined as
2 KUO22+/Mn+ ) (KUO /KM d d ) 2+ n+


2 where KUO and KM are distribution ratios of uranyl and d d foreign ion, respectively. In the column procedure, a cylindrically shaped PTEF microcolumn (30 mm length and 3 mm i.d.) was packed with 100 mg of the imprinted polymer grafted silica gel sorbent. The solutions containing uranyl ion were pumped through the column at various flow rates. The adsorbed uranyl ions were then eluted with dilute HCl. The concentrations of metal ion in the effluent and eluent solutions were determined by ICP.



RESULTS AND DISCUSSION The most established procedure for attachment of MIPs to a silica surface via a covalent approach is radical “grafting from” polymerization of template/functional monomer complexes in the presence of a cross-linker monomer.8,9,11 In an ideal case, the “grafting from” process is expected to be not only surface-initiated but also surface-confined so that the propagation of the polymer chain does not occur in solution. However, due to escaping of the radicals from the surface, it has been reported that the polymerization is no longer surface-confined.23 Nonetheless, the polymer grafting on the surface is associated with favored yields, and nongrafted polymers can be easily removed from the final product by successive washing. It is preferred to initiate the grafting polymerization from azo initiator groups introduced onto the particle surface. In fact, the decomposition kinetics of azo initiators is not affected by the tethering to the surface of silica.13 In this work, the initiator-modified silica particles were prepared in a single step by the treatment of surface amino groups with ACPC in DCM. The “grafting from” polymerization step was then carried out in a stirred solution containing suspended initiator-modified silica particles based on the method introduced by Sellergren et al.9 In preliminary investigations, the grafting step was carried out in a MeOH/DCM (20:80 v/v) solution containing uranyl nitrate, EDMA, and suspended azo-activated particles in the absence of MAA. EDMA was selected because its ester groups increase the interaction of cross-linker with water so that the diffusion of water into the polymeric network can be easily achieved.24 However, after removal of uranyl ion from the matrix, the sorption studies indicated that the resulting polymer-coated silica particles possess very poor uptake of uranyl ions in contact with uranyl-containing solutions at various pHs. In order to overcome this problem and prepare more effective IIPs, the grafting step was done in the presence of MAA as a functional ligand to create a polymerizable uranyl-MAA complex. The cavity-containing polymers were then produced using UO22+ as a template while monomeric complexing ligands were oriented around and later polymerized in the presence of EDMA as a cross-linking agent. After removal of uranyl ion from the matrix, the sorption studies were carried out in batch and column procedures. Characterization Studies. IR Spectra. To ascertain the presence of a polymer layer on the silica support, the FT-IR (ATR) spectra of bare aminopropyl silica gel (APSG) and polymer-grafted silica sorbent were recorded. The spectra revealed two intense bands at 1053 and 801 cm-1 due to the Si-O-Si and Si-OC

vibrations, respectively. The presence of adsorption water was reflected by a broad band between 3000 and 3700 cm-1. The IR spectrum of polymer-grafted silica showed new bands at 1730 cm-1 (carbonyl group) and 2800-2999 cm-1 (C-H stretching), which arise from grafted layers MAA/EDMA copolymers on APSG. SEM. The surface characterization of imprinted polymer grafted silica particles was carried out using scanning electron microscopy, SEM (Figure 1). The modified particles appear as consisting of micrometer-sized particles with no evidence for interparticle aggregation. Comparison of the SEM images of IIPgrafted particles and untreated particles, taken under identical experimental conditions, revealed that surface morphology is affected by the polymer grafting (Figure 1, parts A and B). It is clearly seen that a polymer layer with pore structure has been formed on the silica surface. Meanwhile, the energy-dispersive X-ray (EDX) spectrum of uranyl imprinted polymer grafted silica (Figure 1C) shows a qualitative evidence for the entrapment of uranyl ion in the grafted layers on the silica surface. Thermal and Quantitative Elemental Analyses. The degree of surface functionalization of APSG with ACPC can be determined by EA. The amount of bound initiator, expressed as grafting density (δazo) and area density (Γazo) of the immobilized azo initiator, can be calculated from the increase in carbon and nitrogen content upon initiator immobilization on APSG according to the following equations:10

δazo )

(% C) , MC(100 - (% C)MW/MC) δazo ) (% N) (4) MN(100 - (% N)MW/MN) (5)

Γazo ) δazo/S

where MC and MN are the weight of carbon and nitrogen per mole of azo initiator, MW is the molecular weight of the immobilized initiator, and S is the specific surface area of the APSG (239 m2 g-1). The % C and % N can be obtained from comparison of EA results of APSG before and after azo immobilization. Another way for quantification of azo initiators on the surface of silica is to follow thermal decomposition of attached azo molecules by DSC,13,14,23 as this is a strongly exothermic process. The thermal decomposition of pure ACPA shows a sharp exotherm for decomposition at 120 °C. Noteworthy, it is assumed that the thermal behavior of the azo compounds is not strongly affected by their binding to the surface.13 However, the DSC thermogram of the immobilized azo shows a broad curve at the same temperature as the pure azo compound, most probably due to microenvironmental differences.23 The amount of immobilized azo initiator on the surface can be evaluated by monitoring the signal of the exothermic decomposition in the DSC curve of azoimmobilized silica samples.8,13,23 On the basis of the molar enthalpy of decomposition (?Hdec) of pure ACPA (152.3 kJ mol-1),8 the amount of immobilized azo (δ in ?mol g-1) can be calculated from the ratio of the integral of the DSC signal (Idec) and ?Hdec. The results obtained from the integral of the DSC signal and quantitative elemental analyses are shown in Table 1. From these values, it can be calculated that about 1.0-1.3 initiator molecules/nm2 are immobilized, which is equivalent to the average distance of 0.9-1 nm between two anchoring sites. Meanwhile, the DSC thermogram of polymer-grafted silica showed no exothermic peak at the desired temperature. This is
Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Figure 1. Scanning electron micrograph of bare APSG (A) and uranyl imprinted polymer grafted silica gel (B) surfaces (magnifications: 2 × 104) and EDX spectrum of uranyl imprinted polymer grafted silica gel (C). Table 1. Grafting and Area Densities (δazo and Γazo) and Average Anchor Distance (?) for Immobilized Azo Initiator on Silica applied method DSC (?H)a EA, C content ) 9.57 (5.64)%b EA, N content ) 3.46 (2.09)%b

Table 2. Amounts of Grafted Polymer on Silica Support Determined by Elemental and Thermogravimetric Analyses
EA TGA BET measurements

δazo (?mol g-1) 516 437 414

Γazo (?mol m-2) 2.11 1.83 1.73

? (nm)a 0.89 0.95 0.98

reaction C content grafting grafting S VP dP time (min) (%)a yield (%)b yield (%)a (m2 g-1)c (cm3 g-1)d (?)e 30 20.7 (13.8) 30.9 32.7 289 0.29 41

Idec ) 78.5 J g-1; δazo ) Idec/?Hdec. b Average from three batches, the values in parentheses show the increase in C and N content upon initiator immobilization on APSG.

a Averaged from three batches, the values in parenthesis show the increase in carbon content upon “grafting from” step on azo-APSG. b Results calculated assuming a stoichiometric incorporation of monomers and entrapment of initiator fragments in grafted layers on APSG (grams of total grafted organics from two steps of initiator immobilization and polymer grafting/grams of APSG). c Specific surface area. d Total pore volume. e Average pore diameter.

a good indication for the almost complete decomposition of azo groups during grafting from the polymerization step.23,37 The graft densities of polymer layers on the silica surface were obtained by thermogravimetric analysis (TGA) and EA. The grafting yields from EA were calculated based on the observed increase in carbon content during both steps of initiator immobilization and grafting from polymerization. The TGA experiments were carried out as follows: the bare APSG and the IIP-grafted silica were dried at 65 °C overnight and were cooled in a desiccator to room temperature. The samples were heated by using of thermogravimetric analyzer from 30 to 800 °C at a rate of 10 °C min-1 and kept for 30 min at the final temperature. The amount of grafted polymer was then determined from the difference between the weight loss of the grafted and ungrafted silicas. Table 2 presents the results of thermogravimetric and elemental analyses as well as the degree of “grafting from” MAA/EDMA copolymers on the silica surface. Rebinding Isotherm. Batch rebinding study is a well-known key method for the complete characterization of imprinted polymers.38
(38) Umpleby, R. J., II; Baxter, S. C.; Rampey, A. M.; Rushton, G. T.; Chen, Y.; Shimizu, K. D. J. Chromatogr., B 2004, 804, 141.

A binding isotherm measures the binding efficiency of a polymer over a range of analyte concentrations and is usually plotted as the concentration of analyte bound to a polymer (B) versus the concentration of free analyte remaining in solution (F). However, noncovalently imprinted polymers contain a heterogeneous mixture of cavities with a wide array of binding affinities. It has been demonstrated that the binding behavior of these materials can be accurately modeled by the Langmuir-Freundlich isotherm (LFI) (eq 6).38-40 The binding parameters can be calculated directly using the LFI fitting coefficients:


m N tKm 0F m 1 + Km 0F


The fitting parameters of Nt and K0 are the total number of binding sites (?mol g-1) and median binding affinity (mM-1),
(39) Umpleby, R. J., II; Baxter, S. C.; Chen, Y.; Shah, R. N.; Shimizu, K. D. Anal. Chem. 2001, 73, 4584. (40) Daniel, S.; Babu, P. E. J.; Rao, T. P. Talanta 2005, 65, 441.


Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Figure 2. (A) Experimental isotherm (dot) and LFI fit (line) for IIP-grafted silica and (B) affinity distribution (AD) calculated using eq 7.

respectively, and m is the heterogeneity index, which varies from 0 to 1. In several batch experiments, 200 mg portions of the uranylimprinted sorbent were equilibrated with varying concentrations of uranyl ion (initial concentration range from 20 to 475 ?g mL-1) at a pH of 3.0. After 24 h, the equilibrium concentrations of uranyl (F, mM) in solution were determined by ICP and the amounts of adsorbed uranyl on sorbent (B, ?mol g-1) were calculated from these values. The experimental isotherm data (F and B) were successfully fitted to the LFI (Figure 2) in order to evaluate the Nt, K0, and m values. The resulted fitting coefficients at the desired concentration window were Nt ) 65.459 ?mol g-1, K0 ) 0.439 mM-1, m ) 0.873 (R2 ) 0.9998). As seen, the m value, which is constant regardless of the concentration window, is close to 1. This is an evidence for a relatively homogeneous nature of binding sites on the imprinted sorbent. An affinity distribution (AD) is a plot of the number of binding sites (N) as a function of association constant (K). Recently, Umpleby and co-workers38,39 have proposed a unimodal distribution for binding affinities in noncovalently imprinted polymers and derived a mathematical expression for the AD for an LFI which requires only the experimentally derived LF fitting parameters, Nt, K0, and m. The corresponding mathematical expression derived by Umpleby and co-workers is

Figure 3. Effect of pH on sorption of uranyl ion on imprinted polymer grafted silica gel IIP (A) and control nonimprinted silica gel CP (B) sorbents.

N(K) ) 64.022K-0.873


0.237 + 2.459K-0.873 + 0.056K-1.747 (1 + 0.487K-0.873)4



The resulting AD gives a quantitative measure of the number of binding sites (N(K)) with respect to binding affinity and, at the same time, a measure of the breadth of the heterogeneity. Figure 2 shows the calculated ADs for uranyl IIP-grafted silica sorbent based on eq 7 within the limits of log Kmin and log Kmax, which were set by the concentration ranges of the experimental binding isotherm (Kmin ) 1/Fmax and Kmax ) 1/Fmin). As seen, the sorbent displays an AD with a well-known asymptotically decaying shape, which is characteristic of noncovalently synthesized imprinted polymers.39 Sorption Study. Effect of pH. Uranyl ion is now firmly established as the pentagonal bipyramidal [UO2(H2O)5]2+.41 It is
(41) Cotton, S. Lanthanide and Actinide Chemistry; John Wiley & Sons: London, 2006.

the stable form of uranium in aqueous solutions in the absence of complexing agents for the pH range of 0 to about 4. At higher pH values, the uranyl ion may hydrolyze to form species such as UO2OH+, (UO2)2(OH)22+, and (UO2)3(OH)5+.28 In our work, the effect of pH on the uranyl uptake was evaluated in batch experiments, by equilibrating 100 mg of uranyl imprinted polymer grafted silica sorbent with 10 mL of the buffer solutions containing 50 ?g of uranyl ion at various pHs. The same loading conditions were applied to both the imprinted silica and control silica. The uranyl uptakes were studied in the pH range of 1.5-4.0. As expected, due to the presence of carboxylic acid groups, the uptake of uranyl ion varied significantly with the pH of solution. Because of the weak deprotonation of carboxylate groups at pHs lower than 1.5, no uranyl ion was extracted onto the sorbent. The maximum uptake of uranyl ion was observed at a pH of 3.0. At higher pHs, the percentage of preconcentration decreased, most probably due to competition between acetate ions in solution with immobilized methacrylates for uranyl ion. As seen from Figure 3, in all cases, the imprinted polymers revealed much higher affinity for the uranyl than the control nonimprinted polymers, which is a good indication of the imprinting process and its influence on the formation of the template-specific binding sites. Thus, the quantitative extraction of uranium(VI) ion can be achieved with the IIP-grafted silica particles at weakly acidic solutions. Batch Adsorption Test. The uptake of uranyl ions from solution by imprinted material is fairly rapid. In a typical uptake kinetics test, 100 mg of the sorbent was added to 10 mL of 5 ?g mL-1 uranyl solution at a pH of 3.0. The suspension was stirred for different periods of time (from 2 to 30 min) using a magnetic stirrer. The results indicated that more than 95% uranyl uptake
Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Table 3. Distribution Ratio (Kd) and Selectivity Coefficient (K) Values of Ion Imprinted Polymer Grafted Silica and Control Material and Calculated Relative Selectivity Coefficients (k′)a imprinted sorbent cation UO2 Th4+ Sm3+ Ce3+ Pb2+ La3+ ZrO2+ Y3+ Cu2+ Ni2+ Fe3+ Mn2+ Co2+, Zn2+, Cd2+
a 2+

control sorbent Kd 35.5 30.0 2.6 4.0 1.1 5.2 7.0 7.9 1.2 1.7 18.9 0.9 <0.5 Kcontrol 1.2 13.4 8.9 32.3 6.8 5.1 4.5 29.6 20.9 1.9 39.4 >102 k′ 402.5 >746.3 168.3 92.7 348.5 221.2 185.5 116.1 153.2 234.2 102.2

Kd 4950.0 10.2 <0.5 3.3 1.6 2.1 4.4 5.9 1.4 1.5 11.1 1.2 <0.5

Kimprinted 483 >104 1498 2994 2370 1128 835 3437 3201 445 4026 >104

Figure 4. Elution profile of uranyl ion on a microcolumn containing 100 mg of IIP-grafted silica sorbent (analytical line: 385.957 nm).

was achieved within 5 min. This rapid adsorption equilibrium time is most probably because of the easy diffusion of water onto the polymeric network, due to the presence of cross-linker ester groups24 and also high complexation rate between the uranyl ions and predefined cavities on the imprinted sorbent. Effect of Flow Rate. Rapid extraction kinetics is required due to the high elution rate used in column extractions. It is clear that, at higher flow rates, the contact time of uranyl ions with the column material is shorter. The effect of flow rate on the adsorption of uranyl ions in a packed microcolumn (100 mg of imprinted sorbent) was investigated at optimum conditions by changing the flow rate in the range of 0.2-4.0 mL min-1. It should be mentioned that faster flow rates could not be investigated due to the back pressure generated by the microcolumn. No significant change was observed in uranyl uptake with increasing the effluent flow rate from 0.2 to 4 mL min-1. The extraction of uranyl is clearly quantitative, and the elution volume can be kept a very small value by using a dilute HCl solution (i.e., 1 mL of 1 M HCl). Elution Profile. For column operations, the elution profiles must show complete elution from the column with a very small volume of the eluent solution. Figure 4 displays the results obtained for chromatographic experiments carried out on a microcolumn containing 100 mg of IIP-grafted silica sorbent, which was loaded with a 25 ng mL-1 uranyl solution at a pH of 3.0 and flow rate of 3.4 mL min-1. The flow rate of eluent solution (1 M HCl) was 1.4 mL min-1. As seen, the preconcentrated uranyl ions were easily eluted using this eluent within less than a minute. No significant change in the elution profile was noted when 0.5 and 0.1 M HCl solutions were used to elute uranyl from the microcolumn. Dynamic Capacity. To evaluate the dynamic capacity of the imprinted coated silica gel sorbent, the following procedure was used. A fixed concentration of uranyl (10 ?g mL-1) solution was pumped through a microcolumn containing 100 mg of sorbent at constant flow rates of 0.9, 1.4, 2.5, 3.4, and 4.0 mL min-1. From the exit of the column, 5 mL portions of effluent were collected in several vials. The concentration of uranyl in the collected effluent solutions was analyzed by ICP until the concentration of uranyl in the effluent was stable. Then, the adsorbed uranyl ions were eluted using 10 mL of 1 M HCl, diluted to 200 mL, and analyzed by ICP. It was found that there is no significant change in the dynamic capacity of the sorbent as the sample flow rate increased from 0.9 to 4.0 mL min-1. The average dynamic capacity of the sorbent (mean ( σ) for five sample flow rates was 52.9 ( 3.4 ?mol g-1. The literature-reported sorbent capacities for the bulk-synthesized uranyl-IIPs are 34.0 mg g-1 by Metilda et al.32 and 98.5 mg g-1 by Preetha et al.34 It should be noted that, in the
7122 Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Initial concentrations Ci ) 5 ?g mL-1. b k′ ) Kimprinted/Kcontrol.

present study, the effective sorbent is actually a thin layer of polymer on the inert silica particle support so that the weight of the polymer layer on the silica is only about 30% (CHN and TG) of the total weight of the coated particles. Thus, it is not surprising to observe a lower capacity for the proposed IIP in comparison with those prepared by bulk polymerization method.32,34 Selectivity Study. In IIPs, the cavities created after removal of the template are complementary to the imprint ion in size and coordination geometries. Furthermore, the unique shape of the uranyl ion may be expected to lead to much greater selectivity for the uranyl ion by ion imprinting, in comparison with other metal ions,28 whereas in the case of nonimprinted sorbent, the random distribution of ligand functionalities in the polymeric network results in no specificity in rebinding affinities. In several batch experiments, pairs of uranyl and coexisting cations or other selected inorganic cations (with close atomic radii, identical chemical properties, and high positive charges) were extracted by 100 mg of resin at a pH of 3.0. Table 3 summarizes the distribution ratios (Kd) and selectivity coefficients (k) of uranyl ion over Ce3+, La3+, Sm3+, Th4+, Y3+, ZrO2+, Mn2+, Fe3+, Co2+, Ni2+, Cu2+, Zn2+, Cd2+, and Pb2+ ions, calculated using eqs 2 and 3, respectively. As is obvious from Table 3, the obtained k values, for imprinted sorbent, ranged from >104 for UO22+/Co2+, Zn2+, Cd2+, Sm3+ to 4.45 × 102 for UO22+/Fe3+. The quantitative extraction of uranyl was obtained at an optimum pH of 3.0, while the competitor ions were weakly coextracted together with the target ion. A significant difference between the binding of uranyl ion and other competitor metal ions to the imprinted sorbent clearly suggests that the unique shape of uranyl ion plays an essential role in its selective binding to the polymer.28 The ionimprinting effect is clearly observed by comparing the selectivity results of imprinted and control material in terms of relative selectivity coefficients (k′) in Table 3. The high selectivity coefficients obtained on ion imprinted polymers coated silica particles enable the quantitative removal of uranyl from the metal cations in real samples. In further experiments, the uptake of uranyl ion was investigated in the presence of excess amounts of multicomponent

Table 4. Determination of Uranyl Ion in Spiked Tap and Caspian Sea Waters and Leach Liquor Samples (N ) 3, (σ) sample tap water Caspian Sea water leach liquor uranyl found BLQa BLQ 3.1 (?g mL-1) recovery of spiked uranyl, % 98.8 ( 2.1b 103.8 ( 1.8b 97.9 ( 3.5c

a Below quantification limit. b Spiked amount ) 5 ng mL-1. c Spiked amount ) 3 ?g mL-1.

mixtures of competitor ions at a pH of 3.0 and flow rate of 3.4 mL min-1. The results showed that the concentrations of 25 ?g mL-1 of Co2+, Zn2+, Cd2+, and Sm3+, 20 ?g mL-1 of Mn2+, Cu2+, Pb2+, and Ni2+, 15 ?g mL-1 of Ce3+, La3+, and ZrO2+, and 10 ?g mL-1 of Th4+, Y3+, and Fe3+ do not interfere in column operation for the determination of 25 ng mL-1 UO22+. The tolerance limit was set at the amount of foreign ions causing an error of less than (2% in the extraction of uranyl. Analytical Precision, Detection Limit, and Enrichment Factor. The precision, evaluated as relative standard deviation (RSD) for 12 replicate column extractions of 25 ng mL-1 UO22+, was 1.7%. For the preconcentration of uranyl ions, 100 mL of the aqueous standard solutions of UO22+ was passed through a microcolumn containing 100 mg of the imprinted sorbent at a pH of 3.0 and flow rate of 3.4 mL min-1. The preconcentrated ions were then eluted with 2 mL of 1 M HCl and determined by ICP. The detection limit, defined as the concentration of analyte giving signal equivalent to 3 times the blank standard deviation plus the net intensity of the blank for 100 mL of sample volume, was 1.6 ng mL-1. The enrichment factor, obtained by comparing the slopes of the linear portion of the calibration curves before and after the preconcentration, was 52. In literature, the enrichment factors for preconcentration of uranyl ion on bulk-synthesized IIPs have been reported as 100 by Metilda et al.32 and 20 by Bae et al.28 Determination of Uranyl Ion in Real Aqueous Matrices. To demonstrate the potential applicability of the prepared sorbent to the determination of uranyl in aqueous real matrices, the proposed method was applied to the determination and recovery of uranyl ion from tap and Caspian Sea water samples and an ore leach liquor sample. The aliquots of the samples were spiked with known amounts of uranyl ion. The solutions were adjusted at a pH of 3.0 and introduced into microcolumn containing sorbent as previously described. The obtained results given in Table 4

indicate the high potential of the proposed sorbent for preconcentration and determination of uranyl ion from natural water and environmental samples. Sorbent Stability. The possible role of the silica support on polymer stability and affinity was investigated. It was found that the stabilization of polymer chains on the silica surface has a clear influence on the high polymer affinity toward uranyl ion. It was observed that the bulk-polymerized imprinted materials, prepared under the same conditions as the grafted materials, yielded poor selectivities for the target ion. It capitalizes the key role of the silica support in the creation of a thin layer of rigid polymeric network with stabilized spatial orientation of polymer chains, which results in an improved specificity of binding sites. Similar effects have already been reported for the stabilizing function of the support in the grafting of a thin layer of MIPs onto the polystyrene microplate surface.21,22 The prepared ion imprinted polymer grafted silica sorbent was repeatedly used and regenerated for at least 3 months. No significant decrease was observed in the sorbent affinity over this period of time, and the column efficiency clearly remained stable. This stability of the sorbent without any evidence for polymer destruction is most possibly due to the strong covalent attachment of polymer chains to the surface of silica. CONCLUSIONS In conclusion, the results presented here demonstrate the efficiency of the “grafting from” polymerization procedure for generating a new ion-imprinted surface-coated silica gel sorbent with high affinity toward uranyl ion. The use of immobilized azo initiators provided controllable conditions for the “grafting from” process with high graft densities. Rapid kinetics of adsorption of uranyl ion on the resulting imprinted sorbent provides a fast column preconcentration procedure for uranyl ion in aqueous solutions. The new sorbent revealed an excellent selectivity toward uranyl ion over a range of strong competing metal ions. It was observed that the sorbent affinity was not affected during repeated uses for 3 months. The work is in progress in our laboratories to use the prepared imprinted sorbent in selective online solid-phase extraction, in order to achieve determination of subnanogram per milliliter level of uranyl ion and increased sampling frequency of the method in analysis of complex aqueous samples. Received for review May 13, 2007. Accepted July 18, 2007.

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007