Organo-functionalized mesoporous silicas for efficient uranium extraction

Microporous and Mesoporous Materials 180 (2013) 22–31

Contents lists available at SciVerse ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Organo-functionalized mesoporous silicas for efficient uranium extraction Juan L. Vivero-Escoto 1, Michaël Carboni, Carter W. Abney, Kathryn E. deKrafft, Wenbin Lin ⇑ Department of Chemistry, CB 3290, University of North Carolina, Chapel Hill, NC 27599, USA

a r t i c l e

i n f o

Article history: Received 19 January 2013 Received in revised form 28 May 2013 Accepted 30 May 2013 Available online 7 June 2013 Keywords: Uranium extraction Mesoporous silica Amidoxime Seawater Sorbent

a b s t r a c t A series of new mesoporous silica (MS) sorbents were developed by functionalizing a large-pore 2-D hexagonal MS material, MSU-H, with amidoxime, imide dioxime, phosphonate, and carboxylate functional groups, and characterized by nitrogen adsorption, f-potential, infrared spectroscopy, and thermogravimetric analysis. These MS materials have a grafting density of 0.75 to 1.38 mmol/g, and exhibit BET surface areas of 186–526 m2/g and average pore sizes of 3.8–7.8 nm. The uranyl sorption by the functionalized MS sorbents was investigated in basic water and artificial seawater at pH = 8.3 ± 0.1. The MS materials exhibited a high U sorption capacity in water (>40 lg U/mg sorbent) with Langmuir isotherms suggesting a saturation U sorption capacity of 185.2 lg U/mg sorbent for the phosphonic acidmodified MS material (MSPh-III). The U sorption capacity in artificial seawater was reduced to 12.1 lg U/mg sorbent for MSPh-III. Langmuir isotherms indicated a saturation sorption capacity of 66.7 lg U/mg sorbent for MSPh-III, which also had the greatest binding affinity for U of all sorbents tested, followed by the imide dioxime-functionalized material MSCA-I. Kinetics studies show rapid uranyl sorption and equilibration in less than 40 min. The U was quantitatively eluted from the MS sorbents by washing with strong acid (>0.1 M HCl). This work represents the first comprehensive study of organo-functionalized MS materials for U extraction, and shows that phosphonic acid- and imide dioxime-functionalized MS materials provide excellent platforms for developing novel sorbents for efficient U extraction from seawater. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction The current annual global energy consumption of 13 Terrawatts [1] is projected to double by 2050 due to the increase in global population and standard of living [2]. While 81% of the current power need is supplied from burning fossil fuels [3], a shift towards renewable energy supplies is urgently needed in order to sustain global economic growth and to mitigate the climate change caused by a rapid rise of the atmospheric carbon dioxide level. While efficient solar power generation provides a long-term, sustainable solution to the global energy need, nuclear power remains the only mature, large-scale, non-intermittent, and carbon–neutral means to meet this growing energy demand in the coming decades (and centuries). Development of stable and clean energy supplies through nuclear power is a responsible pursuit. Among all radionuclides, uranium is the predominant fuel for nuclear reactors. For that reason, the extraction and enrichment of uranium has important strategic and economic value. As the ⇑ Corresponding author. Tel.: +1 919 962 6320; fax: +1 919 962 2388. E-mail address: [email protected] (W. Lin). Current address: Department of Chemistry, Burson 200, University of North Carolina at Charlotte, Charlotte, NC 28223, USA. 1

1387-1811/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.micromeso.2013.05.030

uranium in terrestrial ores is limited, extraction from other sources such as waste coal ash and seawater is actively being explored [4– 6]. The oceans contain 4.5 billion tons of uranium, equivalent to 1000 times the currently available terrestrial uranium that can be economically mined. Efficient extraction of uranium from seawater represents an attractive solution for the development of sustainable nuclear fuel cycles [4,7]. Sequestration of uranium from seawater would greatly improve its availability and sustain the fuel supply for nuclear energy, thus removing an important hurdle for increasing nuclear power supply. Because uranium is present at an extremely low concentration of 3 ppb in seawater, economical extraction of uranium from seawater presents a major scientific and technological challenge. Several methods for uranium collection have been introduced and evaluated over the past decades, including ion exchange, solvent extraction, foam separation, co-precipitation, biomass collection, and adsorption [4,5]. Materials such as ion-exchange resins, titanium oxide, activated carbon, galena, and polymers have been used [4,8–11]. However, none of these techniques have been practical for reasons such as ion selectivity, long term stability, and cost effectiveness. Recent research has focused on sorbents containing organic functional groups (such as amidoxime) as promising candidates for uranium sorption [7–16]. Uranium is typically found as

23

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

the uranyl ion (UO22+) in aqueous solutions, which is a hexavalent U species [17,18]. Simple amidoximes are known to bind the uranyl ion in aqueous solution, with recent data suggesting an g2binding mode when grafted onto polymer braids [19–21]. Polymer beads and fibers functionalized with amidoxime groups have shown uranium sorption capacities as high as 1.5 lg U/mg and 6.0 lg U/mg adsorbent in seawater and laboratory conditions, respectively [4]. Further improvements to uranium collection systems in areas of selectivity, loading capacity, sorption kinetics, as well as chemical and mechanical stability must be made in order to make uranium extraction from seawater economically feasible. Mesoporous silica (MS) materials have attracted attention in recent years due to their useful features such as large surface area, tunable pore volume and size, and facile modification of physical and chemical properties through surface functionalization [22,23]. These characteristics make MS materials an ideal platform for generating a variety of hybrid materials for applications ranging from catalysis to nanomedicine [24–29]. MS materials have also been successfully tested in the removal of several heavy metal ions, oxyanions, and toxic organic species from aqueous solutions [30,31]. In the area of actinide sequestration, Fryxell and others recently used surface-modified MS materials for selective U sequestration under acidic conditions [32–35]. Comprehensive studies and comparisons of U extraction with organo-functionalized MS materials have not been carried out to date. Herein we report the synthesis and characterization of a series of surface-modified MS materials and their application as sorbents for U extraction from seawater. A series of amidoxime-, imide dioxime-, phosphonate-, and carboxylate functional groups were grafted onto commercially available MSU-H mesoporous material (Scheme 1), and their sorption performances were investigated in slightly basic water and

artificial seawater. The sorption kinetics, U desorption, and sorption isotherm studies were measured under similar conditions. The phosphonic acid-modified MS material (MSPh-III) exhibited a very high saturation capacity for U sorption in water (185.2 lg U/ mg sorbent) and in artificial seawater (66.7 lg U/mg sorbent) according to the Langmuir adsorption model. Our goal is to compare the performance of these MS materials with different functionalities, thus identifying the best candidates for more thorough evaluations of their potential in uranium extraction from seawater. 2. Experimental 2.1. Materials and methods All reagents including the MSU-H material were purchased from Aldrich and used without further purification, except for 3-(isocyanatopropyl)triethoxysilane (ICP-TES), 3-(cyanopropyl)triethoxysilane (CP-TES), and 3-(iodopropyl)trimethoxysilane (IPTMS) which were purchased from Gelest. Thermogravimetric analysis (TGA) was carried out with a Shimadzu TGA-50 equipped with a platinum pan and using a heating rate of 5 °C/min under air. The surface charge of the synthesized MS materials was measured in 1.0 mM phosphate buffered saline solution on a Malvern ZetaSizer dynamic light scattering instrument. NMR spectra were recorded on a Bruker NMR 400 NB at 400 MHz. Mass spectrometric analyses were conducted using positive-ion electrospray ionization on a Bruker BioTOF Mass Spectrometer. The IR spectra were obtained using an Alpha-T Bruker Fourier Transform Infrared Spectrometer in attenuated total reflectance (ATR) mode. For all the experiments, uranyl concentration (UO22+) in solution was

HO N H2N

OH HO

NH2 H2N

HO

N

N

NH

OH N

O Si O

N

HO

N

O

O

O NH2

OH

H N

NH

O NH

O NH

O

O Si O O

MSA-I

MSA-II

O Si O

O

O O

MSA-III O

Si

P

O

MSGly

MSCA-I OH

HO

O

O Si O

O

OH

O P

NH

NH

O

O P

OH OH P

OH

O NH

O

OH P

O O

Si

MSPh-I

O

O NH

NH

O

O Si O

NH

NH

O

O

MSPh-II

NH

O Si O O

O Si O

MSPh-III

MSPh-IV

O

Scheme 1. Organo-functionalized mesoporous silica sorbents for U sorption from water and artificial seawater.

24

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

measured by the Arsenazo(III) colorimetric assay using a UV-visible (Shimadzu UV-2401 PC) spectrophotometer at a wavelength of 652 nm [13]. Uranyl concentration was also obtained for the MSU-H, MSA-I, and MSPh-III samples by inductively coupled plasma mass spectrometry (ICP-MS) (Varian 820-MS Spectrometer) in order to confirm the results obtained by analysis UV-visible. Nitrogen adsorption experiments were performed with a Quantachrome Autosorb-1C. Size and zeta potential information was obtained on a Malvern ZetaSizer dynamic light scattering instrument. Infrared spectroscopy (IR) was performed using a Bruker Alpha-T Fourier Transform Infrared Spectrometer in attenuated total reflectance (ATR) mode. Powder X-ray diffraction (PXRD) analyses were carried out using a Bruker SMART Apex II diffractometer using Cu radiation. The PXRD patterns were processed with the Apex II package using the phase ID plugin.

2.2. Synthesis of organic ligands Synthesis of triethoxysilylpropyl-3-p-amidoximephenyl-urea (1). A mixture of 0.412 g of 4-aminobenzonitrile (3.5 mmol) and 1 g of sodium carbonate was added to a solution of 20 mL of water:ethanol (3:1 V/V), and heated in an oil-bath at 70 °C. Afterwards, an aqueous solution of 1.4 g hydroxylamine-hydrochloride in 5 mL of water was slowly added. The mixture was stirred overnight at 70 °C. The desired p-amidoxime aniline precipitated during the cooling process, and was filtered off and washed with cold water and ethanol several times. The product was dried to obtain a light beige substance. Yield: 0.45 g (85%). 1H NMR (400 MHz, DMSO-d6):d 5.22 (s, 2H, NH2); 5.48 (s, 2H, NH2 (amidoxime)); 6.59 (d, 2H); 7.31 (d, 2H); 9.15 (s, 1H, OH). MS (ESI positive ion mode): m/Z = 151.89 (Expected 152.07), 303.15 (Expected 303.14) for [M+1]+ and [2M+1]+, respectively. To synthesize the corresponding silane derivative, 0.25 g (1.65 mmol) of p-amidoxime aniline and 0.55 mL (2.2 mmol) of 3-isocyanatopropyltriethoxysilane were added to 15 mL of dry N,N0 -dimethylformamide (DMF) and stirred for 24 h. The sample was then lyophilized for 48 h to remove the solvent. Yield: 0.35 g (51.5%). 1H NMR (400 MHz, DMSO-d6): d 0.51 (m, 2H); 1.13 (t, 9H); 1.51 (m, 2H); 3.06 (m, 2H); 3.73 (q, 6H), 5.44 (s, 2H, NH2 (amidoxime)), 6.53 (d, 2H), 7.46 (d, 2H). MS (ESI positive ion mode): m/Z = 399.23 (Expected 399.21), 421.20 (Expected 421.21) for [M+1]+ and [M+Na]+, respectively. Synthesis of triethoxysilylpropyl-3-pentanyldinitrile-carbamate (2). 0.193 mL (2 mmol) of 3-hydroxyglutaronitrile, 0.5 mL (2 mmol) of 3-isocyanatopropyltriethoxysilane, and 0.348 mL (2 mmol) of diisopropylethylamine were dissolved in 5 mL of dry dimethyl sulfoxide (DMSO). The resultant solution was stirred at r.t. for 48 h. The desired product was obtained after the removal of the solvents with a rotary evaporator at r.t. Yield: 0.46 g (65.0%). 1H NMR (400 MHz, DMSO-d6): d 0.52 (m, 2H); 1.13 (t, 9H); 1.45 (m, 2H); 2.85–3.00 (m, 6H); 3.72 (q, 4H), 5.08 (m, 1H); 7.53 (t, 1H, NH). MS (ESI positive ion mode): m/Z = 277.3 (Expected 277.21) for [M-3(EtO)+1]+. Synthesis of N-trimethoxysilylpropyl-N-phthalimidyl dioxime (3). Phthalonitrile (0.5 g; 3.49 mmol), sodium carbonate (1.0 g; 9.43 mmol), and hydroxylamine hydrochloride (1.4 g; 20.1 mmol) were dissolved in 20 mL of water:ethanol (3:1 V/V). The resultant solution was heated at 75 °C in an oil-bath for 8 h. The product gradually precipitated as yellow crystals upon cooling. The desired 5-amino-phthalimide dioxime was collected by filtration and washed several times with ethanol. Yield: 0.53 g (79%). 1 H NMR (400 MHz, DMSO-d6): d 5.662 (s, 2H, –NH2); 6.661 (dd, 1H); 6.77 (s, 1H); 7.272 (dd, 1H); 8.554 (s, 1H, NH); 10.101 (s, 1H, N-OH); 10.378 (s, 1H, N-OH). 13C NMR (400 MHz, DMSO-d6): d 150.59; 146.47; 132.72; 121.01; 118.55; 115.89; 103.52. MS

(ESI positive ion mode): m/Z = 192.88 (Expected 192.87) for [M+1]+. The silane derivative (3) was synthesized by dropwise addition of 3-iodopropyltrimethoxysilane (389 lL, 1.976 mmol) to a solution of 5-amino-phthalimide dioxime (400 mg, 2.08 mmol) in 16 mL of dry DMF. The resulting mixture was stirred at r.t. for 24 h, and then lyophilized for 36 h to afford 0.455 g of 3 (61.7% yield). 1H NMR (400 MHz, DMSO-d6): d 0.724 (broad peak, 2H); 1.838 (broad peak, 2H); 3.276 (t, 2H); 3.471 (s, 6H); 6.708 (dd,1H); 6.808 (s, 1H); 7.305 (dd, 1H); 8.69 (s, 1H, NH); 10.16 (s, 1H, N–OH); 10.43 (s, 1H, N–OH). MS (ESI positive ion mode): m/Z = 355.17 (Expected 355.16) for [M+1]+. Synthesis of 1-triethoxysilylpropyl-3-ethyl diethylphosphonate-urea (4). To a solution of diethyl(2-aminoethyl)phosphonate [36] (0.3 g; 2.4 mmol) in 6 mL of dry DMF was added 0.6 mL (2.4 mmol) of 3-isocyanatopropyl triethoxysilane dropwise. The resulting mixture was stirred for 24 h, and then lyophilized for 24 h to afford 0.45 g of 4 (43.8% yield). 1H NMR (400 MHz, DMSO-d6): d 0.49 (t, 2H); 1.12 (t, 9H); 1.22 (t, 6H); 1.42 (m, 2H); 1.98 (broad peak, 2H); 2.96 (m, 2H), 3.33 (q, 2H); 3.71 (q, 6H); 3.93 (q, 4H). MS (ESI positive ion mode): m/Z = 158.1 (Expected 158.21) for [M-3(EtO)+1]+. Synthesis of 1-triethoxysilylpropyl-3-ethyl tetraethyl-bisphosphonate urea (5). Tetraethyl(2-aminoethyl)bisphosphonate was synthesized following a procedure already reported in the literature [36,37]. To synthesize the desired silane derivative, 0.85 g (2.7 mmol) of tetraethyl(2-aminoethyl)bisphosphonate was dissolved in 3 mL of dry DMF. To this solution, 0.875 mL (3.5 mmol) of 3-isocyanatopropyltriethoxysilane was added, and the final solution was stirred for 24 h at room temperature under argon atmosphere. The desired product was obtained after lyophilizing the sample for 48 h. Yield: 0.75 g (49.2%). 1H NMR (400 MHz, DMSO-d6): d 0.50 (broad peak, 2H); 1.13 (t, 9H); 1.22 (t, 12H); 1.43 (broad peak, 2H); 1.95 (m, 1H); 2.90–3.06 (broad peak, 2H); 3.43 (t, 2H); 3.73 (q, 6H), 4.02 (q, 8H). 2.3. Synthesis of organo-functionalized mesoporous silica materials In general, for grafting organic functional groups onto MSU-H, 3 mmol of the desired silane derivative was slowly added to 0.5 g of MSU-H dispersed in 50 mL of toluene. In the case of solid silane ligands, the compound was dissolved in 1.0 mL DMSO prior to addition. The final dispersion was refluxed at 90 °C for 12 h. The material was collected by centrifugation at 10,000 rpm for 10 min, washed with methanol and ethanol, and stored in ethanol. MSA-I was synthesized by grafting cyanopropyltriethoxysilane on MSU-H as described above. This material (120 mg) was reacted with hydroxylamine (387.2 mg) in the presence of sodium carbonate (275 mg) in 20 mL water:ethanol solution (5:1.6 V/V) at 70 °C for 12 h to afford the corresponding amidoxime derivative (MSA-I). MSA-II was produced by grafting compound 1 on MSUH as described above. MSA-III was synthesized by grafting compound 2 on MSU-H as described above, followed by treating the grafted material (120 mg) with hydroxylamine (387.2 mg) in the presence of sodium carbonate (275 mg) in 20 mL water:ethanol solution (5:1.6 V/V) at 70 °C for 12 h. MSCA-I was produced by grafting compound 3 on MSU-H as described above. MSPh-I was synthesized by grafting commercially available 3-(trihydroxysilyl)propyl methylphosphonate on MSU-H as described above. MSPh-II was afforded by grafting compound 4 on MSU-H as described above. MSPh-III was synthesized by deprotecting the phosphonate ester ligand of MSPh-II material with trimethylsilyl bromide (TMS-Br) in dry DMF as has been reported previously in the literature [32]. MSPh-IV was produced by grafting compound 5 on MSU-H as described above, followed by deprotection of the bisphosphonate ester ligand with TMS-Br. MSGly was produced

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

by following a procedure reported in the literature, using MSU-H as the MS material [32]. 2.4. Uranium sorption experiments Artificial seawater was prepared as reported in the literature [18], with an initial U concentration (from uranyl acetate) at 5 ppm. The high U concentration was chosen to facilitate rapid screening experiments and accommodate analysis by UV–visible spectroscopy. The solution pH was adjusted to 8.3 ± 0.1 by adding small volumes of 1.0 M sodium hydroxide. In a typical sorption experiment, 4 mg of MS material was added to 40 mL of U solution in a high-density polyethylene bottle. After shaking at 300 rpm with a PRO Scientific VSOS-4P orbital shaker for 1 h at room temperature, the sorbent was separated by centrifugation for 10 min at 12,000 rpm. The U concentration in the supernatant was determined by UVvisible spectroscopy using the Arsenazo (III) assay [13]. To 2 mL of supernatant, 40 lL concentrated nitric acid and 200 lL Arsenazo (III) dye solution were added. The solution was vortexed thoroughly, allowed to sit for 10 min, and analyzed by UV–visible spectroscopy. Concentrations were obtained using a previously measured standard curve. A blank and a sample of U-spiked water without sorbent were analyzed as controls for each sorption experiment. 2.5. Uranium elution from mesoporous silica materials The U absorbed in the MS material was eluted by washing the material with 10 mL 0.01 M, 0.1 M, and 1.0 M HCl aqueous solutions, successively. The washing step consists of suspending an MS material in 10 mL of 0.01 M HCl solution and sonicating for approximately 30 s, followed by collecting the MS material by centrifugation. The supernatant was analyzed by UV-visible spectroscopy as described previously. This procedure was repeated for 0.1 and 1.0 M HCl solutions. The total amount of uranyl eluted from the MS sorbents was the sum of quantities measured in the three eluent solutions. Langmuir sorption isotherms were obtained by slight modification of the procedure described above for sorption experiments. Samples of 1.0 mL water or artificial seawater were prepared containing varied concentrations of U (1, 5, 10, 20, 30, and 40 ppm for water, and 1, 4, 8, 17, 25, and 33 for seawater). Specific U concentration was determined for a control prior to the experiment. MS sorbent (0.1 mg) was added to each sample and shaken at 300 rpm as described previously. The sorbent was separated by centrifugation at 12,000 rpm for 60 s. The U concentration was measured by UVvisible spectroscopy using the Arsenazo(III) assay [13]. Sorption kinetics were determined by slight modification of the procedure described above for sorption experiments. Samples (10 total for 10 different time points) of 1.0 mL water and artificial seawater were prepared at a U concentration of 5 ppm. MS sorbent (0.1 mg) was added to each sample and shaken at 300 rpm. Individual samples were analyzed at fixed time intervals. Sorbent was separated by centrifugation for 60 s at 12,000 rpm, and the U concentration was measured by UV-visible spectroscopy using the Arsenazo(III) assay [13]. 3. Results 3.1. Synthesis and characterization of mesoporous silica-based sorbents A series of amidoxime-, imide dioxime-, phosphonate-, and carboxylate-based MS materials were synthesized and evaluated as

25

sorbents for extracting U from water and artificial seawater (Scheme 1). Previous research has demonstrated the utility of the amidoxime functional group for U sequestration, prompting preparation of mono- and bis-amidoxime functionalized ligands (MSA-I and –III, respectively) [5,6,12–16,19]. Electronic effects of an amidoxime adjacent to an aromatic ring were investigated by MSAII. Recent studies have focused on a cyclized imide dioxime for extracting U [28]. Thermodynamic investigations revealed the simplified glutarimidedioxime binds U in a tridentate-fashion and is enthalpically preferred to the UO2(CO3)34 species dominant in seawater [20]. MSCA-I was prepared to provide a direct comparison of this type of ligand against amidoxime. MSGly was previously grafted on an different mesoporous slilica support and demonstrated to be an effective sorbent for actinide removal [32]. As physical differences between silica supports would prohibit direct comparison, the glycinyl-urea ligand was grafted to MSU to form the material MSGly. The amidoxime-based materials were produced using two different methods. In the first approach, the MS sorbent was synthesized by grafting the corresponding nitrile silane derivative onto MSU-H, followed by conversion of the nitrile to an amidoxime group by treatment with hydroxylamine to afford the MSA-I and MSA-III sorbents. In the second approach, the amidoxime and cyclic amidoxime compounds were first synthesized, and the silane derivative formed by a subsequent reaction with trialkoxysilane. Finally, the silane derivative was grafted onto MSU-H to afford the MSA-II and MSCA-I sorbents. The carboxylate-based sorbent MSGly was obtained by following a procedure reported in the literature while substituting MSU-H as the mesoporous silica support [32]. Phosphorus-based materials such as tributyl phosphate (TBP) and carbamoylmethylphosphine oxide (CMPO) are known to complex strongly to U and other actinides and are used in PUREX and TRUEX decontamination processes. Furthermore, phosphonatefunctionalized mesoporous silica were recently demonstrated to be effective sorbents for removing U from aqueous matricies [38,39]. Phosphonate-based sorbents were prepared to probe the binding motif for this class of materials. A simple methyl phosponate (MSPh-I) was intended to serve as a baseline material and to assess whether the absence of an oxygen would impede U sorption. A diethoxy-protected phosphonate (MSPh-II) was previously studied and published, though on a different solid support [39]. The deprotected mono- and bis-functionalized analogs (MSPh-III and –IV, respectively) were prepared to investigate whether chelation of U might be facilitated by deprotonation of an acidic phosphoryl group. To prepare the phosphonate-functionalized MS materials, phosphonate-containing silanes were first grafted onto MSU-H. The phosphonic acid version of the materials (MSPh–III and –IV) was generated by deprotecting the phosphonate groups with trimethylsilyl bromide (TMS-Br). MSPh-I was synthesized by grafting a commercially available 3-(trihydroxysilyl)propyl methylphosphonate on MSU-H. The structural and chemical properties of the MS sorbents were characterized by nitrogen adsorption (surface areas and pore diameters), f-potential, IR spectroscopy, and thermogravimetric analysis (TGA), as summarized in Table 1. TGA results indicate grafting densities of 0.75 to 1.38 mmol/g for the organo- functionalized MS materials (Table 1 and Figure S1). The nitrogen sorption isotherms are type IV for all MS sorbents, typical of mesoporous materials (Figure 1 and Figures S2–S11). The Brunauer-Emmett-Teller (BET) surface areas range from 186.4 to 526.0 m2/g, and the average pore diameters range from 3.8 to 7.8 nm, indicating their potential as highly efficient sorbents for uranyl ions. The surface areas and average pore diameters of the organo-functionalized MS materials are smaller than those of MSU-H (SA = 648.3 m2/g and average pore size = 9.6 nm), consistent with successful functionalization of the materials.

26

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

Table 1 Structural properties of organo-functionalized mesoporous silica materials. Material

BET surface area (m2/g)

BJH pore size (nm)

Loading (mmol/g)

f-potential (mV)

MSU-H MSA-I MSA-II MSA-III MSCA-I MSPh-I MSPh-II MSPh-III MSPh-IV MSGly

648 494 186 526 277 515 227 406 253 187

9.6 7.8 4.9 7.8 5.6 3.8 6.6 7.8 5.6 4.9

– 0.75 1.09 1.02 1.32 1.38 1.02 0.82 1.02 1.25

27.7 ± 1.3 10.2 ± 2.0 +15.1 ± 0.7 +26.8 ± 1.5 +11.4 ± 0.2 37.1 ± 1.5 20.7 ± 0.9 40.9 ± 1.5 46.6 ± 1.8 44.6 ± 2.5

The FT-IR spectra of the MS materials exhibited the characteristic stretching vibrations at 490–400 and 1090–1030, 850–800, and 1110–1000 cm1, corresponding to Si–O–Si, Si–C, and Si–O–C, respectively (Figures S12–S20). In general, the amidoxime-functionalized MS materials showed diagnostic bands such as a broad peak around 3600–3300 associated with O–H, N–H, free oxime, and oxime stretching modes, as well as a 1671 cm1 peak due to hydroxyimine, characteristic of the amidoxime functional group. In particular, MSCA-III showed a vibration band at 1717.8 cm1 corresponding to the C@O stretch of the carbamate functional group. MSA-II and MSCA-I exhibited a vibration peak at 3025.6 cm1 corresponding to the C–H stretch in the aromatic group. The presence of the phosphonate functional groups in MS

(a)

500 MSPh-II MSPh-III MSA-II

3.2. Uranium sorption in MS-based sorbents

3

N2 Uptake (cm /g)

400

materials is evidenced by the C–H stretching vibrations at 2991.1 and 2955.7 cm1. Moreover, the IR spectra of MSPh-II, MSPh-III, and MSPh-IV showed a band at 1690–1650 cm1 corresponding to the C@O stretch of the urea group. The surface charge of the materials was measured as f-potential by dynamic light scattering measurements. The f-potential varied depending on the type of moiety grafted to the MSU-H. The f-potential for MS materials modified with amidoxime groups was shifted from 27.7 mV for unfunctionalized MSU-H to more positive values following functionalization (10.2 mV, +15.2 mV, +26.8 mV, and +11.4 mV for MSA-I, -II, -III, and MSCA-I, respectively). In contrast, the f-potential for phosphonic acid- and carboxylate-modified materials was shifted toward more negative values than the unfunctionalized MSU-H. The values of f-potential for MSPh-I, MSPh-III, MSPh-IV, and MSGly materials are 37.1, 40.9, 46.6, and 44.6 mV, respectively.

300

200

100

0 0.0

0.2

0.4

0.6

0.8

1.0

Relative Pressure (P/P 0 )

(b) MSPh-II MSPh-III MSA-II

2

Dv(d) (m /g)

1200

800

400

0 4

8

12

16

Pore Diameter (nm) Fig. 1. Representative nitrogen sorption isotherms (a) and pore size distributions (b) of MSA-II (red), MSPh-II (grey), and MSPh-III (black) materials. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

It has been shown in recent reports that MS sorbents with higher surface area, larger pore diameter, and 3-D pore structure possess higher U sorption capacity. MCM-48 materials (with 3-D pores) have higher sorption capacities than MCM-41 materials (with 1-D pores) [40]. Moreover, in water, the U sorption in MS materials increases rapidly with increasing basicity in the pH range of 2–8. This is attributed to the electrostatic interactions between silanol and uranyl species under different pH conditions [34,41,42]. In addition, sorption kinetics are fast with an equilibration time of less than 30 min [34,41,42]. The fast sorption rate suggests that the silanol groups are readily available and easily accessible in MS materials. To evaluate the sorption performance of the new MS sorbents, the U sorption capability was tested in slightly basic water and artificial seawater, both at environmental seawater pH (8.3 ± 0.1). In water, all the MS sorbents showed a high uranyl sorption capability at equilibrium (qe), ranging from 40 to 50 lg U/mg sorbent (Figure 2a). The difference in U sorption between the sorbents under these experimental conditions was not statistically significant. However, in artificial seawater the amount of absorbed U was reduced at least fourfold, with the amount of U adsorption varying between 2 and 13 lg U/mg sorbent (Figure 3a). As reported in the literature, the presence of carbonates in solution reduces the U sorption, presumably due to the formation of anionic U(VI) carbonate complexes that do not bind to the sorbents as strongly [34,42]. The following trend was observed for the U adsorption in seawater by MS sorbents: MSPh-III  MSU-H > MSA-I  MSAII  MSA-III  MSCA-I > MSPh-II > MSGly  MSPh-I  MSPh-IV. Given the much lower surface area of MSPh-III (406.1 m2/g) than MSU-H (648.3 m2/g), the high U sorption capacity exhibited by the MSPh-III material is very impressive. For comparison, amidoxime fibers tested under similar conditions yielded U sorption of 10.5 lg U/mg sorbent, which is statistically equivalent to the

27

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

(a)

(a)

50

15

45

35

qe (µg U / mg sorbent)

qe (µg U / mg sorbent)

40

30 25 20 15 10

10

5

5

(b) 160

(b)

140

hII M SP hIII M SP hIV

M SP

M SP hI

M SG ly

3000

100

2500

Kd (mL/g)

3

4000 3500

120

Kd (10 mL/g)

M SA -II M SA -II I M SC AI

SA -I M

M

M

SP

SP

SP M

0

hIV

hIII

hII

hI

ly

SP M

M

SC M

M

SG

-II I

-II

SA

-I M

SA

SA M

AI

0

80 60

2000 1500

40

1000 20

500 0

IV

III

Fig. 2. (a) Comparison of U sorption capacities for different MS sorbents in water. (b) U distribution coefficients for different MS sorbents in water. pH 8 ± 0.1; msorbent/ vsolution = 0.1 mg/mL; [U]initial = 5 ppm.

MSPh-III material. The U sorption capacity previously reported for amidoxime polymers is 1.5 and 6.0 lg U/g sorbent in seawater and under laboratory conditions, respectively [4]. The fact that the MSA-I, MSA-II, and MSA-III materials do not perform nearly as well as the amidoxime fiber suggests that the uranyl ions do not simply coordinate to amidoxime groups for the amidoxime fiber. Pretreatements of amidoxime fibers with KOH are known to enhance extraction of U, and it is suggested this may alter the actual sorbent group, forming a different binding ligand with improved sorption properties [4,16,28]. More work is needed to elucidate the binding sites in the amidoxime fibers. Distribution coefficients (Kd) for the new MS materials were determined using the following formula:

K d ¼ ðC 0  C e Þ=C e  ðV=mÞ

ð1Þ

where C0 (lg/mL) and Ce (lg/mL) are the initial and equilibrium concentration of U, respectively; V (mL) is the volume of the testing solution, and m (g) is the sorbent dose. The distribution coefficient is a value expressing the sorbent’s sorption capability at a particular concentration. The higher the Kd value, the more effective the sorbent material is at sequestering the target species. Kd values in mL/g above 500 are considered acceptable, those above 5000 are considered very good, and values higher than 50,000 are considered outstanding [32]. Figures 2band 3b show the Kd values for the U sorption in MS-based sorbents in water and artificial seawater, respectively. The Kd values for MS samples in water are remarkable, with all possessing values higher than 30,000 mL/g. In particular, the MSPh-III material exhibited an outstanding Kd value of 160,000 mL/g. However, when tested using artificial seawater, the

I

hIV

-II

SP M

SP h M

M

SP

h-

II

I SP hM

ly M SG

M SA -II M SA -II I M SC AI

h-

h-

SP M

M SA -I

II h-

SP M

SP M

M

SP

h-

I

ly

A-

SG M

I M

SC

-II

-II

SA M

M

SA

-I SA M

I

0

Fig. 3. (a) Comparison of U sorption capacities for different MS sorbents in artificial seawater. (b) U distribution coefficients for different MS sorbents in artificial seawater. pH = 8.3 ± 0.1; msorbent/vsolution = 0.1 mg/mL; [U]initial = 5 ppm.

Kd values for MS sorbents range from 300 to 3200 mL/g, and follow the same trend as the sorption capacity values described above. To elute the adsorbed U from MS sorbents, various concentrations of HCl (0.01 M, 0.1 M and 1.0 M) were used as eluent. A complete elution of U can be achieved using HCl concentrations higher than 0.1 M (Table S1) [12,14,15]. This was confirmed by ICP-MS measurements of digested MS materials after elution showing negligible amounts of U (<0.003 lg U/mg sorbent) remaining in the samples. 3.3. Uranium sorption isotherms and kinetics To further investigate the sorption properties of these MS materials, the sorption isotherms in both water and seawater were determined for six of the MS samples (Figs. 4 and 5). The obtained results were first fit to the Langmuir isotherm model which assumes that sorption occurs on an energetically homogeneous surface by monolayer deposition, and with no interactions between the adsorbates on adjacent sites [43]. Its linear form can be expressed as:

ðC e =qe Þ ¼ ðC e =QÞ þ 1=ðQ  bÞ

ð2Þ

where Ce is defined as above, qe represents the amount of adsorbed metal ion at equilibrium (lg/mg) while Q and b are Langmuir constants related to saturation sorption capacity (lg/mg) and affinity of the binding site on sorbent (mL/mg), respectively. Q and b can be obtained by plotting Ce/qe versus Ce. Table 2 shows the parameters of the Langmuir model for the U sorption by MS sorbents in water.

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

The near unity correlation coefficients for these fits demonstrate that U sorption by MS materials under these conditions follows the Langmuir sorption model. Langmuir model fitting was also performed for U sorption by MS sorbents in artificial seawater, as displayed in Table 3. Under these conditions, MSPh-III also exhibited the highest saturation capacity (66.7 lg U/mg sorbent), followed by MSCA-I (58.1 lg U/ mg sorbent). MSPh-III and MSCA-I also exhibited the highest uranyl binding strength as evidenced by the b values. In an effort to take the non-idealized surfaces into consideration, the data were also modeled using the Freudlich isotherm. The Freundlich isotherm is applicable to sorption on heterogeneous surfaces, and is frequently employed to analyze experimental data for determination of sorption isotherms [44,45]. The linear form can be defined as:

lnðqe Þ ¼ lnðkÞ þ ð1=nÞlnðC e Þ

80

qe (µg U/mg sorbent)

28

60

40

20

0 0

ð3Þ

where Ce and qe are defined as in the Langmuir isotherm, k (lg/mg) and n are the Freundlich constants relating sorption capacity and sorption intensity, respectively. The parameters of the Freundlich model for the U sorption in MS-based sorbents in seawater conditions are shown in Table 4. The sorption kinetics of U in several MS-based sorbents were studied in both water and artificial seawater. An initial [U] = 5 ppm was used with a contact time of 120 min (Figures S21 and S22). The U sorption in MS materials is fast, with most of the U adsorbed in the first 5–15 min, and equilibrium occurring at approximately 40 min. The fast sorption kinetics are consistent with the presence of large channels in the MS materials, and similar sorption kinetics have been reported for mesoporous materials [34,39,41,42]. Modeling studies performed with homogeneously functionalized mesoporous silica indicate pore size and long-range structural ordering to be the dominant influences on accessibility to sorbent sites [46–48]. While beyond the scope of this manuscript, a more detailed knowledge of transport phenomena could be obtained by thorough modeling of the adsorption process for the materials reported herein [49–51]. 4. Discussion 4.1. Grafting of MS materials

MSU-H MSA-II MSA-III MSCA-I MSPh-III MSPh-IV

10

20

30

Ce (µg/mL) Fig. 5. U Sorption isotherms for different MS sorbents in artificial seawater. pH = 8.3 ± 0.1; msorbent/vsolution = 0.1 mg/mL.

Table 2 Langmuir isotherm model parameters and correlation coefficients for U sorption by MS sorbents in water. Sample

1/Q

1/(Q  b)

Q (lg U/mg sorbent)

b (mL/mg)

R2

MSU-H MSA-II MSA-III MSCA-I MSPh-III MSPh-IV

0.012 0.015 0.015 0.016 0.005 0.012

0.017 0.023 0.015 0.006 0.004 0.021

80.6 65.4 67.6 61.0 185.2 84.7

0.74 0.67 1.00 2.78 1.32 0.57

0.993 0.995 0.996 0.999 0.991 0.996

Table 3 Langmuir isotherm model parameters and correlation coefficients for U sorption by MS sorbents in artificial seawater. Sample

1/Q

1/(Q  b)

Q (lg U/mg sorbent)

b (mL/mg)

R2

MSU-H MSA-II MSA-III MSCA-I MSPh-III MSPh-IV

0.023 0.046 0.032 0.017 0.015 0.031

0.18 0.46 0.26 0.10 0.06 0.31

43.1 21.6 31.1 58.1 66.7 32.5

0.13 0.10 0.12 0.18 0.26 0.10

0.933 0.883 0.965 0.975 0.979 0.929

The grafting of organic functionalities onto identical mesoporous silica supports was intended to create materials with similar Table 4 Freundlich isotherm model parameters and correlation coefficients for U sorption by MS sorbents in artificial seawater.

qe (µg U/mg sorbent)

160

MSU-H MSA-II MSA-III MSCA-I MSPh-III MSPh-IV

120

Sample

ln (k)

1/n

k (lg U/mg sorbent)

n

R2

MSU-H MSA-II MSA-III MSCA-I MSPh-III MSPh-IV

1.80 0.98 1.36 2.21 2.71 1.19

0.54 0.55 0.58 0.55 0.45 0.63

6.05 2.67 3.88 9.15 14.97 3.30

1.86 1.80 1.72 1.82 2.24 1.58

0.98 0.92 0.97 0.97 0.98 0.95

80

40

0 0

10

20

30

Ce (µg/mL) Fig. 4. U Sorption isotherms for different MS sorbents in water. pH = 8.3 ± 0.1; msorbent/vsolution = 0.1 mg/mL.

loading, porosity, and surface area (Table 1). The differences in sorption capacity are due to the different affinities of the organic moieties for U, rather than the physical properties of the support. MSU-H with no functionalized sorbent groups was investigated for comparison. As U sorption by this material is due to non-specific binding, discussion will largely focus on comparisons between the functionalized silica materials which absorb U by binding through the grafted functionalities. It is important to note that the high level of sorption shown by MSU-H indicates that non-functionalized space on the silica materials will physisorb U.

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

Analysis of nitrogen isotherms for functionalized materials (Figures S2–S11) reveals minimal tailing in the mesoporous region and indicates homogeneous grafting of the silica support, while similar sorbent loading between all materials suggests thorough surface area coverage. No obvious trends in sorption were apparent from observing surface area, pore size, sorbent loading, or f-potential of the functionalized materials. Taken cumulatively, this indicates that a similar amount of physisorption occurred between all functionalized MS materials, with the differences in U sorption due to specific binding by the grafted moieties. 4.2. Uranium sorption 4.2.1. General trends in sorption The amidoxime-functionalized materials, MSA-I through MSAIII, and the cyclic imine dioxime MSCA-I all adsorbed a statistically equivalent amount of U, though surface area, pore size distribution, loading, and surface charge were varied. Phosphonate-based sorbents in general showed poorer sorption, with the notable exception of MSPh-III. Loading, pore size, and surface area were all similar and relatively comparable with amidoxime-based sorbents. The glycolic acid sorbent, MSGly, had similar physical properties as the phosphonates and showed the lowest sorption of all materials. All phosphonate-based sorbents and MSGly possessed negative surface charge, whereas MSA-I was the only amidoxime-based sorbent with negative surface charge. It was expected a positive surface charge would encourage sorption of U in simulated seawater due to formation of a tetra-anionic uranyl carbonate complex. Though MSPh-III apparently does not fit this generalization, it is likely that incorporation of positive surface charge would be an important characteristic of an ideal sorbent, and could improve the sorption properties of both MSA-I and MSPh-III. 4.2.2. Sorption isotherms Analysis of the sorption isotherms (Figures 4 and 5) provides insight to the adsorption mechanism of U by the functional groups. In water, all materials except MSPh-III had similar saturation capacities. The unfunctionalized material MSU-H can only extract U by physisorption, strongly indicating nonspecific binding to be the primary mode of U extraction for most materials. In contrast, MSPh-III demonstrates a significantly higher saturation capacity compared to all other sorbents. While physisorption undoubtedly contributes, specific binding of U by the phosphonate functional group is the only way to rationalize the twofold increase in saturation capacity for MSPh-III. Measurement of U extraction from seawater simulant provides a more accurate assessment of the tested sorbents, not only because it more closely represents environmental conditions, but also because competing ions prevent the extent of U physisorption as observed in purified water. Using Giles’ classification [24], the isotherm for MSU-H in seawater simulant was type L-4, which occurs when solute affinity for the sorbent is inversely related to concentration due to competition for binding sites. The second plateau suggests either a reorganization of the U on the surface of MSU-H which exposes more binding sites, penetration of U into smaller pores of the support, or the formation of a second layer of U. Given the predominance of large pores in mesoporous silica and that the first plateau accounts for approximately 50% of the total U sorption, the formation of another monolayer is very likely. MSU-H has the only isotherm that displays a second plateau, and this phenomenon is continuous with U sorption occurring exclusively by physisorption. The modest correlation coefficient (R2 = 0.933) observed when modeled by the Langmuir isotherm supports this hypothesis, as a requirement for Langmuir modeling is the presence of a monolayer of sorbent.

29

Sorbents MSA-II, MSA-III, and MSPh-IV all displayed S-type isotherms, showing more facile adsorption as concentration increases. The S isotherm typically appears when there is strong competition for substrate sites, which may be due to Ca2+ and Na+ in the seawater simulant. Sorbent MSA-III and MSPh-IV possess two binding functions, suggesting that sorption of U to one binding site may encourage sorption on the adjacent site, which would yield the observed S-type isotherm. MSA-II, however, is monofunctional, and it is not immediately clear why increasing U concentration facilitates adsorption. The slight tailing observed upon N2 desorption in the BET isotherm (Figure S4) indicates the grafting of MSA-II may be less homogeneous on the pore surfaces than other sorbent groups. It is possible this non-uniform distribution may cause interference between sorbents on MSA-II, which is subsequently alleviated upon binding of one sorbent to U. As the S-type isotherm requires interaction between sorbent functional groups, this explicitly violates the underlying assumptions for Langmuir modeling, yielding poor R2 values (Table 3). Application of the Freudlich model, accounting for the non-idealized surfaces, substantially improves the correlation (Table 4). MSCA-I with the cyclic imide dioxime ligand had a rare C-type isotherm. This would require a constant ratio of binding sites to sorbent despite increasing concentration, which in turn demands that more sites are available as more solute is adsorbed. MSCA-I material showed a decrease in volume of gas adsorbed in the micropore region of the BET isotherm (Figure S6), as compared with MSU-H (Figure S2), which is indicative of sorbent grafting in both meso and micropores. U may rapidly be transported through all pores and channels of the sorbent, binding initially to the functional groups in the large pores, but also readily accessing the sorbents in the smaller pores as well. This could give the appearance of increased site availability as more solute is adsorbed, and saturation would be abrupt, resulting in the C-type isotherm. MSPh-III, a monophosporic acid functional group, had the highest saturation sorption capacity and exhibited an L-type isotherm. The L isotherm occurs when there is competition for binding sites with increasing concentration, resulting in a gradual plateau to saturation. The L-type isotherm is the traditional Langmuir isotherm, which occurs as a monolayer distribution is formed with no interaction between sorbents. As observed by fitting the isotherm to a Langmuir model, the R2 value of 0.979 was the highest of all sorbents investigated and indicative of strong correlation. It is noteworthy that when the isotherm was performed in purified water, an H-type isotherm was observed. The H-curve is a special case of the L-curve occurring when the solute has such high affinity for the sorbent that in dilute solutions it appears completely adsorbed. Better fittings of the U adsorption isotherms were obtained with the Freudlich model with most sorbents (Table 4). The constant k provides insight into the affinity of MS-based sorbents toward U. For the samples tested, the order of affinity was obtained as follows: MSPh-III > MSCA-I > MSU > MSA-III > MSPh-IV > MSA-II. Fitting with both the Langmuir model and the Freundlich model identified phosphonic acid- and imide dioxime-grafted MS materials as the most promising sorbents for U extraction from seawater. 4.3. Elution studies Elution studies with 0.01–1 M HCl indicated that while relatively weak HCl (0.1 M) elutes virtually all of the U from most materials, 1.0 M HCl was needed to desorb the significant fraction of U that remained in MSPh-III and MSCA-I after the 0.1 M HCl elution (Table S1). This indicates that the functional groups grafted in these two materials have stronger uranyl affinity than unfunctionalized MSU-H and other functionalized materials (Table S1). The U

30

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31

elution results are thus consistent with the fitting results of U sorption isotherms. 4.4. Phosphonate sorbents In seawater, MSPh-III exhibited the highest saturation capacity (66.7 lg U/mg sorbent), followed by MSCA-I (58.1 lg U/mg sorbent). This observation is not surprising; recently MS materials have been modified with a (2-diethylphosphatoethyl) triethoxysilane (DTPS) functional moiety imparting favorable characteristics for actinide extraction from water. Kleitz and coworkers functionalized a large pore 3-D cubic MS material (KIT-6) with DTPS [52], which exhibited excellent extraction properties towards U(VI) and Th(IV) (40 times higher than the commercially available EXC U/TEVA resin), and was superior to a 2-D structural equivalent (SBA-15). Shi and coworkers synthesized phosphonate(DTPS)modified MCM-41 type spherical nanoparticles (NP10) by a cocondensation method [39]. This material showed a high sorption capacity of 303 mg U/g sorbent and a fast equilibrium time of 30 min at slightly acidic pH at room temperature. The motif for UO22+ binding to phosphonates is known to be monodentate coordination through the phosphonyl oxygen [53]. The absence of hindering functional groups in the vicinity of the phosphonyl oxygen of MSPh-III likely played a major role in the superior sorption as compared to MSPh-II and MSPh-IV. Further investigations of the phosphonate-based MSPh-III and cyclic imine dioxime MSCA-I are needed to determine the effects of pore size, loading, and surface charge upon U sorption. 5. Conclusions A series of new mesoporous silica (MS) sorbents with amidoxime-, imide dioxime-, phosphonate-, and carboxylate-functional groups were examined for U sorption in basic water and artificial seawater. The MS sorbents exhibited a high U sorption capacity in water (>40 lg U/mg sorbent) with Langmuir isotherms suggesting a saturation U sorption capacity as high as 185.2 lg U/mg sorbent for the MSPh-III material. The U sorption capacity in artificial seawater is reduced by fourfold, and the maximum U sorption achieved was 12.1 lg U/mg sorbent. This loading capacity is comparable with that obtained for amidoxime fibers under similar conditions. Fitting of U sorption isotherms in artificial seawater with both the Langmuir model and the Freundlich model identified MSPh-III and MSCA-I as the best sorbents with the highest saturation capacities and greatest binding affinity. Additionally, U is quantitatively desorbed from loaded MS sorbents by washing with HCl at concentrations higher than 0.1 M. These results suggest that organo-functionalized MS materials are promising alternatives for U extraction from seawater. Acknowledgements This work was supported by the DOE Office of Nuclear Energy’s Nuclear Energy University Program (Sub-Contract - 20 #120427, Project - #3151). We thank Dr. Yatsandra Oyola at Oak Ridge National Laboratory for providing the amidoxime fiber. The authors thank H. Lau for experimental help. J.L.V.-E thanks the Carolina Postdoctoral Program for Faculty Diversity for a postdoctoral fellowship. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.micromeso.2013. 05.030.

References [1] M.I. Hoffert, K. Caldeira, A.K. Jain, E.F. Haites, L.D.D. Harvey, S.D. Potter, M.E. Schlesinger, S.H. Schneider, R.G. Watts, T.M.L. Wigley, D.J. Wuebbles, Nature 395 (1998) 881–884. [2] K.K. Mohanty, AIChE J. 49 (2003) 2454–2460. [3] S.J. DeCanio, A. Fremstad, Energy Policy 39 (2011) 1144–1153. [4] H. Sodaye, S. Nisan, C. Poletiko, S. Prabhakar, P.K. Tewari, Desalination 235 (2009) 9–32. [5] M. Kanno, J. Nucl. Sci. Technol. 21 (1984) 1–9. [6] R.V. Davies, J. Kennedy, R.W. McIlroy, R. Spence, K.M. Hill, Nature 203 (1964) 1110–1115. [7] S.-H. Choi, M.-S. Choi, Y.-T. Park, K.-P. Lee, H.-D. Kang, Radiat. Phys. Chem. 67 (2003) 387–390. [8] T. Hirotsu, S. Katoh, K. Sugasaka, N. Takai, M. Seno, T. Itagaki, Ind. Eng. Chem. Res. 26 (1987) 1970–1977. [9] M. Nogami, S.Y. Kim, N. Asanuma, Y. Ikeda, J. Alloy Compd. 374 (2004) 269– 271. [10] T. Suzuki, K. Saito, T. Sugo, H. Ogura, K. Oguma, Anal. Sci. 16 (2000) 429– 432. [11] K. Saito, T. Hori, S. Furusaki, T. Sugo, J. Okamoto, Ind. Eng. Chem. Res. 26 (1987) 1977–1981. [12] A. Goto, S. Morooka, M. Fukamachi, K. Kusakabe, T. Kago, Sep. Sci. Technol. 28 (1993) 2229–2235. [13] O. Hazer, S. Kartal, Talanta 82 (2010) 1974–1979. [14] T. Hirotsu, S. Katoh, K. Sugasaka, N. Takai, M. Seno, T. Itagaki, Sep. Sci. Technol. 22 (1987) 1725–1736. [15] T. Hirotsu, N. Takagi, S. Katoh, K. Sugasaka, N. Takai, M. Seno, T. Itagaki, Sep. Sci. Technol. 22 (1987) 2217–2227. [16] A. Zhang, T. Asakura, G. Uchiyama, React. Funct. Polym. 57 (2003) 67–76. [17] G.R. Choppin, Mar. Chem. 99 (2006) 83–92. [18] K. Saito, T. Miyauchi, J. Nucl. Sci. Technol. 19 (1982) 145–150. [19] Y.-Y. Park, S.-Y. Kim, J.-S. Kim, M. Harada, H. Tomiyasu, M. Nogami, Y. Ikeda, J. Nucl. Sci. Technol. 37 (2000) 344–348. [20] G. Tian, S.J. Teat, Z. Zhang, L. Rao, Dalton Trans. 41 (2012) 11579–11586. [21] S. Vukovic, L.A. Watson, S.O. Kang, R. Custelcean, B.P. Hay, Inorg. Chem. (Washington, DC, United States) 51 (2012) 3855–3859. [22] Y. Wan, D. Zhao, Chem. Rev. (Washington, DC, United States) 107 (2007) 2821– 2860. [23] F. Hoffmann, M. Cornelius, J. Morell, M. Froeba, Angew. Chem. Int. Ed. 45 (2006) 3216–3251. [24] C.H. Giles, T.H. MacEwan, S.N. Nakhwa, D. Smith, J. Chem. Soc. (Resumed) (1960) 3973–3993. [25] D. Pimentel, S. Cooperstein, H. Randell, D. Filiberto, S. Sorrentino, B. Kaye, C. Nicklin, J. Yagi, J. Brian, J. O’Hern, A. Habas, C. Weinstein, Hum. Ecol. 35 (2007) 653–668. [26] K.M.L. Taylor-Pashow, J. Della Rocca, R.C. Huxford, W. Lin, Chem. Commun. 46 (2010) 5832–5849. [27] B.P. Hay, T.K. Firman, Inorg. Chem. 41 (2002) 5502–5512. [28] S.O. Kang, S. Vukovic, R. Custelcean, B.P. Hay, Ind. Eng. Chem. Res. 51 (2012) 6619–6624. [29] S. Aksoyoglu, J. Radioanal. Nucl. Chem. 134 (1989) 393–403. [30] P. Makowski, X. Deschanels, A. Grandjean, D. Meyer, G. Toquer, F. Goettmann, New J. Chem. 36 (2012) 531–541. [31] A. Walcarius, L. Mercier, J. Mater. Chem. 20 (2010) 4478–4511. [32] G.E. Fryxell, Y. Lin, S. Fiskum, J.C. Birnbaum, H. Wu, K. Kemner, S. Kelly, Environ. Sci. Technol. 39 (2005) 1324–1331. [33] H.I. Lee, J.H. Kim, J.M. Kim, S. Kim, J.-N. Park, J.S. Hwang, J.-W. Yeon, Y. Jung, J. Nanosci. Nanotechnol. 10 (2010) 217–221. [34] Y. Liu, L. Yuan, Y. Yuan, J. Lan, Z. Li, Y. Feng, Y. Zhao, Z. Chai, W. Shi, J. Radioanal. Nucl. Chem. 292 (2012) 803–810. [35] M. Polanyi, Trans. Faraday Soc. 28 (1932) 316–333. [36] A. Rosowsky, R.A. Forsch, R.G. Moran, W. Kohler, J.H. Freisheim, J. Med. Chem. 31 (1988) 1326–1331. [37] C.R. Degenhardt, D.C. Burdsall, J. Org. Chem. 51 (1986) 3488–3490. [38] J.C. Birnbaum, B. Busche, Y. Lin, W.J. Shaw, G.E. Fryxell, Chem. Commun. (Cambridge, United Kingdom) 38 (2002) 1374–1375. [39] L.-Y. Yuan, Y.-L. Liu, W.-Q. Shi, Y.-L. Lv, J.-H. Lan, Y.-L. Zhao, Z.-F. Chai, Dalton Trans. 40 (2011) 7446–7453. [40] K. Vidya, S.E. Dapurkar, P. Selvam, S.K. Badamali, N.M. Gupta, Micropor. Mesopor. Mater. 50 (2001) 173–179. [41] H. Sepehrian, A. Malekynejad, Int. J. Ind. Chem. 2 (2011) 93–101. [42] K. Stamberg, K.A. Venkatesan, P.R. Vasudeva Rao, Colloids Surf. A 221 (2003) 149–162. [43] I. Langmuir, J. Am. Chem. Soc. 40 (1918) 1361–1402. [44] S. Aytas, M. Yurtlu, R. Donat, J. Hazard. Mater. 172 (2009) 667–674. [45] G. Tian, J. Geng, Y. Jin, C. Wang, S. Li, Z. Chen, H. Wang, Y. Zhao, S. Li, J. Hazard. Mater. 190 (2011) 442–450. [46] A. Walcarius, M. Etienne, B. Lebeau, Chem. Mater. 15 (2003) 2161–2173. [47] A. Walcarius, C. Delacôte, Chem. Mater. 15 (2003) 4181–4192. [48] J. Chung, J. Chun, J. Lee, S.H. Lee, Y.J. Lee, S.W. Hong, J. Hazard. Mater. 239–240 (2012) 183–191. ˇ erny´, Micropor. Mesopor. Mater. 142 (2011) 405–422. [49] I. Medvedˇ, R. C [50] M. Manzano, G. Lamberti, I. Galdi, M. Vallet-Regí, AAPS PharmSciTech 12 (2011) 1193–1199.

J.L. Vivero-Escoto et al. / Microporous and Mesoporous Materials 180 (2013) 22–31 [51] I. Galdi, G. Lamberti, Heat Mass Transfer 48 (2012) 519–528. [52] P.J. Lebed, K. de Souza, F. Bilodeau, D. Lariviere, F. Kleitz, Chem. Commun. (Cambridge, United Kingdom) 47 (2011) 11525–11527.

31

[53] C.-Z. Wang, J.-H. Lan, Y.-L. Zhao, Z.-F. Chai, Y.-Z. Wei, W.-Q. Shi, Inorg. Chem. 52 (2012) 196–203.