Mesoporous Zirconium Titanium Oxides. Part 3. Synthesis and

Nov 9, 2009 - Part 3. Synthesis and Adsorption. Properties of Unfunctionalized and ... process was by Grun et al.3 and then later by others.4-6 Imhof ...
0 downloads 0 Views 2MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Mesoporous Zirconium Titanium Oxides. Part 3. Synthesis and Adsorption Properties of Unfunctionalized and Phosphonate-Functionalized Hierarchical Polyacrylonitrile-F-127-Templated Beads G. Devlet Sizgek, Christopher S. Griffith, Erden Sizgek, and Vittorio Luca* Australian Nuclear Science and Technology Organisation, Institute of Materials Engineering, New Illawarra Road, Lucas Heights, NSW 2234, Australia Received April 29, 2009. Revised Manuscript Received August 17, 2009 A method is presented for the preparation of zirconium titanate mixed oxides in bead form having hierarchical pore structure. This method entailed the use of both preformed polyacrylonitrile (PAN) polymer beads and surfactants as templates. The templates were removed by calcination at temperatures below about 500 °C, resulting in mixed oxide beads with trimodal pore size distributions and interconnected pores. The pore size distributions as determined using nitrogen adsorption-desorption showed clear maxima at 4.5 and 45 nm length scales and also clear evidence of microporosity. The macroporous framework morphology was a replica of the PAN beads with radial structure. The mesoporous framework possessed wormhole-like pores with pore size of about 6 nm that was consistent with the F-127 triblock copolymer template used. The mixed oxide beads exhibited surface areas of 215 and 185 m2/g after calcination at 500 and 600 °C. Thermal stability up to 650 °C is unprecedented for bulk systems. The adsorption properties were characterized using uranyl as the target cation and the mass transport in the beads with the present hierarchical architectures has been shown to be exceptional. The beads were functionalized with 4-amino,1-hydroxy,1,1-bis-phosphonic acid (HABDP) and aminotris-methylene phosphonic acid (ATMP) and the adsorption properties for the extraction of uranyl sulfate complexes from acidic solution examined. Of the two molecules investigated, ATMP functionalization resulted in the best extraction efficiency with equilibrium uptake of about 90% of uranium available in solution between pH 1 and 2. The beads could potentially be utilized as catalysts, catalyst supports, adsorbents, and separation materials.

1. Introduction Supramolecular templating of porous metal oxides entails the use of an organic molecule or an assembly of molecules or macromolecules that act as templates for the assembly of the inorganic framework from a solution containing an assortment of inorganic molecular fragments. In the synthesis of zeolites, it is a single molecule that directs the structure of a particular zeolite type. Alternatively, multiple structure types can be prepared from the same amphiphilic molecule. Supramolecular templating using amphiphilic molecules was first demonstrated in the early nineties, and, generally, surfactant and surfactant-like aggregates act as templates or structure-directing agents for the assembly of mesoporous metal oxides with pore sizes in the 2-50 nm range.1 For still larger structures, researchers have naturally resorted to polymeric or macromolecular motifs.2 Indeed, inspiration for such chemical assembly has, as is often the case, come from nature. However, to be useful in many applications, such templated metal oxides often need to be obtainable in bead or granular form. One of the first reports of the preparation of micrometer sized mesoporous silicate spheres by a modification of Stober’s emulsion *Corresponding author. Present address: Comision Nacional de Energı´ a Atomica, Centro Atomico Constituyentes, Av. General Paz 1499,1650 San Martin, Provincia de Buenos Aires, Argentina. Tel. 54-11-6772 7018. E-mail: [email protected]. Phone: 54 11 6772 7018. (1) Davis, M. E. Nature 2002, 417, 813. (2) Stamberg, K.; Venkatesan, K. A.; Rao, P. R. V. Colloids Surf., A 2003, 221, 149. (3) Grun, M.; Lauer, I.; Unger, K. K. Adv. Mater. 1997, 9, 254. (4) Huo, Q.; Feng, J.; Schueth, F.; Stucky, G. D. Chem. Mater. 1997, 9, 14. (5) Buchel, G.; Grun, M.; Unger, K. K.; Matsumoto, A.; Tsutsumi, K. Supramol. Sci. 1998, 5, 253. (6) Stevens, W. J. J.; Mertens, M.; Mullens, S.; Thijs, I.; Van Tendeloo, G.; Cool, P.; Vansant, E. F. Microporous Mesoporous Mater. 2006, 93, 119. (7) Imhof, A.; Pine, D. J. Nature 1997, 389, 948.

11874 DOI: 10.1021/la9015299

process was by Grun et al.3 and then later by others.4-6 Imhof and Pines7 developed a fractionated emulsion templating procedure for the preparation of ordered macroporous titanate and other metal oxide microspheres with narrow size distribution. Aerosol and spray drying techniques have been reported for the production of mesoporous silicate microspheres when a suitable amphiphile or porogen was used as a template.8-10 Mircoemulsion techniques can also be employed to generate mesoporous silicate microspheres.11 It has been demonstrated that bimodal porosity can be engineered in such materials using a polymeric template and employing organic molecules that would usually produce a zeolite.12 Thus the concept of hierarchal porosity has developed. The use of preformed polymeric spheres as templates has also become popular and can potentially generate spherical materials with sizes exceeding hundreds of micrometers. For instance, latex and polymethyl methacrylate particles can be used to template a variety of ordered macroporous metal oxides.13 The use of micrometer-sized polymeric beads as templates of inorganic oxides such as silica and titania has been reported by numerous investigators.14-16 Caruso et al.17 combined this colloidal latex (8) Lu, Y. F.; Fan, H. Y.; Stump, A.; Ward, T. L.; Rieker, T.; Brinker, C. J. Nature 1999, 398, 223. (9) Andersson, N.; Alberius, P. C. A.; Pedersen, J. S.; Bergstrom, L. Microporous Mesoporous Mater. 2004, 72, 175. (10) Baccile, N.; Grosso, D.; Sanchez, C. J. Mater. Chem. 2003, 13, 3011. (11) Lee, Y. G.; Oh, C.; Yoo, S. K.; Koo, S. M.; Oh, S. G. Microporous Mesoporous Mater. 2005, 86, 134. (12) Holland, B. T.; Abrams, L.; Stein, A. J. Am. Chem. Soc. 1999, 121, 4308. (13) Holland, B. T.; Blanford, C. F.; Do, T.; Stein, A. Chem. Mater. 1999, 11, 795. (14) Meyer, U.; Larsson, A.; Hentze, H. P.; Caruso, R. A. Adv. Mater. 2002, 14, 1768. (15) Shchukin, D. G.; Caruso, R. A. Chem. Mater. 2004, 16, 2287. (16) Deshpande, A. S.; Shchukin, D. G.; Ustinovich, E.; Antonietti, M.; Caruso, R. A. Adv. Func. Mater 2005, 15, 239. (17) Caruso, F.; Caruso, R. A.; Mohwald, H. Science 1998, 282, 1111.

Published on Web 09/11/2009

Langmuir 2009, 25(19), 11874–11882

Sizgek et al.

particle templating approach with layer-by-layer self-assembly to generate submicrometer hybrid hollow spheres. Grosso et al.18 utilized multimodal templating of polymeric latex particles and surfactant supramolecular assembly together with aerosol and spray drying to produce very small transition metal oxides spheres with multi scale porosity. However, in this case, the collapse of the mesoporosity in this system was observed above 400 °C. All of the systems mentioned thus far typically produce very small particles with monomodal porosity. One of the first reports of large (millimeter) mesoporous silica spheres appears to have been by Huo et al.4 Cooper et al.19 have recently produced hierarchically porous uniform beads (diameters 1-1.5 mm) containing silica and single and selected metal oxide nanocrystalites using emulsion-templated porous poly(acrylamide) beads as scaffolds. More recently, Wang et al.20 reported a method of preparing metal oxide nanocrystalite macro beads (500 μm range) using cation-exchange resin with sulfonic acid group as templates and aqueous solution of metal salts as precursors. The main interest of the program of work of which this study forms a part is in the development of robust and selective adsorbents for the extraction of actinide species from nuclear waste or fuel reprocessing streams. The particular composition under study here has the advantage that, once the adsorbent is saturated, it can be disposed of by conversion to a stable zirconium titanate ceramic in which the same actinides remain immobilized. A major impediment to the deployment of novel ion exchange materials in this and related areas is the difficulty in preparing materials that, in their native state, are obtained in the form of very fine grained materials rather than in the form of beads or granules. For large-scale applications, these beads should have diameters of several hundreds of micrometers to millimeters, which should, at the same time, retain their inherent ion selectivity or recognition as well as display good mass transport characteristics for rapid kinetics. In adsorption and other applications, this requirement places constraints on the architecture of the porous network. Polymeric ion exchange resins and chelating resins in contrast can be routinely prepared in the form of beads displaying macro and mesoporosity. While useful for many applications, polymeric ion exchange resins have their limitations. Such materials often have to be applied in aggressive media, and, in this regard, the nuclear industry perhaps represents one of the most aggressive settings. Here, not only is chemical and mechanical stability important, but it is also essential to have good radiation stability. High radiation fields can rapidly break bonds responsible for cross-linking of resin backbones causing structural collapse or those of the functional groups resulting in loss of selectivity. Such limitations have probably been the principal causes for a tendency to move to liquid-liquid extraction techniques as opposed to solid-liquid extraction using ion exchange columns. Yet there is no dearth of industrial application of column chromatographic separations. The uranium mining industry offers one such industrial scale example. In catalysis applications, the need for materials with hierarchical porosity is also becoming increasingly recognized.21 If inorganic resins with properties similar to those of their organic counterparts could be engineered they would be expected to have superior mechanical and radiation resistance. These (18) Grosso, D.; de A.A. Soler-Illia, G. J.; Crepaldi, E. L.; Charleux, B.; Sanchez, C. Adv. Func. Mater 2003, 13, 37. (19) Zhang, H.; Hardy, G. C.; Khimyak, Y. Z.; Rosseinsky, M. J.; Cooper, A. I. Chem. Mater. 2004, 16, 4245. (20) Wang, M. L.; Wang, C. H.; Wang, W. J. Mater. Chem. 2007, 17, 2133. (21) Vantomme, A.; Leonard, A.; Yuan, Z. Y.; Su, B. A. L. Colloids Surf., A 2007, 300, 70.

Langmuir 2009, 25(19), 11874–11882

Article

would also have to have high selectivity and capacity for target species, display rapid kinetics, and be extremely stable and perhaps be readily disposed of once used. In order to achieve rapid kinetics, macroporosity may be required, while, in order to achieve high capacity and selectivity, there is a need to, at the same time, control porosity on a finer length scale. Indeed it has recently been demonstrated that larger mesopores give rise to faster mass transport.22 The ability to build in coordinating groups is also necessary. These requirements are difficult to meet in a single material and present an enormous chemical challenge. Polyacrylonitrile (PAN) has been promoted extensively as a universal polymeric support for nanocrystalline inorganic ion exchangers, especially within the nuclear industry. This is due to its physicochemical properties such as strong adhesive force with inorganic materials, good solubility in organic solvents, and chemical stability.23-25 PAN beads have a unique radial structure that apparently imbues the composite resin with excellent kinetic properties. In this case, the PAN component serves a template only and enables the preparation of different inorganic ionexchangers in granular form. From resin beads that are loaded with 80 wt % inorganic, we have measured kinetic properties to be as good as, if not better than, the neat nanoparticulate adsorbent in batch contact experiments.26 This is impressive indeed and is probably attributable to the macroporous radial structure. The aim of the present study was to extend the work in earlier parts of this series27 on xerogel powders and beads with monomodal porosity to beads with hierarchical pore structures having improved mass transport characteristics. To achieve these aims, multiple templates have been used, including PAN and the block copolymer mesogen (F-127) together with a cationic surfactant (palmitic acid) or nonionic surfactant (Tergitol) so as to generate, in addition to the macroporosity associated with the polymer template, mesoporosity and microporosity associated with the auxiliary surfactants. A further aim of the study was to evaluate the ion exchange and mass transfer properties of the hierarchical mixed oxide beads since such properties are critical in separation applications. For this, uranyl and uranyl sulfate have been used as target species since uranium is one of the most important and ubiquitous actinide elements. Uranyl is chemically extraordinarily robust owing to its strongly covalent U-O bonds.28 Therefore, it is a problematic contaminant, especially in uranium mining and processing sites. Finally, functionalization of the pore surfaces with polyphosphonic acid has been attempted, and preliminary investigations of the adsorption of uranyl sulfate complexes are presented.

2. Experimental Section 2.1. Synthesis. Materials. For the production of porous PAN beads as templates, PAN powder and dimethylsulfoxide (DMSO) were used. Zirconium titanate mixed oxide precursor was prepared by using titanium(IV) isopropoxide (97%), zirconium propoxide (70% in propanol), titanium tetrachloride, zirconium (22) Shin, Y. S.; Burleigh, M. C.; Dai, S.; Barnes, C. E.; Xue, Z. L. Radiochim. Acta 1999, 84, 37. (23) Sebesta, F. J. Radioanal. Nucl. Chem. 1997, 220, 77. (24) Xue, T. J.; McKinney, M. A.; Wilkie, C. A. Polym. Degrad. Stab. 1997, 58, 193. (25) Sebesta, F.; John, J. An Overview of the Development, Testing, and Application of Composite Absorbers; Los Alamos National Lab.: Los Alamos, NM, 1995; LA-12875-MS. (26) Griffith, C. S.; Luca, V.; Yee, P.; Sebesta, F. Sep. Sci. Technol. 2005, 40, 1781. (27) Griffith, C. S.; Sizgek, G. D.; Sizgek, E.; Scales, N.; Yee, P. J.; Luca, V. Langmuir 2008, 24, 12312; Sizgek, G. D.; Sizgek, E.; Griffith, C. S.; Luca, V. Langmuir 2008, 24, 12323. (28) Denning, R. G. J. Phys. Chem. A 2007, 111, 4125.

DOI: 10.1021/la9015299

11875

Article tetrachloride, and ethanol, Pluronic F-127 (block copolymer surfactant) was used as the mesogen, and palmitic acid was added to adjust the viscosity and maintain metal homogeneity. All chemicals were sourced from Sigma-Aldrich and were used as received. Preparation of Porous PAN Beads. PAN beads were prepared using a home-built droplet generator assembly. Initially, a PAN solution was prepared by dissolving PAN powder in DMSO (3.9% by weight). The PAN solution was placed in the reservoir of a pressure-driven solution feed system and was pumped to the nozzles (constructed from modified 21 gauge hypodermic syringe needles) of the vibrating droplet generator head at a constant flow rate. The flow of the PAN solution through the vibrating nozzles produced laminar flow, breaking into a stream of droplets with uniform size. The free-falling droplets were solidified to form PAN polymer beads on contact with a stirred bath. The bath was prepared by dissolving 1 g of Triton-X100 surfactant in 5 L of deionized water. The precipitated PAN beads were washed with deionized water until all DMSO was removed and subsequently dried in air. The PAN beads were then used as templates for the preparation of mixed oxide inorganic beads. Preparation of Mixed Oxide Precursor Solution. A typical synthesis of the precursor solution for infiltration was accomplished in two stages. First, a mixed metal chloride solution was prepared using the following precursor chemistry: ZrCl4/TiCl4/F127/EtOH/H2O with a mole ratio of 0.33:0.67:0.005:40:10. The final Ti/(ZrþTi) was 0.67 as in earlier work.29,30 The dissolution of TiCl4 and ZrCl4 in EtOH is strongly exothermic and generates HCl, and the resulting solution was strongly acidic. Second, a mixed metal alkoxide solution was prepared using the previously reported simple procedure.22,29 Titanium isopropoxide and zirconium propoxide with titanium mole fraction, Ti/(ZrþTi), of 0.67 were mixed. Palmitic acid was added to this solution to achieve a metal-to-surfactant ratio of 2. The alkoxide-surfactant mixture was kept at a 90 °C in a sealed Pyrex bottle for 12-16 h. After cooling to ambient temperature, the mixture was transferred to a Pyrex container and placed in a humidity- and temperaturecontrolled environment. Typically, a relative humidity of 50%, temperature of 30 °C, and air flow of 950 mL/min was used for volatilization of alcohol, and partial hydrolysis was used to form a partially condensed mixed metal oxide framework. Once the viscosity of the precursor solution reached between 1000 and 1500 cP, it was removed from the humidity chamber and placed into a sealed plastic bottle. Finally, this semi-hydrolyzed mixed metal alkoxide precursor solution was mixed with the initial highly acidic mixed chloride precursor solution to obtain the precursor solution containing a metal alkoxides-to-metal chloride molar ratio of 6.3. The resulting mother solution was highly acidic (pH: -0.65) with a viscosity of about 15 cP. Templating Procedure. About 100 g of porous PAN beads were infiltrated with 1200 mL of the mixed oxide precursor solution. The mixture was degassed using vacuum and ultrasonic agitation to accomplish complete filling of the pores and was kept sealed at 30-32 °C for about 24 h. Then the soaked PAN beads were separated from the excess solution using a strainer and were placed in the humidity chamber at a relative humidity of 50%, temperature of 32 °C, and air flow of 950 mL/min for the elimination of all solvent. Finally, the beads were calcined in air at 500 °C to remove the PAN and surfactant templates. 2.2. Polyphosphonate Functionalization of Beads. Surface functionalization of the beads was carried out with the two polyphosphonic acids 4-amino,1-hydroxy,1,1-bis-phosphonic acid (HABDP) and amino-tris-methylene phosphonic acid (ATMP). ATMP and HABDP were sourced from Aldrich as a hygroscopic powder and an aqueous solution (ATMP, 50% w/w), respectively, and used without further purification. (29) Luca, V.; Bertram, W. K.; Widjaja, J.; Mitchell, D. R. G.; Griffith, C. S.; Drabarek, E. Microporous Mesoporous Mater. 2007, 103, 123. (30) Luca, V.; Soler-Illia, G. J. A. A.; Angelome, P. C.; Steinberg, P. Y.; Drabarek, E.; Hanley, T. L. Microporous Mesoporous Mater. 2008, 118, 443.

11876 DOI: 10.1021/la9015299

Sizgek et al. Functionalization was typically undertaken by contacting a portion of the beads (1 g) with a solution (41 mM) of the phosphonic acid for 12 h in a sealed vessel. The pH of the mixture was not adjusted. The solid was then filtered, washed with deionized water (18.2 MΩ) until the filtrate was neutral, and then washed with about 20 mL of ethanol. Finally, the solid was dried at 70 °C for 12 h, then in vacuum (10 mmHg, 50 °C) for 12 h. 2.3. Characterization. A JEOL 6400 scanning electron microscope (SEM) fitted with an energy dispersive X-ray (EDX) analysis accessory was used to investigate bead morphologies. For SEM investigation, individual beads were sectioned with a razor to reveal internal structure and were mounted on aluminum stubs using carbon adhesive tape before being coated with thin film of gold. SEM investigation of uranium adsorbed beads was carried out by mounting into resin and polishing. Transmission electron microscopy (TEM) was conducted using a JEOL 2000FXII instrument operation at 200 keV to investigate the nonstructural features. The mixed oxide spheres were ground to fine powders before analysis, and the sample grids were prepared by sonicating the powdered samples in methanol and evaporating one drop of this suspension on a carbon-coated grid. X-ray powder diffraction (XRD) data of the samples were collected on a Scintag X1 diffractometer. The diffragtrograms were recorded using Cu KR radiation over a 2θ range of 0.7-40°. The specific surface area, mean pore diameter, and specific desorption pore volumes were determined by nitrogen physisorption at 77 K using a Micromeretics ASAP 2010 unit. Prior to measurements, the samples were degassed at 200 °C.

2.4. Uranium Batch Kinetic and Adsorption Experiments. Independent-batch kinetic experiments were carried out using batch equilibrium tests for equilibration with contact times of 5 min to 48 h. Each test used 0.1 g of mixed oxide beads and a uranium ion solution volume of 10 mL, giving a solution-to-solid ratio of 100. The initial concentration of uranium in solution was kept at 52 ppm (0.22 mM), and the pH was kept at 3.8. The uranium-containing solution and mixed oxide spheres were contacted in 20 mL screw-cap vials. In each series of experiments, samples were mixed on a shaker with a constant speed of 150 rpm. The effluent in each vial was withdrawn at suitable time intervals and filtered through 0.45 μm filters to remove any solid material. Uranium ion concentration was measured using inductively coupled plasma mass spectrometry (ICP-MS) on a PerkinElmerSCIEX Elan 6000 instrument. Nitric acid was also added to adjust the pH of the stock solution. For the adsorption isotherms, batch adsorption experiments with uranium stock solutions at different concentrations varying from 10 ppm (4.2  10-2 mM) to 1000 ppm (4.2 mM) were carried out. Uranyl stock solution was prepared by dissolving uranyl nitrate hexahydrate, UO2(NO3)2 3 6H2O (Merck), in Milli-Q water. The pH of each stock solution was adjusted to 3.8. Screw-cap vials for the uranyl adsorption tests were placed on a platform shaker for 48 h. After 48 h of contact, the effluent in each vial was filtered through 0.45 μm syringe filters and analyzed for uranium using ICP-MS.

3. Results and Discussions 3.1. Structural Properties. In this work we have utilized the well-developed radial pore structure and very high macroporosity (about 70%) of PAN polymer beads as templates to produce mixed oxide beads with hierarchical pore structure. A crosssectional view of the air-dried PAN polymer beads produced on a kilogram scale with required particle size using a home-built droplet generator setup are shown in Figure 1. The beads possessed the unique radial pore structure typical of PAN beads.31 (31) Moon, J.-K.; Kim, K.-W.; Jung, C.-H.; Shul, Y.-G.; Lee, E.-H. J. Radioanal. Nucl. Chem. 2000, 246, 299.

Langmuir 2009, 25(19), 11874–11882

Sizgek et al.

Article

An SEM image of the sectioned zirconium titanite bead after removal of the PAN and surfactant templates by calcination at 500 °C (Figure 2a) reveals that the resulting structure is a direct replica of the PAN polymer beads shown in Figure 1. No surface particulate deposits and no agglomeration of the beads have been observed, as demonstrated in the photographic image of Figure 2b. To be able to compare structural properties, the mixed oxide powder prepared from the metal oxide precursor solution without templating with PAN beads was subjected to the same experimental procedure followed by calcination at 500 °C. Nitrogen adsorption-desorption isotherms and the corresponding pore size distribution (PSD) for zirconium titanate powder prepared from precursor solution after drying and calcining at 500 °C (1), PAN polymer beads (2) and PAN-templated zirconium titanate beads calcined at 500 °C (3) are given in Figures 3a,b. These isotherms are all of type IV according to the IUPAC classification.32 The nitrogen adsorption-desorption isotherm of zirconium titanate powder exhibits one large hysteresis loop at intermediate partial pressures (P/P0 of 0.4 and 0.8), which is characteristic of the filling of the framework-confined mesopores.33 The PSD as determined using the BarrettJoyner-Halenda (BJH) technique indicates a mean pore diameter of 4.5 nm, confirming the mesoporosity of the structure. The nitrogen adsorption-desorption isotherm of the PAN polymer bead shows one distinct hysteresis loop in the high P/P0 region (above 0.8), usually indicating the existence of macropores, and the PSD is centered at about 90 nm. On the other hand, the isotherms of the PAN-templated zirconium titanate beads exhibited two distinct hysteresis loops. In other words the adsorption branch shows two well-defined capillary condensation steps, which can be attributed to two different pore sizes. Their calculated Brunauer-Emmett-Teller (BET) surface area at 500 °C was 215 m2/g, and the two peaks in the PSD (curve 3 in Figure 3b) have maxima at 4.5 and 45 nm. These two pore sizes are consequences of mesopores created by the inclusion of amphiphilic F-127 molecules in the precursor solution, while macropores are created by the PAN polymer template, resulting

in a hierarchical pore structure of the PAN templated mixed metal oxide beads. Also, the zirconium titanate powder and the PANtemplated zirconium titanate beads (Figure 3a) display significant volume adsorption at P/P0 of about 0.1. This has been attributed to microporosity present in the framework or the mesoporous wall structure.18,34 The influence of various thermal treatments on the XRD patterns of the mixed oxide beads calcined in air from 500 to 700 °C can be observed in Figure 4. The hierarchical zirconium titanate beads demonstrated high structural integrity and thermal stability upon calcination in air. They showed pronounced lowangle XRD peaks up to at least 600 °C as seen in Figure 4. No high-angle peaks were observed in the XRD patterns up to at least this temperature, making phase segregation and crystallization of titanium and zirconium oxides or zirconium-titanium mixed oxides unlikely, and confirming the attribution of the low-angle reflection to the mesophase. Calcination at 650 °C resulted in a significant diminution of the low-angle reflection, suggesting partial degradation of the mesostructure. A further increase in calcination up to 700 °C resulted in complete disappearance of the low-angle reflection and the appearance of two high angle XRD peaks at about 25.2 and 30.9 2θ, which can be attributed to the formation of crystalline titania (anatase) and zirconia phases. Nitrogen adsorption-desorption isotherms and PSDs of the PAN-templated mesoporous zirconium titanate beads calcined at various temperatures are given in Figure 5. For calcination temperatures of 500 and 600 °C, the shape of the isotherm and the PSDs are quite similar. For instance, at 500 °C, the BET surface area of the zirconium titanate beads calcined at 500 °C was 215 m2/g, and the total pore volume was 0.37 cm3/g. After calcining at 600 °C, the BET surface area and total pore volume decreased slightly to 185 m2/g and 0.35 cm3/g, respectively (Table 1). Calcining the material at 650 °C did not significantly alter the shape of the isotherm or the PSD. The hysteresis loop was still in the P/P0 range from 0.4 to 0.8, which is typical of mesoporous materials. However, the BET surface area and total pore volume decreased to 140 m2/g and 0.32 cm3/g, respectively. Clearly these changes represent the start of the collapse of the mesoporous structure. For the sample calcined at 700 °C, the BET surface area rapidly collapsed to 74 m2/g, and the hysteresis loop shifted to higher relative pressures indicating the formation of larger pores upon degradation of the mesoporous structure. The TEM of the materials treated at different temperatures are shown in Figure 6 and provide further insight into the structure and textural mesoporosity. The sample calcined at 500 °C possessed a wormhole-like pore structure. Analysis by TEM confirmed the retention of textural pores up to 600 °C and maintenance of a homogeneous composition on the scale of at least tens of nanometers. In contrast, both TEM and XRD confirmed degradation of the mesophase by 700 °C. The degradation of mesostructural texture can be associated with the crystallization of the inorganic framework such that pore walls can not accommodate the crystallites formed, resulting in pore collapse.30,35 EDX analysis indicated the presence of both titanium and zirconium in all samples, and their relative amounts corresponded well with the initial stoichiometry in the precursor used for synthesis indicating homogeneity at least on the length scale of the analysis beam (a few nanometers). The stability observed for the present materials up to 650 °C is excellent for bulk mesoporous titania-zirconia metal oxides.

(32) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603. (33) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area, and Porosity, 2nd ed.; Academic Press, Ltd.: London, 1982; pp 111-112.

(34) Ryoo, R.; Ko, C. H.; Kruk, M.; Antochshuk, V.; Jaroniec, M. J. Phys. Chem. B 2000, 104, 11465. (35) Crepaldi, E. L.; Soler-Illia, G. J. d. A. A.; Grosso, D.; Sanchez, C. New J. Chem. 2003, 27, 9.

Figure 1. SEM image of a cross-section of a PAN polymer template bead before precursor infiltration.

Langmuir 2009, 25(19), 11874–11882

DOI: 10.1021/la9015299

11877

Article

Sizgek et al.

Figure 2. (a) SEM image of a cross-section of the zirconium titanate bead after removal of PAN and surfactant templates and (b) photo of zirconium titanate mixed oxide beads (calcined at 500 °C).

Figure 4. Effect of various thermal treatments on the XRD patterns of zirconium titanate beads. The high angle reflections after calcination at 700 °C at about 25.2 o2θ and 30.9 2θ are due to titania and zirconia crystalline phases, respectively.

Figure 3. (a) Nitrogen adsorption-desorption isotherms with (b) corresponding BJH desorption dV/d log(D) pore volume plots of (1) zirconium titanate mixed oxide from the precursor solution calcined at 500 °C, (2) PAN polymer beads, and (3) PANtemplated zirconium titanate mixed oxide beads calcined at 500 °C.

Typically such materials show structural degradation by about 400 °C.18,36,37 The high thermal stability of the mesostructured phases is one of the most important requirements, especially in catalysis applications. This subject has been discussed in details by Carreon and Gulliants for mesostructured transition metal oxides.38 (36) Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D. Chem. Mater. 1999, 11, 2813. (37) Yuan, Z.-Y.; Ren, T.-Z.; Vantomme, A.; Su, B.-L. Chem. Mater. 2004, 16, 5096. (38) Carreon, M. A.; Guliants, V. V. Eur. J. Inorg. Chem. 2005, 2005, 27.

11878 DOI: 10.1021/la9015299

Figure 5. N2-BET adsorption-desorption isotherms with corresponding BJH desorption dV/d log(D) pore volume plot (inset) for PAN-templated zirconium titanate beads at various temperatures.

The present synthesis involves hydrolysis and condensation of mixed metal alkoxides with aqueous mixed metal chlorides in the presence mixed surfactants. During partial hydrolysis of mixed zirconium titanium oligomers containing alkoxide and long chain carboxylate ligands a zirconium titanium oxide network develops.27,29 The extension to these previous approaches Langmuir 2009, 25(19), 11874–11882

Sizgek et al.

Article

Figure 6. TEM images of meso zirconium titanate after various thermal treatments. Table 1. Porosity of the Hierarchical Zirconium Titanate Beads After Various Thermal Treatments temperature BET surface BJH desorption BJH desorption average (°C) area (m2 /g) pore volume (cm3/g) pore diameter (nm) 500 600 650 700

215 185 140 74

0.37 0.35 0.32 0.29

5.1 5.4 6.5 11.6

involves partially hydrolyzed mixed metal alkoxides to ethanol, water, F127, and the metal chloride mixture (highly acidic and containing metal chloro-alkoxides species39) such that enhanced cross-linking of the inorganic framework occurs. This is in part confirmed with the preservation of the mesoscopic structure of our materials at high temperatures exceeding those normally observed for similar mesoporous titanate materials. The advantages of using partially hydrolyzed alkoxides together with inorganic salts are the formation of homogeneous building blocks with mesostructure during partial hydrolysis of alkoxides and further control of hydrolysis/condensation rates of metal oxide species by addition of metal chlorides. Although chloride ions play a fundamental role in the formation of the hybrid mesostructure, they also act as counteranions, reducing the oxygen coordination numbers of the metal atoms.40 Therefore, having alkoxides as extra oxygen donors41 with the lower content of chloride in the mother solution could promote higher inorganic framework linkages. Using the present method, we were able to combine highly water sensitive, partially hydrolyzed metal alkoxides with aqueous metal chlorides without any precipitation. The HCl generated from the reaction of metal chlorides serves to mitigate the rate of hydrolysis-condensation. The resulting precursor solution was very transparent, homogeneous, and extremely stable. The precursor had low viscosity and was not water sensitive, which was well suited to PAN bead infiltration in large quantities. The PAN polymer bead production involves water, hence there is potential for residual moisture inside the beads even after drying. Therefore, water sensitive precursor solutions can cause immediate precipitation of the precursor during infiltration. Precursor solutions with low viscosity are also required to maximize penetration and filling of the pores of the template. 3.2. Uranyl (UO22þ) Adsorption on Unfunctionalized Beads. As mentioned in the introduction, the present study forms part of a program of research to develop porous zirconium titanate materials for the adsorption of actinide species. There have been numerous studies of the adsorption of uranium on various (39) Brinker, C. J.; Scherer, G. W. Sol-Gel Science; Academic Press: San Diego, CA, 1990. (40) Grosso, D.; de A. A. Soler-Illia, G. J.; Babonneau, F.; Sanchez, C.; Albouy, P.-A.; Brunet-Bruneau, A.; Balkenende, A. R. Adv. Mater. 2001, 13, 1085. (41) Tian, B.; Liu, X.; Tu, B.; Yu, C.; Fan, J.; Wang, L.; Xie, S.; Stucky, G. D.; Zhao, D. Nat. Mater. 2003, 2, 159. (42) Vidya, K.; Dapurkar, S. E.; Selvam, P.; Badamali, S. K.; Gupta, N. M. Microporous Mesoporous Mater. 2001, 50, 173.

Langmuir 2009, 25(19), 11874–11882

Figure 7. Kinetics of uranyl uptake (initial concentration 52 mg/L (0.22 mM) at pH 3.8) for (4) zirconium titanate mixed oxide powder (from mother solution), (() PAN templated zirconium titanate mixed oxide hierarchical beads, and (b) monomodal zirconium titanate spheres. Solid lines are pseudo-second-order fits.

unfunctionalized and functionalized mesoporous silicates2,42-45 and titanates.22 Batch U(VI) (in the form of UO22þ) kinetics and adsorption experiments were carried out to test the ion exchange/adsorption and mass transport properties of the bimodal zirconium titanate beads. The results were also compared with the kinetic properties of zirconium titanate xerogel powder produced from the same precursor solution and monomodal zirconium titanate beads with the same Zr mole fraction.22,27 The adsorption capacity of the materials was calculated based on the mass balance of uranyl ion expressed as qt ¼

VðCi -Ct Þ m

ð1Þ

Here, qt is the amount of uranyl adsorbed onto the unit mass of the beads (mmol/g) at time t; Ci and Ct are the concentrations of the uranyl in the initial solution (mmol/L) and in the aqueous phase after treatment for a certain period of time (mmol/L), V is the volume of the aqueous phase (L); and m is the mass of the beads (g) added to the solution. The uptake kinetics of uranyl ions measured at pH 3.8 at an initial uranyl concentration of 52 ppm (0.22 mM) for the hierarchical zirconium titanate beads ((), the zirconium titanate xerogel powder obtained from the precursor solution (4), and the monomodal zirconium titanate beads (b) are given in Figure 7. The best fit to the uranyl adsorption data was obtained using a pseudo-second-order kinetic model.46 The pseudo-second-order (43) Dudarko, O. A.; Goncharik, V. P.; Semenii, V. Ya.; Zub, Yu. L. Prot. Met. 2008, 44, 193. (44) Yantasee, W.; Lin, Y.; Fryxell, G. E.; Wang, Z. Electroanalysis 2004, 16, 870. (45) Ju, Y. H.; Webb, O. F.; Dai, S.; Lin, J. S.; Barnes, C. E. Ind. Eng. Chem. Res. 2000, 39, 550. (46) Ho, Y. S.; McKay, G. Water Res. 2000, 34, 735.

DOI: 10.1021/la9015299

11879

Article

Sizgek et al. Table 2. Pseudo-Second-Order Kinetic Constants and Characteristics of the Materials Shown in Figure 7 zirconium titanate powder (1)

PAN-templated zirconium titanate hierarchical beads (3)

monomodal zirconium titanate beads

0.021 1.51 6.74  10-4 0.99

0.021 0.88 4  10-4 0.96

0.016 0.08 2.05  10-5 0.96

>80 216 4.28 0.32 4.55

500-800 215 4.5 and 45 0.37 4.55

900-1000 340 2.8 0.31 4.6

Pseudo-Second-Order Constants qe (mmol/g) k h0 (initial sorption rate) R2 Characteristics of the Materials particle size (micrometer) surface area (m2/g) pore size (nm) pore volume (cm3/g) point of zero charge

Figure 8. Cross-sectional back-scattered electron SEM image of (a) monomodal and (b) hierarchical beads loaded with uranium.

reaction rate expression can be written as qt ¼

kqe 2 t ð1þkqe tÞ

ð2Þ

where k is the pseudo-second-order rate constant, qe is the amount adsorbed at equilibrium, and kqe2 (h0) is the initial sorption rate. Equation 2 was solved using a least-squares routine (downhillsimplex) available in EASY PLOT graphic software. While hierarchical zirconium titanate beads displayed fast uranium uptake rates similar to that of the zirconium titanate xerogel powders (with extraction of uranium close to 99%), the uranium uptake rate for monomodal zirconium titanate beads was about 1 order of magnitude slower (Table 2). The adsorption of solutes from solutions by porous adsorbents involves bulk transport of the solute in the solution, film diffusion of the solute, and finally diffusion of the solute inside the exchange bead (intraparticle transport). The rate controlling step is usually either diffusion of ions within the exchanger itself or across the thin film of solution surrounding the exchanger material.47 At high solute concentrations, intraparticle diffusion becomes a rate-determining step where particle size is quite important. The zirconium titanate xerogel powder consisting of particles below about 80 μm material has faster adsorption rates, because the uranium species have shorter average distances to diffuse through these relatively small particles. Therefore, obtaining very similar adsorption rates for zirconium titanate beads (500-800 μm) demonstrates that acces(47) Amphlett, C. B. Inorganic Ion Exchangers; Elsevier: New York, 1964.

11880 DOI: 10.1021/la9015299

sibility to adsorption sites in the mesopores has been achieved with the open hierarchical pore structure. In other words, the open structure of hierarchical beads has given rise to enhanced reaction kinetics as compared with those of monomodal zirconium titanate beads. To further investigate mass transport in the beads, the uranyl exchanged beads were sectioned and polished, and the location of the uranium was monitored using back-scattered electron imaging on an SEM microscope fitted with an energy dispersive analysis unit. These sections are shown in Figure 8. Areas of light and dark electron contrast can be easily discerned. Using EDX, each of these regions was analyzed, and this demonstrated that regions of dark contrast contained uranium, titanium, and zirconium. Regions of light contrast contained only Zr and Ti. These images indicate that, for the monomodal beads (Figure 8a), uranium penetrates only along cracks in the bead surfaces, whereas the hierarchical beads (Figure 8b) have extensive uranium-containing regions in the interior of the bead. There was no obvious sign of uranium precipitates larger than several micrometers that would have been visible using SEM. Since not all of the porosity in the hierarchical beads can be easily accessed via the bead periphery, it is easy to comprehend that there might be areas where U does not penetrate. Uranyl adsorption isotherms for the zirconium titanate hierarchical beads were obtained by increasing uranyl concentrations from 10 ppm (4.2  10-2 mM) to 1000 ppm (4.2 mM) with pH at 3.8 for each solution. Although successfully fitting isotherms to adsorption data does not necessarily provide information on the mechanism of adsorption, this procedure is useful for elucidating Langmuir 2009, 25(19), 11874–11882

Sizgek et al.

Article Table 3. Langmuir and Freundlich Constants from Adsorption Isotherms for Samples Langmuir qmax mmol/g PAN templated zirconium titanate bimodal beads

Figure 9. Uranyl adsorption isotherms at pH 3.8 for PANtemplated zirconium titanate mixed oxide hierarchical beads.

the adsorption mechanism and capacity as well as for facilitating comparisons. Experimental data were compared to Langmuir and Freundlich sorption isotherm models, and the adsorption isotherms are given in (Figure 9). The Langmuir isotherm is (eq 3) Cads ¼

qmax bCeq 1þbCeq

ð3Þ

where Ceq (mg/L) and Cads (mmol/g) are the equilibrium concentrations in the liquid and the solid phases, respectively. Here, qmax (mmol/g) is a Langmuir constant that expresses the maximum uptake, and b is also a Langmuir constant related to the energy and affinity of the sorbent. The Freundlich model is, on the other hand, based on the following relation (eq 4): 1=n Cads ¼ Cm Ceq

ð4Þ

where Cm and n are Freundlich parameters related to adsorption capacity and intensity, respectively. The Langmuir relation is a hyperbolic expression that demonstrates saturation in the surface monolayer and it is mathematically equivalent to the simplest form of the surface complexation model for a fixed pH. The Freundlich isotherm is an empirical expression that assumes that the adsorbent surface sites have an exponential distribution of binding energies. The Freundlich model can represent multilayer coverage situations in which bound surface molecules interact with each other. The adsorption data of Figure 9 overall are better described by the Freundlich rather than the Langmuir expression (Table 3). This suggests that the sorption of U(VI) on the zirconium titanate materials is concentration dependent, in turn suggesting secondary surface interactions between solution uranyl and adsorbed species, resulting in greater than monolayer surface coverage as uranium concentration is increased. Clearly a more detailed spectroscopic analysis would be required to confirm this. Although, as already stated, adsorption isotherms do not necessarily provide direct evidence of the adsorption mechanism, the good fit of the adsorption data to the Freundlich model suggests that the surface sites on the adsorbent have a range of binding energies, and it almost certainly excludes the possibility of a precipitative mechanism. Noteworthy is the value of the capacity of the present hierarchical beads. Assuming a Langmuir model, the saturation capacity has been determined to be 0.14 mmol U/g (33 mg U/g). This value of uranium adsorption capacity appears to be somewhat higher than that recently reported for a range of other adsorbent materials, including ion exchange resins.48 Typical ion exchange (48) Phillips, D. H.; Gu, B.; Watson, D. B.; Parmele, C. S. Water Res. 2008, 42, 260.

Langmuir 2009, 25(19), 11874–11882

0.14

b

Freundlich R2

Cm

1/n

R2

0.014 0.937 0.015 0.339 0.995

resins used in acid leach recovery of uranium from ore bodies are based on quarternary ammonium or tertiary pyridinium functionalities which, are able to bind uranyl sulfate complexes when sulfuric acid is the leaching medium. Uranyl speciation changes according to the pH and uranyl concentration of the solution. As the pH of the solution increases, the proportion of oligomeric species increases with increasing hydrolysis of the uranyl.49 The transition occurs at lower pH values, as the uranyl concentration of the solution is higher. For the concentrations and pH used in the present experiments, uranyl is predicted to be the predominant species (∼95%).50 We have measured the pH at point of zero charge for the zirconium titanate bimodal beads to be 4.55 (Table 2). This value is slightly below the commonly accepted value of 5.1 for pure anatase, as measured by us as well. Therefore, the surface of the present materials is slightly more basic in character. Since we carried out our measurement at a pH value (3.8) below the pHPZC of the zirconium titanate, its surface should carry a net positive charge. UO22þcation is the dominant species under the experimental conditions, and adsorption occurs on a positively charged surfaces despite the electrostatic repulsion. Such adsorption is difficult to rationalize on the basis of the surface charge of the oxide alone. Using spectroscopic techniques, others have suggested that uranyl cation forms inner-sphere surface complexes on silica and γ-alumina and TiO2 surfaces.51,52 A similar adsorption mechanism should be expected for the present zirconium titanate beads. 3.3. U(VI)-Sulfate Complex Adsorption on Polyphosphonate Functionalized Beads. While uranyl adsorbs strongly to the surfaces of the as-prepared zirconium titanate beads, at low pH values, in the presence of excess sulfate, the major uranium(VI) species are likely to be [UO2(SO4)2]2- or [UO2(SO4)4]4- as opposed to UO22þ.53 Speciation calculations based on the conditions employed in our studies, indicate that the former predominates. Given that the pH at the point of zero charge (pHPZC) for zirconium titanate materials is around 4.6, this suggests that at low pH values, the present materials will bear a positive surface charge and should therefore bind such uranyl sulfate complexes by a predominantly electrostatic mechanism. Here we present the kinetics and acid dependency of uranium sulfate complex adsorption for the native beads and those functionalized with the two polyphosphonic acids HABDP and ATMP. Figure 10 demonstrates that, by incorporating surface functionality using polyphosphonic acids, it is possible to dramatically enhance the adsorption of the uranyl sulfato complex as a function of increasing pH. Little “nonselective” adsorption of [UO2(SO4)2]2-, or for that matter [UO2]2þ, is displayed by the (49) Baes, C. F.; Mesmer, R. E. The Hydrolysis of Cations; Wiley: New York, 1976. (50) Chisholm-Brause, C. J.; Berg, J. M.; Matzner, R. A.; Morris, D. E. J. Colloid Interface Sci. 2001, 233, 38. (51) Sylwester, E. R.; Hudson, E. A.; Allen, P. G. Geochim. Cosmochim. Acta 2000, 64, 2431.  C. R. Chim. 2007, 10, 1078. (52) Drot, R.; Roques, J.; Simoni, E. (53) Moll, H.; Reich, T.; Hennig, C.; Rossberg, A.; Szab0 o1, Z.; Grenthe, I. Radiochim. Acta 2000, 88, 559.

DOI: 10.1021/la9015299

11881

Article

Sizgek et al.

4. Conclusions

Figure 10. Dependence of [UO2(SO4)2]2- uptake on pH for unfunctionalized hierarchal beads (O) and beads functionalized with pendent amino- (9, HABDP) or phosphonic acid (2, ATMP) moieties; V/m = 100 mL/g, [U] = 1 mM, [SO42-] = 0.25 M.

unfunctionalized titanate surfaces when in competition with either SO42- or H3Oþ. Selectivity for [UO2(SO4)2]2- under acidic conditions is therefore hypothesized to be due to the formation of an electrostatic interaction with the protonated amino group (-NH3þ) in the case of HABDP, while for ATMP we postulate that a ligand substitution reaction might involve the free phosphonic acid moiety of coordinated ATMP molecules. The kinetics of [UO2(SO4)2]2- adsorption on the polyphosphonate functionalized hierarchical titanate beads were found to be comparable to what was observed with UO22þ (Section 3.2). This general methodology of using polyphosphonic acid anchor groups has allowed us ready access to numerous acid-stable hybrid inorganic/organic phases with affinity for a wide range of elements, which we will be reported on in the near future. What separates the present hybrid titanate-based materials from silicate-based variants is that the present adsorbents are in a readily deployable form suitable for industrial scale, fixed-bed column chromatographic separations, especially in the nuclear field. In the latter regard, once saturated with actinides the present materials can, in principle, be converted to titanate ceramics (e.g., zirconolite) suitable for repository disposal. Of primary concern in the future development of these materials will be issues of hydrolytic and radiolytic stability of the functional phosphonate molecules and whether pore environments containing molecules with binding groups can display the desired selectivity in relevant applications.

11882 DOI: 10.1021/la9015299

We have demonstrated a robust procedure for synthesis of hierarchical titanium zirconate beads using PAN polymer beads together with surfactants as the templates for engineered porosity. A very high reproducibility (surface area, average pore diameter, pore volume, adsorption characteristics) was obtained from a number of samples produced in various batch sizes. In this method removal of excess precursor from the PAN template was facile as a result of the low viscosity of the precursor solution. No dense film formation and no agglomeration of the product beads were observed. Zirconium titanate beads were robust upon calcination and demonstrated high thermal stability, retaining wormhole mesoporosity at temperatures as high as 600 °C. This enhanced stability is probably due to the formation of a compositionally homogeneous and highly condensed structure as well as presence of surface hydroxyl groups, which condense further during drying and calcination. The novel hierarchical beads have already shown highly improved adsorption characteristics due to the coexistence of macro and mesopores and have a mass transfer efficiency comparable with xerogel powders in batch experiments. Surface functionalization of the hierarchical beads has been realized with functionalized phosphonic acids, which have imparted affinity to the materials for anionic uranyl sulfate complexes. Similarly enhanced selectivity could be displayed for other metal solution species that have an affinity for the amine and/or phosphonate binding group. The large spherical beads prepared here also provide ideal flow dynamics, provide for simple handling, and are easy to separate. The rate constant for adsorption on the hierarchical beads was an order of magnitude greater than that for monomodal beads and comparable to that of the xerogels powders. Therefore the hierarchical beads could offer excellent advantages as catalysis, catalysis support, adsorbents, and separation materials where optimization of diffusion, selectivity, and adsorption regimes is required. In principle the synthesis method could be extended to prepare a range of mixed oxide materials. For example, we are currently working on mesoporous yttria-stabilized zirconia (YSZ) material preparation. Acknowledgment. The authors are indebted to Elizabeth Drabarek of ANSTO for assistance with the nitrogen adsorption-desorption measurements.

Langmuir 2009, 25(19), 11874–11882