Contaminant Adsorption on Nanoscale Particles: Structural and

Oct 21, 2013 - Contaminant Adsorption on Nanoscale Particles: Structural and Theoretical Characterization of Cu2+ Bonding on the Surface of Keggin-Typ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Contaminant Adsorption on Nanoscale Particles: Structural and Theoretical Characterization of Cu2+ Bonding on the Surface of Keggin-Type Polyaluminum (Al30) Molecular Species Samangi Abeysinghe, Katie W. Corum, Diane L. Neff, Sara E. Mason,* and Tori Z. Forbes* Department of Chemistry, University of Iowa, Iowa City, Iowa 52242, United States S Supporting Information *

ABSTRACT: The adsorption of contaminants onto metal oxide surfaces with nanoscale Keggin-type structural topologies has been well established, but identification of the reactive sites and the exact binding mechanism are lacking. Polyaluminum species can be utilized as geochemical model compounds to provide molecular level details of the adsorption process. An Al30 Keggin-type species with two surface-bound Cu2+ cations (Cu2Al30-S) has been crystallized in the presence of disulfonate anions and structurally characterized by single-crystal X-ray diffraction. Density functional theory (DFT) calculations of aqueous molecular analogues for Cu2Al30-S suggest that the reactivity of Al30 toward Cu2+ and SO42− shows opposite trends in preferred adsorption site as a function of particle topology, with anions preferring the beltway and cations preferring the caps. The bonding competition was modeled using two stepwise reaction schemes that consider Cu2Al30-S formation through initial Cu2+ or SO42− adsorption. The associated DFT energetics and charge density analyses suggest that strong electrostatic interactions between SO42− and the beltway of Al30 play a vital role in governing where Cu2+ binds. The calculated electrostatic potential of Al30 provides a theoretical interpretation of the topology-dependent reactivity that is consistent with the present study as well as other results in the literature.

■

INTRODUCTION Poorly crystalline iron and aluminum oxyhydroxide nanoparticles with Keggin-type structural features are considered some of the most effective materials for adsorption of toxic heavy metals, radionuclides, and oxyanions.1,2 Ferrihydrite is a nanomineral that is ubiquitous in environmental systems and an effective scavenger for a wide range of contaminants.1 The structure of ferrihydrite has recently been the subject of debate within the literature, but a majority of the current evidence points to a topology with similarities to the mineral akdalaite (Al10O14(OH)2).3−5 This proposed core structure contains 20% tetrahedrally and 80% octahedrally coordinated iron with topological similarities to Keggin-type molecular species.3,6 Polyaluminum Keggin-type clusters, such as the ε-Al13, have also been identified as the building blocks of amorphous aluminum oxyhydroxide flocculants that precipitate from streams impacted acid mine drainage and are widely used as adsorbent in water treatment applications.2,7 Several previous studies have demonstrated that the 1 nm ε-Al13 clusters perform better than alum (KAl(SO4)2·nH2O) in clarifying potable water sources, but chemical agents containing the 2 nm Keggin-type Al30 species improves the removal efficiency of contaminants over a broader dosage and pH range.8−10 While the adsorptive capabilities of Keggin-type metal oxyhydroxides have been established, identification of specific © 2013 American Chemical Society

reactivity factors and mechanistic understanding of the process are lacking. The surface features of poorly crystalline nanominerals, such as 2-line ferrihydrite, are difficult to study by traditional X-ray and spectroscopic techniques due to lack of long-range ordering within the material. However, the use of geochemical model compounds provides a unique means to investigate adsorption at the molecular level. Polyaluminum Keggin-type species are an ideal experimental system because they form well-defined species that are between 1 and 2 nm in diameter that have been extensively characterized over the past 50 years.2,11 Casey and colleagues have pioneered the use of Keggin-type species to investigate exchange kinetics and have utilized these molecules to investigate similar reactions occurring at extended mineral surfaces.2 In addition, these species can be ordered into three-dimensional crystalline arrays that can be characterized using X-ray diffraction, providing concrete structural information regarding binding modes of the contaminant. The resulting experimental structures comprise a basis for atomistic calculations, providing an opportunity to investigate mechanistic details of contaminant adsorption onto nanoparticulate surfaces. Received: July 23, 2013 Revised: September 6, 2013 Published: October 21, 2013 14124

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

80.078(10)°, and γ = 62.392(9)°. H atoms associated with the naphthalene rings of the disulfonate anion were constrained using a riding model. Disorder was present for several of the 2,6-NDS anions due to free rotation about the S−C bond; therefore, O atoms associated with the sulfonate functional group (O52, O53, O55, and O74) were modeled as split sites with 50% occupancy. The presence of large void space (2113 Å3) within the crystalline lattice also resulted in the presence of disordered solvent (water) molecules. This diffuse electron density was modeled using the SQUEEZE command in the PLATON software,17 reducing the R1 value from 11.89% to 8.8% and accounting for 1090 electrons within the cavity. The crystallographic information files (CIFs) for Cu2Al30-S and additional details and results of the chemical characterization of the material (thermogravimetric analysis, SEM-EDS, and ICP-OES) are available in the Supporting Information. Computational Methodology. To better understand how the observed Cu2Al30 solid structure is formed, aqueous Al30 and several Cu2Al30 models were simulated using DFT and the COSMO dielectric continuum model18 as implemented in the DMol3 code developed by Delley.19,20 Aperiodic all-electron spin-polarized DFT calculations were performed using the generalized gradient approximation (GGA) of Perdew, Burke, and Ernzerhof (PBE)21 with a double-numeric-plus polarization atom-centered basis set. A real-space basis set cutoff of 3.5 Å was used for efficiency and did not have a significant effect on total energies, reaction energies, or the magnitude of forces, as further reported in the Supporting Information. For geometry optimizations, each system was embedded in the COSMO model such that the calculated gradients include forces between the solute and screening charges. All structures were optimized until an energy tolerance of 0.003 eV was obtained, which resulted in an average residual force magnitude of 0.0020 eV/Å on aluminum and oxygen atoms or 0.0005 eV/Å on hydrogen atoms. The oxygen functional group categorization and labeling scheme of Rustad,22 summarized in Figure 1a, is used to describe our Al30 and

We have recently begun a combined experimental and theoretical investigation on nanoscale Keggin-type clusters as a starting point for developing a molecular-level understanding of contaminant adsorption on small metal oxyhydroxide nanoparticles. Our initial study focuses on the adsorption of Cu2+ on the surface of Keggin-type polyaluminum species. Of the various heavy metal contaminants, Cu2+ is of particular interest as it is a constituent of widely used fungicides, pesticides, and algaecides.12,13 Herein, we have provided the synthesis and structural characterization of a Keggin-type Al30 molecule with two surface-bound Cu2+ atoms (Cu2Al30-S) and explored the adsorption process through density functional theory (DFT) studies.

■

EXPERIMENTAL METHODS

Synthesis of Cu2Al30. A stock solution of the partially hydrolyzed aluminum solution was prepared by heating a 25 mL of 0.25 M AlCl3 (6.25 mmol) solution to 80 °C in a water bath, followed by dropwise addition of 60 mL of 0.25 M NaOH (6.25 mmol) to a hydrolysis ratio (OH−/Al3+) of 2.4. A 7 mL aliquot of the cooled Al13 stock solution was mixed with 0.0349 g of Trizma base buffer and 0.0382 g of CuCl2. The resulting dark blue solution was stirred for 20 min to allow for complete dissolution of the starting materials and loaded in to a 23 mL Teflon-lined Parr reaction vessel. This solution was heated to 80 °C in a gravimetric oven for 24 h, cooled slowly to room temperature, and transferred to a glass scintillation vial. 3 mL of a 0.1 M 2,6napthalenedisulfonate (2,6-NDS) solution was added as a crystallization agent, and after 14 days of slow evaporation at room temperature, light-blue crystalline blades formed on the bottom of the reaction vial. Structural Characterization. Suitable crystals were separated from the mother liquor, coated in Infinium oil, and mounted on a Nonius Kappa CCD single crystal X-ray diffractometer equipped with Mo Kα radiation (λ = 0.7107 Å) and a low temperature cryostat. Data collection, cell refinement, data reduction, and absorption corrections were performed using Collect14 and APEX II15 software. The structure was solved using direct methods and refined on the basis of F2 for all unique data using the Bruker SHELXTL version 6.10 programs.16 Al, S, and Cu atoms were located in the direct methods solution, and the O and C atoms were identified in the difference Fourier maps calculated following refinement of the partial-structure models. Selected data collection parameters are given in Table 1, and relevant bond distances for Cu2Al30-S are provided in the Supporting Information, Table S1. Cu2Al30-S crystallized in the triclinic space group P-1 with a = 18.323(5) Å, b = 18.469(5) Å, c = 24.985(8) Å, α = 73.966(10)°, β =

Table 1. Selected Crystallographic Information for Cu2Al30-S a (Å) b (Å) c (Å)

18.323(5) 18.469(5) 24.985(8)

α (deg) β (deg)

73.966(10) 80.078(10)

γ (deg)

62.392(9)

V (Å3)

7191(4)

Z FW (g mol−1) space group

1 6218.5

ρcalc (g/cm3)

1.291

P1

μ (mm−1) F(000) crystal size (mm) θ range data collected completeness to θ = 26.07° reflections collected/ unique GOF on F2 final R indices [I > 2σ(I)] R indices (all data) largest peak and hole (Å3)

0.471 2798 0.192 × 0.18 × 0.095 1.48°−25.06° −21 < h < 21, −21 < k < 21, −29 < l < 29 99.2%

Figure 1. (a) Al30O8(OH)56(H2O)2618+ (Al30) structure, with aluminum represented as blue polyhedral, showing functional group categorization and labeling scheme of Rustad.34 (b) Al30 structure showing the tetrahedral aluminum in dark blue bound to four (a, b, c, and d) oxygen atoms.

135761/25292

Cu2Al30 models. As discussed in the Supporting Information, a stable structure for Al30 in the 18+ state could not be found. Five Al30 structures in a 16+ state, denoted Al3016+, were generated by deprotonating symmetry-equivalent 1-, 2-, 3-, 4-, and 5-ηH2O groups in the 18+ structure, and all were optimized to yield stable structures as confirmed by vibrational analysis. Similar to what is reported for classical molecular dynamics simulations of Al30,22 deprotonation of 2-

1.082 R1 = 0.0818, wR2 = 0.2456 R1 = 0.1025, wR2 = 0.2634 2.141 and −1.077

14125

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Figure 2. Optimized geometries of different configurations for Al30 in the 16+ state. From left to right structure correspond to the deprotonation of 1-ηH2O through 5-ηH2O. In the deprotonation of 2-ηH2O and 3-ηH2O, the resulting OH group a hydrogen bond from a nearby 4-ηH2O group, labeled, to form H3O2− bridges, indicated with dashed lines. The color scheme for the Al, O, and H atoms are blue, red, and gray, respectively. Atom labeling scheme is consistent throughout. ηH2O or 3-ηH2O sites results in OH groups pairing with a neighboring 4-ηH2O groups to form bridged H3O2− ligands as shown in Figure 2. Details of the optimized geometries of the five Al3016+ models are presented in Table 2, and the structures are shown in Figure 2.

Scheme 1. Formation of Cu2(SO4)2Al30 by Al3016+ Activation (Step 1), Inner-Sphere Adsorption of Cu(II) (Step 2), Followed by Outer-Sphere Adsorption of Sulfate (Step 3)

Table 2. Details of Experimental and Theoretical Al30 Geometries in Terms of the Tetrahedral Al−O Distancesa Al−Ox O on tetrahedral Al (Å) precipitated out of solution (expt) 1 2 3 4 5 Gaussian/DFT-PBE0+PCM a

Oa

Ob

Oc

Od

av

1.837

1.782

1.774

1.796

1.797

1.885 1.867 1.857 1.860 1.870 1.854

1.804 1.811 1.812 1.811 1.805 1.806

1.809 1.810 1.809 1.814 1.814 1.808

1.801 1.817 1.813 1.810 1.810 1.810

1.823 1.826 1.823 1.824 1.825 1.820

Scheme 2. Formation of Cu2(SO4)2Al30 Proceeding by Outer-Sphere Adsorption of 2 SO42− Anions by Al30 16+ (Step 1), Followed by Activation of (SO4)2Al30 (Step 2), and Ending with Inner-Sphere Adsorption of Cu(II)

The oxygen labeling scheme is shown in Figure 1b.

Benchmarking comparisons to the calculations carried out at the DFTGGA+COSMO level were made using the Gaussian computational package,23 employing a DFT-PBE024 exchange-correlation functional and the PCM solvation model25,26 with the Merz−Kollman atomic radii and using the TZVP basis set27,28 to expand the wave function solutions. Following in the style of Yang et al.,29 Table 2 reports the distances between the tetrahedral aluminum and the four oxygen atoms to which it is bound, using a labeling scheme defined in Figure 1b. Excellent agreement between Al3016+ structures modeled with the two computational methods is achieved. Further justification of the computational methodologies necessary for quantum-based modeling of aluminum hydroxide structures is made by considering literature results of DFT studies of aluminum hydroxide complexes (up to pentameric in size), which report a high level of structural agreement as a function of various basis sets and the choice of DFT exchangecorrelation functional.30 Additional comparisons between DMol3/ DFT-GGA+COSMO and Gaussian/DFT-PBE0+PCM were carried out on Cu2Al30 systems as reported further below or provided in the Supporting Information. As the synthesis of Cu2Al30-S involves charge balancing with a complexing agent, bonding competition between aqueous cations and anions was considered in the adsorption modeling. Different pathways from common starting reactants to end products were proposed, as summarized in Schemes 1 and 2. DFT reaction energies for theoretical mechanistic steps, Erxn, were calculated by using total energy

information for each reactant and product species, weighted by the appropriate stoichiometric coefficients. In order to simulate Schemes 1 and 2, the following Al30-based molecular aqueous species were modeled: Al3016+, Al3014+, (SO4)2Al30 (with a net charge 12+ or 10+), Cu2Al30 (net charge 18+), and Cu2(SO4)2Al30 (net charge 14+). A key difference between the modeled species and the experimental structure is the identity of the anion, which serves to charge-balance and aid in the crystallization of the synthesis product. The rationale for including two SO42− anions in the calculations as opposed to fully modeling all nine of the 2,6-NDS anions of the crystal structure was based on a compromise between model tractability and a physically sufficient representation of the synthesis experiment. Because of the possibility of conjugated ring− Al30 interactions, plus the higher computational cost of the greater number of atoms involved, calculations with 2,6-NDS were deemed prohibitive. The approximate aqueous analogue to the crystal structure was instead built by considering that in the solid hydrate, each adsorbed Cu2+ is associated with one sulfonate functional group. The model developed for the final product, denoted Cu2(SO4)2Al30, has one explicit SO42− group for each bound Cu2+. Mechanistically, sulfate addition was modeled as occurring after (Scheme 1) or before (Scheme 2) Cu2+ adsorption. Based on the crystal structure, SO42− is assumed to have an outer-sphere relationship to the Al30, with one 14126

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Figure 3. Structural characterization of the Cu2Al30-S compound reveals two Cu2+ cations in a square-planar configuration bonded to the Keggintype Al30 cation on either side of the central beltway. An additional water molecule and the sulfonate group of the 2,6-NDS anion are positioned approximately 2.6 Å away from the metal center. The color scheme for the Cu, Al, S, O, C, and H atoms are green, blue, orange, red, black, and gray, respectively. The atom labeling scheme is consistent throughout. oxygen atom in an axial position relative to adsorbed Cu2+, and these details are reproduced in the theoretical structures. To form a basis of comparison, several hypothetical Cu2Al30 and (SO4)2Al30 configurations were designed, starting from each of the five Al3016+ structures. The model (SO4)2Al30, Cu2Al30, and Cu2(SO4)2Al30 configurations are labeled in terms of the functional groups that form inner-sphere bonds to copper. Bidentate adsorption of Cu2+ groups at inversion symmetry-related sites was modeled in configurations involving distinct pairs of Al30 functional groups. The model that reflects the experimental structure, labeled as A, has Cu2+ reacting with type 3- and 4-ηH2O groups. The remaining hypothetical structures are defined as follows: B, two type 1-ηH2O groups; C, type 2- and 4-ηH2O groups; and D, two type 5-ηH2O groups. With the exception of the edge-sharing configuration in D, all of the structures have Cu(II) attaching to Al30 in a corner-sharing bidentate fashion. The A and C configurations bind copper through two types of ηH2O groups, which could result from either of two starting 16+ structures, and only the lowest energy results are reported. The electronic structure of the Al30-based models was analyzed in terms of the adsorption-induced change rearrangements or induced charge density Δρ. Δρ was calculated for different fragmentations of Cu2Al30 or Cu2(SO4)2Al30 to isolate specific bonding interactions. Additionally, visualization of the electrostatic potential, Vel, of Al30 was used to explain trends in DFT reactivity in terms of a governing physical property of the nanoparticle.

octahedrally coordinated Al3+ cations. From these components, five Al13 isomers can be constructed and are delineated by the number of hydroxyl edges shared between the trimeric units. The ε-Al13 isomer dominates in the initial partially hydrolyzed aluminum solution used in the synthesis of Cu2Al30-S and contains edge-sharing between all four of the [Al3(μ2OH)6(H2O)3] trimers. Upon heating, the δ-Al13 isomer is created when one trimer rotates by 60°, resulting in the transformation of the hydroxyl edges into six shared vertices. The larger Al30 species is then formed when two metastable δAl13 moieties are linked by four additional octahedrally coordinated Al3+ cations. Two Cu2+ cations are also bonded to the Al30 species in a bridging bidentate configuration to create the Cu2Al30-S polynuclear species. The Cu2+ atoms are located on opposite sides of the central beltway region of the cluster and share two hydroxyl groups with two neighboring octahedrally coordinated Al3+ cations. Bond lengths for the Cu2+-μ2-OH bridges are 1.921(3) and 1.948(3) Å, and two additional water molecules form bonds at 1.969(4) and 2.007(4) Å, resulting in a squareplanar geometry about the metal center. An additional water molecule and an O atom associated with a sulfonate functional group of the 2,6-NDS molecule are positioned 2.571 and 2.589 Å above and below Cu2+ cation, suggesting the formation of a distorted octahedron. These bond lengths imply the presence of Jahn−Teller distortions, which are commonly observed in Cu2+ species owing to the d9 configuration of the metal. Compounds with elongated axial Cu−O bonds have previously been reported in the literature with distances ranging between 2.1 and 2.7 Å, but the presence of orbital interactions between the metal center and ligand have not been confirmed computationally.31−37 The overall Cu2Al30-S polynuclear species is approximately 2 nm in diameter and possesses an 18+ formal charge based upon structural characterization of the compound.

■

RESULTS AND DISCUSSION Structural Description. The core feature of the Cu2Al30-S compound is the Al30 polynuclear species, which is built upon two smaller Keggin-type δ-Al13 moieties (Figure 3). Polyaluminum Al13 clusters possessing Keggin topologies are composed of a central tetrahedrally coordinated Al3+ cation surrounded by 12 octahedrally coordinated Al atoms. These 12 aluminum polyhedra are arranged into four planar [Al3(μ2OH)6(H2O)3] trimer through edge-sharing of hydroxyl groups and link to the tetrahedrally coordinated Al3+ cation through four μ4-O atoms. Bond lengths for the Al(O)4 tetrahedron range from 1.775(3) to 1.837(3) Å, and slightly longer distances (1.811(3) to 2.073(3) Å) are observed for the 14127

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Crystallization of the Cu2Al30-S species occurs through charge balancing with nine 2,6-NDS anions and hydrogen bonding between the deprotonated oxygen on disulfonate anions and the hydrogen atoms of the OH and H2O molecules of the aluminum polycations. The O−H···O distances associated with hydrogen bonding range between 2.66 and 2.87 Å, and additional π−π interactions may also aid in crystallization, as the naphthalene rings of the 2,6-NDS molecules are separated by approximately 3.5 Å. Solvent water molecules are also present in the void spaces, forming additional hydrogen bonds with the aluminum polycations and 2,6-NDS anions. The exact amount of water present in the structure was verified by thermogravimetric analysis, resulting in a final composition of [(Cu(H 2 O) 2 (Al 2 O 8 Al 2 8 (OH)60(H2O)22](2,6-NDS)9(H2O)52. Both Al13 isomers and Al30 molecule, along with related Al26 and Al32 species, have been previously isolated and characterized by a variety of solid-state techniques, usually with the addition of sulfate and selenate anions.2,38−41 Armstrong et al.42 reported the ΔHf,el for ε-Al13 clusters were exothermic based upon the initial oxide-based components, indicating that the formation of these clusters are stable in the presence of the SO42− and SeO42− crystallizing agents. We have recently utilized a supramolecular approach with the larger 2,6-NDS molecule that has resulted in more control over the crystallization of the δ-Al13 and Al30 molecules and the identification of the novel Al26 species.41 Bulk aluminum oxide materials are known to possess an affinity for Cu2+ cations and generally form an inner-sphere complex with the mineral surface.43 McBride et al. performed electron spin resonance (ESR) spectroscopy to provide insight into the adsorption of Cu2+ on the surface of gibbsite and observed square-planar coordination geometry about the metal center.44 The C4 rotation axis of the square planar Cu2+ species is modeled as perpendicular to the gibbsite surface and complexation is expected to occur through adjacent hydroxyl groups, or in a bridging bidentate manner. Similar results have been reported for amorphous aluminum hydroxide, Al2O3, boehmite, and clays, such as hydroxyl-Al3+-monotmorilonite and hectorite.45−50 Structural characterization of the polyaluminum species suggests that the central beltway of Al30 is the most reactive area of the cluster, as evidence by Cu2+ adsorption to symmetric sites within that region. This idea is substantiated with other structural models including Al32 clusters that contain two additional Al3+ cations bonding in an identical fashion to the same adsorption site.40,51 In both cases, the additional Al3+ cations form a ternary complex with either a sulfate anion or iminodiacetate ligand, decreasing the overall charge of the cluster. In addition, the Zn2Al32−NTA cluster has also been identified, which contains the Al32 polyaluminum cluster with two Zn2+ cations complexed by nitrilotriacetate (NTA) bonded in a monodentate manner within the reactive central beltway region.51 Computational Analysis. In Table 3, the geometry of the A model Cu2Al30 structure optimized at the DMol3/DFT-GGA +COSMO is compared to that obtained from optimization at the Gaussian/DFT-PBE0+PCM level. The agreement between the two theoretical structures is excellent, for example the distance between copper and the oxygen of its associated water ligands, d(Cu−OH2O), is 2.006 Å for both. Key bond distances and angles for all of the Cu2Al30 and Cu2(SO4)2Al30 structures

Table 3. Key Bond Distances in the Experimental Cu2Al30-S and Theoretical Cu2Al30/Cu2(SO4)2Al30 Modelsa av distances (Å) and angle (deg) comparisons Cu2Al30-S Cu2Al30 A Cu2(SO4)2Al30 A Cu2Al30 B Cu2(SO4)2Al30 B Cu2Al30 C Cu2(SO4)2Al30 C Cu2Al30 D Cu2(SO4)2Al30 D Gaussian/DFTPBE0+PCM for Cu2Al30 A

d (Cu− OAl30)

d (Cu− OH2O)

a (OAl30− Cu− OAl30)

a (OH2O− Cu− OH2O)

1.935 1.960 1.981 2.112 1.997 1.983 2.023 1.953 1.987 1.960

1.988 2.006 2.067 2.029 2.063 2.009 2.042 2.002 2.031 2.006

93.430 95.856 97.064 82.801 95.882 99.315 105.243 79.339 79.516 95.857

90.112 92.488 89.423 94.986 90.269 86.652 86.254 88.923 88.798 92.486

d (Cu··· OSO42−) 2.589 2.238 2.333 2.292 2.324

a Theoretical results are for structures optimized using DMol3/DFTGGA+COSMO unless otherwise labeled.

are given in Table 3, and the structures are shown in Figure 4. Specifically we report on the distance between the copper and oxygen functional groups of Al30 d(Cu−OAl30) and d(Cu− OH2O), as well as the angle of the copper and the two oxygen functional groups of Al30 a(OAl30−Cu−OAl30), and the angle of the copper and the two oxygen atoms of water ligands a(OH2O− Cu−OH2O). For Cu2(SO4)2Al30, the distance between the adsorbed copper and the axial sulfate oxygen atom, d(Cu··· OSO42−), is also given and is in great agreement (Figure 5). In the Cu2(SO4)2Al30 optimized geometries, d(Cu···OSO42−) is 2.238 Å for the A structure, while the corresponding distance in the experimental structure is 2.589 Å. As all other Cu−O distances are in good agreement between theory and experiment, this difference in d(Cu···OSO42−) distance is attributed to not simulating the full 2,6-NDS counterion. Overall, the DFT energetics provide theoretical evidence for preferential formation of an aqueous Cu2(SO4)2Al30 species with copper adsorbed in the same fashion as in the crystal structure. The total energy information on the Cu2Al30 structures shows that the A and D models are nearly degenerate, with B and C higher in energy by 0.43 and 0.81 eV, respectively. The DFT values for Erxn are reported in Tables 4 and 5, including the overall and stepwise values for Erxn. No stable Al3014+ geometry leading to the final D structure could be found, which further confirms the relative stability of the A model. Erxn values depending on the total energy of the Al3014+ D structure are unavailable and are left blank in Tables 4 and 5. To summarize the two model pathways from starting Al3016+ to final Cu2(SO4)2Al30, the Scheme 1 mechanism begins with deprotonation of 2 η-H2O groups at each pair of symmetryrelated sites to form Al3014+, followed by reaction with Cu(H2O)4 to form Cu2Al30 (an inner-sphere complex) and subsequent outer-sphere SO 4 2− adsorption to yield Cu2(SO4)2Al30. Alternatively, Scheme 2 starts with SO42− and Al3016+ reacting to form outer-sphere (SO4)2Al30, followed by deprotonation of the respective reacting oxygen functional groups and ending with the inner-sphere adsorption of Cu(H2O)4 to produce Cu2(SO4)2Al30. Chemically speaking, the two schemes describe the same net reaction but are broken down into different elementary steps that allow us to consider 14128

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Figure 4. Optimized geometries of model Cu2Al30 configurations. From left to right A, B, C, and D structures (defined in the text) are shown.

how cation and anion adsorption compete or cooperate in the formation of Cu2(SO4)2Al30. Going through the results for Scheme 1, the DFT values of Erxn step 1 Scheme 1 show that the reacting oxygen functional groups of the A configuration have a lower barrier to activation through deprotonation, suggesting that the 4-ηH2O groups are the most acidic. However, in step 2 of Scheme 1, the A model subsequently shows the least favorable uptake of Cu2+. The same qualitative behavior is seen in the energetics of the C model. In contrast, the B model exhibits a relatively high endothermic activation (step 1 Scheme 1) compared to A/C, while copper addition (step 2 of Scheme 2) is predicted to be significantly more exothermic at B than in the A/C structures. For step 3 of Scheme 1, grouping of like trends between A/C compared to B persists, with the former showing more favorable reactivity toward SO42− than the latter. The trends in Erxn for Scheme 2 pose contrast to those reported for Scheme 1. Erxn values for step 1 Scheme 2 span a range of 2.44 eV, with the B model Al3016+ showing the lowest reactivity toward sulfate. Subsequently in step 2 of Scheme 2, the B model exhibits the lowest energetic cost to deprotonate (as opposed to the A model in step 1 of Scheme 1). In step 3 of Scheme 2, Cu 2+ adsorption is the most favorable in configuration A, which again counters the analogous step in Scheme 1 for which A was the least reactive toward Cu2+. Another noteworthy comparison of the DFT Erxn values is the fact that the outer-sphere adsorption energy of SO42− (step 3 Scheme 1, step 1 Scheme 2) is greater in magnitude than the inner-sphere adsorption energies of Cu2+ (step 2 Scheme 1, step 3 Scheme 2). This goes against the chemically intuitive prediction that chemisorption is stronger than physisorption and will be later discussed by relating the local Al30 geometry of beltway sites to electrostatic arguments. A fundamental question posed by the experimental structure is whether the interaction between adsorbed Cu2+and SO42− is covalent or electrostatic in nature. The position of one of the SO42− oxygen atoms appears to be axial to Cu2+, intuitively suggesting a covalent relationship. However, the isosurface of ΔρCuSO4 (calculated as the difference of ρAl30/Cu/SO4 − ρAl30/Cu − ρSO4), shown in Figure 6, exhibits a void in between the Cu and “axial” sulfate oxygen in the B model. In the A model, the isosurface between the Cu and “axial” sulfate oxygen is

Figure 5. Optimized geometry of the Cu2(SO4)2Al30-A structure.

Table 4. DFT Reaction Energies, in eV, for the Scheme 1 Formation of Cu2(SO4)2Al30 Scheme 1 config

eq 1

eq 2

eq 3

total

A B C D

1.37 3.03 2.19

−0.43 −2.52 −0.44

−5.23 −2.60 −5.36 −4.88

−4.29 −2.09 −3.61 −4.56

Table 5. DFT Reaction Energies, in eV, for the Scheme 2 Formation of Cu2(SO4)2Al30 Scheme 2 config

eq 1

eq 2

eq 3

total

A B C D

−3.93 −1.75 −4.19 −3.78

2.74 1.28 2.34

−3.10 −1.62 −1.76

−4.29 −2.09 −3.61 −4.56

14129

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Figure 6. Induced density ΔρCuSO4 (defined in the text) for Cu2(SO4)2Al30 structures. From left to right A, B, C, and D. Charge loss is shown in black, and charge gain is shown in yellow. To better show the isosurfaces, bonds and H atoms are not shown.

indicative of charge loss, as opposed to the buildup of charge that would characterize covalent interaction. On the other hand, oxygen atoms of sulfate not in the Cu axial site exhibit electrostatic interactions with the water ligands of bound Cu2+ and/or with functional groups of Al30. Thus, the occupation of the Cu2+ axial site by a sulfate oxygen atom seems coincidental, with sulfate interacting predominantly through electrostatic interactions. While the net reaction to form Cu2(SO4)2Al30 is the same for both schemes, reversing the order of anion/cation adsorption steps reveals opposing trends in DFT reaction energies. The various trends in Erxn initially seem to be opposing but can go on to be explained consistently through the single physical property of electrostatic potential. Visualization of Vel of Al30 mapped onto an electron density isosurface is presented in Figure 7. Owing to the 16+ charge of the molecule, Vel > 0 throughout, and qualitatively the red surfaces represent regions with higher values of Vel while the blue surfaces represent regions with lower values. The highest values are found in the beltway, while the caps of Al30 exhibit the lowest. The driving force for Vel variation on the Al30 exterior is likely due to the cooperative contributions of inwardly folded oxygen functional groups, or “pockets”, in the beltway, while the distribution of functional groups on the caps of Al30 is more consistent with what is seen in extended surfaces. Similar “pockets” in zeolites have been linked to highly reactive surface sites in EU-1 and may contribute to enhanced catalytic processes, such as the hydroisomerization of n-hexane contained within mordenite.52,53 Concave curvature has also been implicated in the enhanced interactions with gaseous molecules on the surface of TiO2 nanotubes and may play a minor role on the adsorption of water vapor within these materials.54,55 The outer-sphere adsorption of SO42− onto Al30 (as in step 1 of Scheme 2) is expected to be electrostatic in nature. This was confirmed by visualizing the sulfate adsorption-induced charge density ΔρSO4 taken as ρAl30/SO4 − ρAl30 − ρSO4, shown for (SO4)2Al30 A-D in Figure 8. The large volume enclosed by the negative ΔρSO4 isosurface surrounding the sulfate group indicates that electron density goes from the anion into the Al30 core, and the alternating positive/negative ΔρSO4 isosurface

Figure 7. Electrostatic potential of Al3016+ mapped onto a charge density isosurface. The electrostatic potential is positive throughout, with higher values shown in red and lower values shown in blue. The structure shown is the configuration that is the initial structure for the A and C sites.

volumes between the sulfate and functional groups of Al30 is characteristic of Coulombic interactions between oppositely charged groups. Furthermore, there is an obvious correlation between the plots of ΔρSO4 and Erxn for step 1, Scheme 2: Adsorption of SO42− in the B model results in the least favorable value of Erxn (−1.75 eV), and ΔρSO4 in Figure 8 visualizes the relatively weak sulfate−Al30 interaction through the smaller ΔρSO4 volumes enclosed by the isosurfaces. A physical interpretation is that the flat, surface-like topology at the reactive sites in the B model prevents more extensive SO4− Al30 interaction. The configurations with reactive sites in the beltway are A, C, and D, for which the corresponding values of Erxn are −3.93, −4.19, and −3.78 eV. In this set, greater ΔρSO4 isosurface volumes reflect the greater extent of SO4−Al30 interaction. The limitation of Coulombic Al30−SO42− inter14130

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

Figure 8. Induced density ΔρSO4 for (SO4)2Al30 configurations. From left to right A, B, C, and D. Charge loss is shown in black, and charge gain is shown in yllow. To better show the isosurfaces, bonds and H atoms are not shown.

action sterically imposed by the reactive functional group topology in the B model also manifests in step 3 Scheme 1, in which sulfate adsorption is modeled after Cu2+adsorption. As in Scheme 2, the sulfate adsorption in Scheme 1 is the least favorable in the B model, and the associated plots of ΔρCuSO4 (Figure 6) visualize the lesser extent of SO 4 −Cu 2 Al 30 interaction in that structure. The question of how inner-sphere cation adsorption depends on the Al30 adsorption site in the absence of SO42− is addressed by comparing the values of Erxn for step 2 of Scheme 1. As summarized above, Erxn for the specific adsorption of Cu2+ is the least favorable for the A configuration (−0.43 eV). On the other hand, Erxn is the most favorable for the B configuration (−2.52 eV). This demonstrates a reversal in sitewise reactivity trends of Al30 toward SO42− vs Cu2+. As the D configuration has the unique trait of edge-sharing Cu2+, and also produced unstable intermediate structures which prevented reliable calculations of Erxn for several steps, we form an interpretation based on comparison of the A, B, and C models. The calculated Vel correlates with Erxn such that stronger Cu2+ adsorption is associated with the less positive Vel at the ends of Al30. The preference for anion adsorption in the beltway again manifests in step 3 of Scheme 1 (outer-sphere adsorption of SO42− onto Cu2Al30), which is predicted to be more favorable in configurations A and C. Common to both proposed schemes is the need to activate the Al30 through deprotonation of ηH2O groups as depicted in step 1 of Scheme 1 and step 2 of Scheme 2. In both schemes, the activation step is endothermic. For Scheme 1, deprotonation is the most favorable in the A configuration with reactive sites in the beltway, while in Scheme 2, the configuration with reactive sites on the Al30 caps, B, has the lowest energetic cost to deprotonation. Though opposing, the trends in the deprotonation step of each scheme conform to the emerging overall description of Al30 reactivity: The above comparisons of Cu2+ and SO42− adsorption show that the functional groups of the Al30 cap show greater affinity for cations, while those of the beltway show greater affinity for anions. An alternative but consistent statement is that the strong SO4−Al30 interaction in the A model impedes subsequent deprotonation of (SO4)2Al30

in step 2 of Scheme 2 relative to the deprotonation of Al30 in step 1 of Scheme 1. The electronic structure analysis suggests that, regardless of the order of cation and anion adsorption, the formation of Cu2Al30-S is influenced by cooperative effects between the adsorption processes as well as the topological variation in the electrostatic properties of Al30. Experimental data regarding the impact of counterions on bulk Al- and Fe-oxides suggest similar enhancement of adsorption of a variety of metals (Cd, Cu, Pb, Cd, and Zn), particularly in the presence of sulfate and phosphate anions.43 Sulfate itself is expected to form outersphere complexes with the mineral surface, although innersphere interactions may occur with decreasing pH and increasing sulfate concentrations in solution. 56−58 The enhanced metal adsorption on the surface of Al- and Fe-oxides in the presence of anions, such as sulfate, has been described in three ways:59 (1) electrostatic enhancement: reduction of the positive surface charge by initial adsorption of the anion, resulting in a more attractive surface for the metal cation; (2) formation of ternary cation−anion−surface complex: a cation− anion complex binds to the surface by inner- or outer-sphere interactions; (3) surface precipitation: addition of the crystalline lattice that contains the metal species. Surface precipitation can be ruled out in the Cu2Al30 complex, but the interaction between the Cu2+ and SO42− could be modeled by either electrostatic enhancement or formation of ternary complex. Elongation of the Cu−O−SO3 bond in Cu2Al30-S and limited covalent interactions between Cu2+and sulfate as revealed in the analysis of ΔρSO4 suggests electrostatic enhancement to be the major mechanism for the synergetic metal-anion adsorption.60 The electronic structure analysis not only elucidates that the SO4−Al30 and the SO4−Cu2Al30 interactions are dominated by electrostatics but also shows how the concavity of the Al30 beltway allows for enhanced anion−Al30 interactions with the inwardly folded oxygen functional groups. Collecting our interpretations of the individual trends in Erxn in terms of adsorption site, a universal interpretation emerges: In Al30, Vel is less positive at the caps and more positive in the beltway, which consistently explains the relative ease of 14131

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

deprotonating the η-H2O groups in the A model (as in step 1 of Scheme 1) or those in the B model when sulfate is preadsorbed (as in step 2 of Scheme 2), the less favorable SO42− adsorption in the B configuration with or without Cu2+ (as in step 3 of Scheme 1 and step 1 of Scheme 2, respectively), and the less favorable Cu2+ adsorption in the A model relative to the B model in step 2 of Scheme 1, etc. Thus, a concise way to summarize the trends in Erxn is that the variation in Vel between the cap and beltway regions is sufficient to result in opposite trends in cation/anion affinity at adsorption sites in the two topologically distinct regions (beltway and caps) of Al30. The DFT-based interpretation of Al30 reactivity helps to interpret the factors governing formation of the Cu2−Al30-S structure synthesized here and also relate to the results of classical molecular dynamics simulations by Rustad,22 which show greater water and perchlorate ion distributions in the beltway region. Our results have a direct impact on the current understanding of contaminant transport within the environment, particularly the role that counterions can play in adsorption processes. Competitive adsorption processes are key for predicting uptake of metals, oxyanions, and organic ligands and play a major role in element cycling within natural systems.43 A complete understanding of favorable interactions between cation and anions on mineral surfaces can lead to increased control over contaminant mobility and enhanced environmental remediation efforts. As highlighted in ref 43, a more complete understanding of the mechanisms for these interactions at the molecular level is critical and requires an interdisciplinary research approach. Our current study demonstrates a powerful synergy between experimental and computational approaches that lead to unique insights into important geochemical phenomena.

computational methods. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Corresponding Authors

*E-mail [email protected] (T.Z.F.). *E-mail [email protected] (S.E.M.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS T.Z.F. acknowledges the University of Iowa Vice President of Research Mathematical and Physical Science Funding program for partial support of this work. T.Z.F. and S.E.M. thank the College of Liberal Arts and Sciences for additional funding. We thank the UI Central Microscopy Facility for training and use of the SEM and Paul Mueller and Dr. Sarah Larsen, UI Department of Chemistry, for providing the ICP-OES data. We also thank Randall T. Holmes, an undergraduate summer intern, for assisting in the analysis and discussion of computational results. R.T.H. was supported by the Center for Global and Regional Environmental Research at the University of Iowa. Many of the calculations contributing to this work were carried out through the Arctic Region Supercomputing Center at the University of Alaska Fairbanks.

■

REFERENCES

(1) Brown, G. E.; Calas, G. Environmental mineralogy - Understanding element behavior in ecosystems. C. R. Geosci. 2011, 343, 90− 112. (2) Casey, W. H. Large aqueous aluminum hydroxide molecules. Chem. Rev. 2006, 105, 1−16. (3) Michel, F. M.; Ehm, L.; Antao, S. M.; Lee, P. L.; Chupas, P. J.; Liu, G.; Strongin, D. R.; Schoonen, M. A. A.; Phillips, B. L.; Parise, J. B. The structure of ferrihydrite, a nanocrystalline material. Science 2007, 316, 1726−1729. (4) Peak, D.; Regier, T. Direct observation of tetrahedrally coordinated Fe(III) in ferrihydrite. Environ. Sci. Technol. 2012, 46, 3163−3168. (5) Harrington, R.; Hausner, D. B.; Xu, W. Q.; Bhandari, N.; Michel, F. M.; Brown, G. E.; Strongin, D. R.; Parise, J. B. Neutron pair distribution function study of two-line ferrihydrite. Environ. Sci. Technol. 2011, 45, 9883−9890. (6) Michel, F. M.; Ehm, L.; Lui, G.; Han, W. Q.; Antao, S. M.; Chapas, P. J.; Lee, P. L.; Knorr, K.; Eulert, H.; Kim, J.; Grey, C. P.; Celestian, A. J.; Gillow, J.; Schoonen, M. A. A.; Strongin, D. R.; Parise, J. B. Similarities in 2- and 6-line ferrihydrite based on pair distribution fucntion analysis of X-ray total scattering. Chem. Mater. 2007, 19, 1489−1496. (7) Furrer, G.; Phillips, B. L.; Ulrich, K. U.; Pothig, R.; Casey, W. H. The origin of aluminum flocs in polluted streams. Science 2002, 297, 2245−2247. (8) Mertens, J.; Casentini, B.; Masion, A.; Poethig, R.; Wehrli, B.; Furrer, G. Polyaluminum chloride with high Al30 content as removal agent for arsenic-contaminated well water. Water Res. 2012, 46, 53− 62. (9) Mertens, J.; Rose, J.; Kaegi, R.; Chaurand, P.; Ploetze, M.; Wehrli, B.; Furrer, G. Adsorption of arsenic on polyaluminum granulate. Environ. Sci. Technol. 2012, 46, 7310−7317. (10) Stewart, T. A.; Trudell, D. E.; Alam, T. M.; Ohlin, C. A.; Lawler, C.; Casey, W. H.; Jett, S.; Nyman, M. Enhanced water purification: A single atom makes a difference. Environ. Sci. Technol. 2009, 43, 5416− 5422. (11) Mensinger, Z. L.; Wang, W.; Keszler, D. A.; Johnson, D. W. Oligomeric group 13 hydroxide compounds - a rare but varied class of molecules. Chem. Soc. Rev. 2012, 41, 1019−1030.

■

CONCLUSIONS The combined structural characterization and computational study of the Cu2Al30-S polyaluminum species provides consistent evidence for the preferred Cu2+ adsorption site on the Al30 cluster. The net formation of Cu2(SO4)2Al30 was studied via two theoretical stepwise mechanisms differing in the order of anion outer-sphere and cation inner-sphere adsorption steps and revealed opposing trends in anion/cation affinity as a function of adsorption site on Al30. We conclude that the coupling of Al30 geometry and electronic structure results in distinct structure−reactivity relationships for adsorption sites in the beltway versus the caps of Al30, with the former exhibiting greater reactivity toward outer-sphere adsorption of SO42− and the latter showing greater reactivity toward specific adsorption of Cu2+. The DFT calculations for coadsorption suggests that the electrostatic interactions govern the bonding competition, as Cu2+ prefers adsorption at the caps of Al30 in the absence of SO42−, or at the beltway in its presence. The inwardly folded oxygen functional groups present in the beltway contrast to the more ideal surface-like functional group topology on the caps. The reversal of anion/cation affinity in Al30 may be conceptually extendable to understand fundamental differences between extended surface and nanoparticle aluminum hydroxide reactivity.

■

AUTHOR INFORMATION

ASSOCIATED CONTENT

S Supporting Information *

Crystallographic information file (CIF) of Cu2Al30-S, TGA, ICP-OES, SEM-EDS results, and additional details of the 14132

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

arsenatomolybdates and Cu-II-organic units. Cryst. Growth. Des. 2009, 9, 5206−5212. (32) Kawata, S.; Kumagai, H.; Adachi, K.; Kitagawa, S. Novel layered structures constructed from metal(II)-chloranilate monomer compounds. J. Chem. Soc., Dalton Trans. 2000, 2409−2417. (33) Rademeyer, M.; Turnbull, M. M.; Landee, C. P.; Kunene, M.; Liles, D. C. Crystal structure and magnetic properties of catena-poly bromido [trans(1S,2S)/(1R,2R)-cyclohexane-1,2-diamine copper(II)μ3-bromido]. Solid State Sci. 2012, 14, 1106−1110. (34) Yasodha, V.; Govinclarajan, S.; Low, J. N.; Glidewell, C. Cationic, neutral and anionic metal(II) complexes derived from 4-oxo4H-pyran-2,6-dicarboxylic acid (chelidonic acid). Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 2007, 63, M207−M215. (35) Keene, T. D.; Hursthouse, M. B.; Price, D. J. Stabilization of discrete [Cu(C2O4)2(H2O)2]2− dianions in the solid state by an extensive hydrogen bonded network. Z. Anorg. Allg. Chem. 2004, 630, 350−352. (36) Choudhury, S. R.; Bhattacharyya, J.; Das, S.; Dey, B.; Mukhopadhyay, S.; Lu, P.; Zhu, M. L. Bis(2-aminopyridinium) diaqua-bis(malonato-κ2 O,O′)cuprate(II). Acta Crystallogr., Sect. E: Struct. Rep. Online 2007, 63, M1331−U1370. (37) Yan, B. B.; Goh, N. K.; Chia, L. S. Hydrothermal syntheses and crystal structures of novel one- and two-dimensional organic-inorganic hybrid materials composed of paradodecatungstate-B clusters and [Cu(en)2]2+ complex groups. Inorg. Chim. Acta 2004, 357, 490−494. (38) Allouche, L.; Gerardin, C.; Loiseau, T.; Ferey, G.; Taulelle, F. Al30: A giant aluminum polycation. Angew. Chem., Int. Ed. 2000, 39, 511−514. (39) Rowsell, J.; Nazar, L. F. Speciation and thermal transformation in alumina sols: Structures of the polyhydroxoaluminum cluster [Al30O8(OH)56(H2O)26]18+ and its δ-Keggin Moieté. J. Am. Chem. Soc. 2000, 122, 3777−3778. (40) Sun, Z.; Wang, H.; Tong, H.; Sun, S. A giant polyaluminum species S-Al32 and two aluminum polyoxocations involving coordination by sulfate ions S-Al32 and S-K-Al13. Inorg. Chem. 2011, 50, 559− 564. (41) Abeysinghe, S.; Unruh, D. K.; Forbes, T. Z. Crystallization of Keggin-type polyaluminum species by supramolecular interactions with disulfonate anions. Cryst. Growth Des. 2012, 12, 2044−2051. (42) Armstrong, C. R.; Casey, W.; Navrotsky, H. A. Energetics of Al13 Keggin cluster compounds. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 14775−14779. (43) Violante, A. Elucidating Mechanisms of Competitive Sorption at the Mienral/Water Interface. In Advances in Agronomy; Sparks, D. L., Ed.; Academic Press: San Diego, CA, 2013; Vol. 118, pp 111−176. (44) McBride, M. B. Cu2+ - adsorption characteristics of aluminum hydroxide and oxyhydroxides. Clays Clay Miner. 1982, 30, 21−28. (45) Weesner, F. J.; Bleam, W. F. X-ray absorption and EPR spectroscopic characterization of adsorbed copper(II) complexes at the boehmite (AlOOH) surface. J. Colloid Interface Sci. 1997, 196, 79− 86. (46) Gessa, C.; Decherchi, M. L.; Melis, P.; Micera, G.; Erre, L. S. Anion-induced metal binding in amorphous aluminum hydroxide. Colloids Surf. 1984, 11, 109−117. (47) Xia, K.; Taylor, R. W.; Bleam, W. F.; Helmke, P. A. The distribution of Cu(II) on boehmite and silica surfaces: Correlating EPR signal loss with the effective Bohr magneton number of sorbed ions. J. Colloid Interface Sci. 1998, 199, 77−82. (48) Rudin, M.; Motschi, H. A molecular model for the structure of copper complexes on hydrous oxide surfaces: An ENDOR study of ternary Cu(II) compelxes on δ-alumina. J. Colloid Interface Sci. 1984, 98, 385−393. (49) Harsh, J. B.; Doner, H. E.; McBride, M. B. Chemisorption of copper on hydroxy-aluminum-hectorite: An electron spin resonance study. Clays Clay Miner. 1984, 32, 407−413. (50) Motschi, H. Correlation of EPR-parameters with thermodynamic stability constants for copper(II) complexes Cu(II)-EPR as a probe for the surface complexation at the water/oxide interface. Colloids Surf. 1984, 9, 333−347.

(12) Dubbin, W. E.; Sposito, G. Copper-glyphosate sorption to microcrystalline gibbsite in the presence of soluble Keggin Al13 polymers. Environ. Sci. Technol. 2005, 39, 2509−2514. (13) Kabata-Pendias, A. Soil-plant transfer of trace elements - an environmental issue. Geoderma 2004, 122, 143−149. (14) Hoft, R. W. W. COLLECT; Nonius, B. V., Ed.; Delft, The Netherlands, 1998. (15) Sheldrick, G. M. APEX II; Bruker AXS: Madison, WI, 1996. (16) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, 64, 112−122. (17) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool; Utrecht University: Utrecht, The Netherlands, 2005. (18) Klamt, A.; Schuurmann, G. Cosmo - a new approach to dielectric screening in solvents with explicit expressions for the screening energy and its gradient. J. Chem. Soc., Perkin Trans. 1993, 799−805. (19) Delley, B. An all-electron numerical-method for solving the local density functional for polyatomic-molecules. J. Chem. Phys. 1990, 92, 508−517. (20) Delley, B. From molecules to solids with the DMol3 approach. J. Chem. Phys. 2000, 113, 7756−7764. (21) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (22) Rustad, J. R. Molecular dynamics simulation of the titration of polyoxocations in aqueous solution. Geochim. Cosmochim. Acta 2005, 69, 4397−4410. (23) Gaussian09: Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.; , Jr., Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S., Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Ö . Farkas, Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc., Wallingford, CT, 2009. (24) Adamo, C.; Barone, V. Toward reliable density functional methods without adjustable parameters: The PBE0 model. J. Chem. Phys. 1999, 110, 6158−6170. (25) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. New developments in the polarizable continuum model for quantum mechanical and classical calculations on molecules in solution. J. Chem. Phys. 2002, 117, 43−54. (26) Scalmani, G.; Frisch, M. J. Continuous surface charge polarizable continuum models of solvation. I. General formalism. J. Chem. Phys. 2010, 132, 114110−15. (27) Schafer, A.; Horn, H.; Ahlrichs, R. Fully optimized contracted Gaussian-basis sets for atoms Li to Kr. J. Chem. Phys. 1992, 97, 2571− 2577. (28) Schafer, A.; Huber, C.; Ahlrichs, R. Fully optimized contracted Gaussian-basis sets of triple zeta valence quality for atoms Li to Kr. J. Chem. Phys. 1994, 100, 5829−5835. (29) Yang, W. J.; Qian, Z. S.; Lu, B. M.; Zhang, J.; Bi, S. P. Density functional theory study and kinetic analysis of the formation mechanism of Al30O8(OH)56(H2O)2618+ (Al30) in aqueous solution. Geochim. Cosmochim. Acta 2010, 74, 1220−1229. (30) Saukkoriipi, J.; Sillanpaa, A.; Laasonen, K. Computational studies of the cationic aluminium(chloro) hydroxides by quantum chemical ab initio methods. Phys. Chem. Chem. Phys. 2005, 7, 3785− 3792. (31) Li, L. L.; Liu, B.; Xue, G. L.; Hu, H. M.; Fu, F.; Wang, J. W. Three Hybrid organic-inorganic assemblies based on different 14133

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134

Langmuir

Article

(51) Abeysinghe, S.; Unruh, D. K.; Forbes, T. Z. Surface modification of the Al30 Keggin-type polyaluminum molecular clusters. Inorg. Chem. 2013, 52, 5991−5999. (52) Chiang, H.; Bhan, A. Catalytic consequences of hydroxyl group location on the kinetics of n-hexan hydroisomerization over acidic zeolites. J. Catal. 2011, 283, 98−107. (53) Martins, J.; Birot, E.; Guillon, E.; Lemos, F.; Ramoa Ribeiro, F.; Magnoux, P.; Laforge, S. Sodium exchange over H-EU-1 zeolite. Part I: Physicochemical characterization. Microporous Mesoporous Mater. 2013, 171, 230−237. (54) Liu, H.; Tan, K. Adsorption of water on single-walled TiO2 nanotube: A DFT investigation. Comput. Theor. Chem. 2012, 991, 98− 101. (55) Pan, X.; Cai, Q.; Chen, W.; Zhuang, G.; Li, X.; Wang, J. A DFT study of gas molecules adsorption on the anatase (001) nanotube arrays. Comput. Mater. Sci. 2013, 67, 174−181. (56) Zhang, P. C.; Sparks, D. L. Kinetics of selenate and selenite adsorption desorption at the geothite water interface. Environ. Sci. Technol. 1990, 24, 1848−1856. (57) Turner, L. J.; Kramer, J. R. Sulfate ion binding on geothite and hematite. Soil Sci. 1991, 152, 226−230. (58) Peak, D.; Ford, R. G.; Sparks, D. L. An in situ ATR-FTIR investigation of sulfate bonding mechanisms on goethite. J. Colloid Interface Sci. 1999, 218, 289−299. (59) Collins, C. R.; Ragnarsdottir, K. V.; Sherman, D. M. Effect of inorganic and organic ligands on the mechanism of cadmium sorption to goethite. Geochim. Cosmochim. Acta 1999, 63, 2989−3002. (60) Brown, I. D.; Shannon, R. D. Empirical bond-strength bondlength curves for oxides. Acta Crystallogr. 1973, A29, 266−282.

14134

dx.doi.org/10.1021/la402736t | Langmuir 2013, 29, 14124−14134