Large Aqueous Aluminum Hydroxide Molecules

Large Aqueous Aluminum Hydroxide Molecules William H. Casey* Department of Chemistry and Department of Geology, University of California, Davis, Calif...

0 downloads 95 Views 794KB Size
Volume 106, Number 1

Large Aqueous Aluminum Hydroxide Molecules William H. Casey* Department of Chemistry and Department of Geology, University of California, Davis, California 95616 Received July 8, 2005

Contents 1. Introduction 2. Classes of Large Aqueous Aluminum Hydroxide Clusters 2.1. Structures with a Central Tetrahedral M(O)4 Site 2.1.1. The -Al13 Molecule 2.1.2. Heteroatom -MAl12 Structures 2.1.3. Transformations of the -Al13 to Other Oligomers 2.1.4. The δ-Al13 Molecule 2.1.5. R-Al13 Structures 2.1.6. The Al2O8Al28(OH)56(H2O)2618+(aq) (Al30) Molecule 2.1.7. Heteroatom Derivatives of the Al30 2.2. Molecular Clusters Based upon Brucite-like Al3(OH)45+ Cores 2.3. Alumoxanes 2.4. New Methods of Aluminum and Heteroatom Cluster Isolation 3. Kinetics of Ligand-Exchange Reactions in Aluminum and Heteroatom Clusters 3.1. The Analogy to Aluminum Hydroxide Mineral Surfaces 3.2. Oxygen-Isotope Exchange Rates 3.3. Kinetic Data for Other Ligand Exchanges on the MAl12 3.3.1. Reactions at Bound Water Molecules 3.3.2. Reactions at Bridging Hydroxyls 3.3.3. Dissociation Rates and Pathways 3.3.4. Formation Pathways 4. Uses and Environmental Significance of the Aqueous Aluminum Hydroxide Clusters 4.1. Water Treatment 4.2. Pillaring Agents 4.3. Environmental Significance 5. Conclusions 6. Acknowledgments 7. References * E-mail: [email protected] Phone: 530-752-3211.

1 2 2 2 3 3 3 4 4

7

William Casey (born 1955) received his Ph.D. degree in Mineralogy and Geochemistry from The Pennsylvania State University in 1986 under the direction of Prof. Antonio Lasaga. After graduating, he worked as a research geochemist at Sandia National Laboratories in Albuquerque, New Mexico, for several years and jointed the faculty of the University of California in 1991. He has published over 130 scientific articles on subjects relating to aqueous solution chemistry of natural waters, mineral surface chemistry, and reaction kinetics.

7

1. Introduction

5 6 7 7

7 10 10 11 12 12 13 13 13 13 13 14 14

One of the first subjects introduced to students of environmental chemistry is the aqueous chemistry of aluminum. This metal is the third most abundant element in the shallow Earth, where it hydrolyzes in water to produce a rich array of solute molecules and solids, including clays and aluminum hydroxide phases. Although these materials are ubiquitous, we are just beginning to understand the kinetic properties of their surfaces at the molecular scale. The problem is experimentalsthe solids are too unwieldy, even as colloids, for detailed spectroscopy. One recent approach has been to use 1-2 nm aqueous Al(III) molecules as experimental models to determine reaction rates and pathways at a fundamental level. The 1-2nm-sized clusters are useful because they expose functional groups that resemble those found on the minerals, yet reactions at these functional groups can be studied at the molecular scale using relatively simple methods of solution spectroscopy, such as NMR. The 1-2 nm ions are sufficiently small that reactions can then be simulated at a high

10.1021/cr040095d CCC: $59.00 © 2006 American Chemical Society Published on Web 12/07/2005

2 Chemical Reviews, 2006, Vol. 106, No. 1

Casey

level using ab initio and molecular-dynamics methods [e.g., ref 1]. In addition to their use as experimental models, these 1-2nm-sized aluminum hydroxide molecules are familiar to us as the key ingredient in antiperspirants [e.g., ref 2], dye mordants,3,4 taste astringents, and surfactants [e.g., ref 5]. The polyoxocations are also present in catalysts and claypillaring agents [e.g., refs 6-11] and in water-treatment plants. We add them to water to eliminate organic macromolecules and metal pollutants [e.g., refs 12 and 13], and these aluminum hydroxide molecules are not too different from some vaccine adjuvants that we inject into ourselves [e.g., refs 14 and 15]. This paper begins by reviewing the known classes of aluminum hydroxide polyoxocations. It then presents recent kinetic information about their interactions in water.

2. Classes of Large Aqueous Aluminum Hydroxide Clusters The aluminum hydroxide molecules fall into two broad structural classes. Most familiar are derivatives of the BakerFiggis-Keggin isomers [Figure 1] that have central metals

Figure 2. An 27Al NMR spectrum of a hydrolyzed AlCl3 solution showing peaks corresponding to three of the large aluminum molecules, -Al13, δ-Al13, and Al30. These molecules have a diagnostic peak in the 27Al NMR spectrum because of their relatively symmetric Al(O)4 sites in the center of the molecules. The more abundant Al(O)6 sites yield a broad peak near +10 ppm that is not helpful. These spectra were taken at elevated temperature (∼80 °C) to make the peak near +10 ppm particularly conspicuous.

because only complexes with a tetrahedral site yield diagnostic peaks in 27Al or 71Ga NMR spectra [Figure 2]. In the regime of extreme narrowing but slow chemical exchange, the NMR peak widths for 17O and 27Al, the principle NMR nuclei, are dominated by quadrupolar relaxation:

[

]

3 2I + 3 η2 1 ) π(FWHM) ) 1 + [2πCq]2τc (1) T2 40 I2(2I - 1) 3

Figure 1. The Baker-Figgis-Keggin isomers shown in polyhedral representation.16,214 The isomers can be understood as the stepwise rotation of trimeric groups of Al(O)6 octahedra that share corners (light gray) about the µ4-O so that they share edges with one another (darker gray).

tetrahedrally coordinated to oxygens [M(O)4 sites]. The Baker-Figgis-Keggin isomers are familiar structures among scientists who study polyoxometalates [e.g., refs 16 and 17] and form aluminum molecules having the stoichiometry MO4Al12(OH)24(H2O)127+(aq) [M ) Ge(IV), Ga(III), or Al(III)]. The second class of oligomers have a characteristic core of edge-shared Al(O)6 octahedra organized into cubane-like moieties that are linked together in a structure similar to the mineral brucite (Mg(OH)2). These molecules are most commonly synthesized with an aminocarboxylate ligand that reduces the overall charge [e.g., refs 18 and 19] but have been found in purely inorganic solutions as well,20,21 and extensive Fe(III) analogues exist.22 It is important to note that these are not the only large aluminum oligomers that are present in a concentrated aluminum solution (see below) but are just the oligomers that can be isolated and for which structural data are available. Most large polymers are yet uncharacterized and unidentified

where I is the spin quantum number (I ) 5/2 for both 17O and 27Al; I ) 3/2 for 69Ga and 71Ga), Cq is the nuclear quadrupolar coupling constant (the product of the nuclear quadrupolar moment and the maximum component of the electric-field gradient at the nucleus, in hertz), η is the asymmetry of the electric-field gradient (unitless and generally assumed to be zero unless otherwise indicated), and τc is the molecular rotational correlation time (s).23,24 The electric-field gradients at the octahedral aluminum sites [Al(O)6] in most of the oligomers are sufficiently large to yield broad 27Al NMR peaks at δ ≈ 10 ppm that are not diagnostic in identification. Only structures with a relatively symmetric 27Al(III), which usually means a tetrahedral Al(O)4 site, yield diagnostic peaks in 27Al NMR spectra.

2.1. Structures with a Central Tetrahedral M(O)4 Site 2.1.1. The -Al13 Molecule The -Al13 ion was originally isolated as sulfate and selenate salts25-28 and has an -Keggin-like structure of Td symmetry, containing a central tetrahedral Al(O)4 unit surrounded by twelve Al(O)6 octahedra. The structure of the -Al13 can be viewed [Figure 3] as consisting of four planar trimeric Al3(OH)6 groups that are linked to the central Al(O)4 site via four µ4-O. The molecule has 12 η-OH2 sites and two structurally distinct sets of 12 µ2-OH at the shared edges of Al(O)6 octahedra. These two sets of µ2-OH differ

Large Aqueous Aluminum Hydroxide Molecules

Chemical Reviews, 2006, Vol. 106, No. 1 3

Reports also exist for aluminum-free versions of the -isomer of the Keggin structure, but none have been crystallized for structural analysis and confirmation. Complexes of Ga(III), (GaO4Ga12(OH)24(OH)12)7+,48,62 have been suggested but not yet isolated [see ref 63]. Similarly, the Fe(III) and Cr(III) analogues52,53 are suspected but not separated or crystallized.

2.1.3. Transformations of the -Al13 to Other Oligomers Figure 3. The -Keggin isomer of the MO4Al12(OH)24(H2O)127+(aq) series of molecules (e.g., -Al13) shown in polyhedral representation (left) and as a ball-and-stick model (right). The Al(III) atoms are green, the oxygens are red, and the hydrogens are portrayed as uncolored spheres. The molecule can be viewed as four Al3(OH)6(H2O)3 trimeric groups linked together at polyhedral edges around the central M(O)4 site.

in their positions relative to the µ4-O groups. One site, labeled µ2-OHa, lies cis to two µ4-O groups [Figure 3]. The other site, labeled µ2-OHb, lies cis to one µ4-O site. One can view these two sites as either linking two trimeric groups together (µ2-OHa) or linkages within a single trimeric group (µ2-OHb). The -Al13 molecule is usually synthesized in a relatively concentrated aluminum solution but can be made over a broad range in concentrations (1 > ∑[Al] >10-5 M) by titrating to 2.1 e [OH]/[Al] e 2.5 at 80-90 °C, followed by crystallization with added selenate or sulfate ions. The -Al13 complex in aqueous solution is easily established from the distinct and narrow peak at 62.5 ppm in 27Al NMR spectra [e.g., refs 29-35]. It can also be detected indirectly by the uptake by phenolic sulfonate ligands and spectrophotometric detection [e.g., ref 36]. The ligand ferron (8-hydroxy-7-iodo-5-quinoline)32,37-40 and pyrocatechol violet41 have been used to determine the amount of polymerized aluminum in solution and rely upon the kinetics of reaction [e.g., refs 37 and 42]. Akitt et al.43 reported a partial molar volume for the -Al13 molecule, and equilibrium constants are available,44,45 although it remains questionable as to whether this molecule is truly in reversible exchange equilibrium with other species in solution or it is only a persistent metastable product [see ref 42 and references therein].

2.1.2. Heteroatom -MAl12 Structures Two decades after Johanssen’s work on the -Al13, considerable progress was made by the research teams at the University of Freiburg in Germany and Calgary in Canada who independently published a series of papers that examined metal substitutions and polymerization.46-56 Singleatom substituents include Ge(IV) (yielding GeO4Al12(OH)24(H2O)128+ in solution ) GeAl12, refs 46 and 57] and Ga(III) (GaO4Al12(OH)24(H2O)127+ ) GaAl12, refs 47-51). Crystals of Na[GaO4Al12(OH)24(H2O)12(SeO4)4]‚x(H2O), [GeO4Al12(OH)24(H2O)12(SeO4)4]‚x(H2O), and Na[AlO4Al12(OH)24(H2O)12(SeO4)4]‚x(H2O) can be grown at 80-90 °C by hydrolysis of the appropriate AlCl3 + MCl3 solutions followed by filtration, dilution, cooling, and addition of selenate to induce crystallization. The GeAl12 molecule has a slight structural distortion from the cubic symmetry exhibited by -Al13 and GaAl12 but is overwhelmingly similar.57 There is indirect evidence for single-atom substitution of Mn(II)58 and Fe(III)59,60 although when Parker et al.61 attempted to synthesize these and other MAl12 molecules they concluded that only the GaAl12 was unequivocal. This small series has since been augmented with the GeAl12.57

Fu et al.64 and Nazar et al.65 showed that other peaks appear downfield in the 27Al NMR spectra as solutions containing the -Al13 are heated for days at 80-95 °C [see also refs 66 and 67]. Peaks appear near δ ) 64.5 ppm, δ ) 71.2 ppm, and δ ) 75.6 ppm in the 27Al NMR spectra. These authors could enrich solutions in the molecules that yield these peaks by gel-permeation chromatography but could not crystallize the molecules into a material suitable for a structural analysis. The peaks were assigned to AlP1, AlP2, and AlP3 (aluminum peak no. 1, etc.). Fu et al.64 documented a progressive reaction series where the -Al13 reacts to form the AlP1 as an intermediate, then the AlP2 molecule plus aluminum monomers (the monomers yield a peak near δ ) 0 ppm and correspond to Al(H2O)63+ and its conjugate base (Al(H2O)5OH2+) at 4 < pH < 6). Assignment of the AlP1 peak is presented in section 2.1.6. The AlP2 molecule was more stable than the AlP1 or -Al13, and they interpreted it as a dimer of one of the smaller molecules, probably the -Al13 cluster.64 Because the resonance at +64.5 ppm appears and then disappears as peaks at +70.2 and 0 ppm grow, Fu et al.64 concluded that the AlP1 complex is a transient intermediate. Parker et al.61 showed that the GaAl12 molecule could not be converted into an AlP2 equivalent, indicating that the central Ga(III) atom considerably stabilized the structure. Both Fu et al.64 and Parker et al.61 report apparent pseudo-first-order rate coefficients for the polymerization of -Al13 into the AlP2 oligomer at 80 °C of about (0.036-0.05) × 10-2 h-1 (see also refs 68 and 69).

2.1.4. The δ-Al13 Molecule Of the five Baker-Figgis-Keggin isomeric structures [Figure 1], the δ-Keggin isomer of the Al13 molecule (δ-Al13) is the only other aluminum oligomer besides the -Al13 that has yet been synthesized in isolation and structurally characterized (see below). The δ-Al13 molecule has a single [Al3O13] trimeric group rotated 60° around the µ4-O [Figure 4] so that it bonds at polyhedral corners, not at

Figure 4. The δ-Al13 molecule shown in polyhedral representation (left) and as a ball-and-stick model (right).70 The molecule is similar to the -Al13 but with one trimeric group (shown in brown) rotated 60° so that it bonds via octahedral corners, not edges. Some of the structurally distinct hydroxyl bridges are identified. The Al(III) atoms are green, the oxygens are red, and the hydrogens are portrayed as uncolored spheres.

4 Chemical Reviews, 2006, Vol. 106, No. 1

Casey

the edges. All three other [Al3O13] trimers in the molecule join at shared edges, as in the -Al13. Rowsell and Nazar70 isolated this δ-Al13 isomer structure, along with the large Al30 molecule discussed in the next section. The δ-Al13 structure has more types of oxygens than the -Al13 molecule because of the introduction of the cornershared µ2-OH and the resulting reduction in symmetry from Td to C3V. For example, the δ-Al13 contains three distinct sets of η-OH2 and two types of µ4-O, based upon their positions relative to the rotated trimer. Rowsell and Nazar70 conclude that the δ-Al13 molecule accounts for the AlP1 peak of Fu et al.64 and yields a peak in the 27Al NMR spectrum near δ ) 64.5 ppm [see also refs 68, 69, and 71].

2.1.5. R-Al13 Structures The only aluminum oligomer that forms in one of the Baker-Figgis-Keggin isomers is in the mineral zunyite [Figure 5], which has the stoichiometry Al13Si5O20-

Figure 6. The Al2O8Al28(OH)56(H2O)2618+(aq) (Al30) molecule shown in polyhedral representation (top left), as a ball-and-stick model (top right) and in a polyhedral exploded view (bottom). Hydrogens are eliminated from the structure for the sake of clarity.70,71 The Al30 molecule can be viewed as two δ-Al13 molecules joined via a belt of four joining Al(O)6 octahedra. The blue polyhedra are organized in a similar fashion as the -Al13. In the ball-and-stick diagram, the Al(III) atoms are green and the oxygens are red. Figure 5. The mineral zunyite, with the nominal stoichiometry Al13Si5O20(OH)16F2Cl, is the only occurrence of the R-Al13 isomer of the Baker-Figgis-Keggin series. The mineral contains R-Al13 clusters (shown here in both light and dark gray) with silicate chains bonded sharing some hydroxyl bridges in µ3-OH coordination. The silicate groups are shown as hatched-pattern tetrahedra.

(OH)16F2Cl and is a modified form of the R-Al13. This mineral is found near hydrothermal ore deposits, can be synthesized hydrothermally [e.g., ref 72], and was first studied by Pauling.73 The mineral is used extensively for structural [e.g., ref 74] and spectroscopic studies [e.g., refs 75-78], but the R-Al13 has not yet been isolated as a solute or as a molecule in a simple salt. The silicate groups are directly bonded to the R-Al13 in zunyite.

2.1.6. The Al2O8Al28(OH)56(H2O)2618+(aq) (Al30) Molecule An exceptional advance in identifying aqueous aluminum polymers was the isolation and structural characterization by two independent groups70,71 of the largest aluminum

polyoxocation yet characterized, the Al30 molecule [Al30 ) Al2O8Al28(OH)56(H2O)2618+(aq)], which also accounts for the AlP2 peak of Fu et al.64 and Nazar et al.65 The Al30 is ∼2 nm in length [Figure 6] and exposes oxygens in many different coordination environments to the aqueous solution. The 27Al NMR spectrum shown by Akitt and Mann66 showed the presence of this Al30 molecule as a broad peak near 71 ppm, which is also evident in Akitt and Farthing29 and Akitt et al.67 Similarly, Allouche and Taullele79 show conspicuous peaks at 76 and 81 ppm in the 27Al NMR spectra of their Al30 solutions. Shafran and Perry68 also report minor peaks at 74, 48, and 81 ppm. As all the authors state, these peaks correspond to polymers with a higher molecular weight than the Al30 that have yet to be isolated and crystallized. Fu et al.64 suspected that the AlP3 peak corresponded to a dimer made of two -Al13 that were bonded and rotated to yield a compound with molecular weight near 1500-3000 Da, having the approximate stoichiometry Al24O72. These conclusions were insightfulsthe structure of the Al30 is best

Large Aqueous Aluminum Hydroxide Molecules

understood as two δ-Al13 molecules that face one another at the rotated trimers and are bonded via a belt of additional Al(O)6 linkages [Figure 6]. One set of linkages consists of Al(O)6 groups that share three edges at the apices of the two δ-Al13 units, forming a nonplanar tetrameric cap on each of the two δ-Al13-like molecules. These linkages form three adjacent µ3-OH groups on each tetrameric subunit. The second linkage set consists of Al(O)6 groups that connect the two modified δ-Al13 molecules to one another at the tetramer caps via four corner-shared µ2-OH bridges. The disruption of symmetry creates many sets of different oxygens. In fact, Allouche et al.71 found evidence that the Al30 sulfate salt exhibits Cc space-group symmetry, which requires all 88 oxygens to be inequivalent. If instead we can assume that the molecule has C2h symmetry, we can identify 15 sets of inequivalent hydroxyl bridges in the Al30 and eight distinct sets of bound waters [see ref 80]. The eight total µ4-O sites in the Al30 link the two tetrahedrally coordinated aluminums [Al(O)4] to the outer part of the molecule. Allouche and Taullele79 and Shafran and Perry68,69 conducted time-series studies and showed that the Al30 complex can be easily formed by heating a solution of -Al13 at 85 °C for a few days. However, it also forms during storage of a stock millimolar solution of -Al13 for a decade or so (see Figure 9 in ref 81), and this Al30 molecule can be a component of some commercial aluminum chlorhydrate [Al2(OH)5Cl], along with the -Al13 [e.g., ref 81]. As mentioned above, there is apparently no Ga(III)-centered analogue of the Al30, although the GaAl12 is apparently more stable than the Al30. It worth revisiting this question of a GaAl28 molecule with new experiments since Parker et al.61 report a decrease in intensity of the peak associated with the Ga(O)4 in the 71Ga NMR spectra with aging at 80-90 °C, suggesting that polymerization into a larger molecule was occurring. A more detailed study might uncover this polymerization and isolate the resulting molecule using one of the new supramolecular63 or column82 methods that are being developed. Allouche et al.83 employed a triple-quantum 27Al-MAS NMR method to assign eight peaks in the 27Al NMR spectrum. Narrow peaks at δ ) 68.8 ppm and δ ) 69.9 ppm were assigned to the two Al(O)4 and indicate that the structure is not centrosymmetric. The remaining Al(O)6 peaks exist as six groups with peak positions ranging from δ ≈ 4 to δ ≈ 12 ppm and with quadrupolar-coupling constants ranging from ∼500 to ∼1600 kHz.83

2.1.7. Heteroatom Derivatives of the Al30 Recent efforts created nanocluster composites using the -Al13,84-86 the δ-Al13 and Al30 molecules,87 and the -GaAl12 molecules,88 along with tungstate or molybdate polyoxoanions. The clusters bond to one another in the composite by electrostatic interaction and via hydrogen bonding, and they create a porous solid with interesting properties. Son et al.87 showed that the Al30 molecule exchanges metals in a composite with the H2W12O406- (W12) polyoxoanion to form a new compound, W2Al28, that is nearly isostructural with the Al30 structure [Figure 7]. This new compound differs in that a W(O)6 group replaces one of the capping Al(O)6 groups at each end of the Al30 molecule. The fact that this W2Al28 molecule broadly resembles the dimer-like structure of the wider family of tungstate clusters [e.g., refs 16, 89, and 90] suggests that more varieties of tungstoaluminate clusters could be synthesized [Figure 7].

Chemical Reviews, 2006, Vol. 106, No. 1 5

Figure 7. The Al(III) f W(VI) substitution (and the reverse) into the largest tungstate and aluminate clusters might be a fruitful area for research. The structure of the W2Al28O18(OH)48(H2O)2412+ ion (W2Al28) shown in polyhedral representation (top) is nearly isostructural with the Al30 ion. The Al(O)6 groups are shown in dark gray, and the W(O)6 groups are shown as light gray. The Al(O)4 site in the center of the molecules is hatched. In the bottom figure, the H14Si2W18Al6O37(H2O)12 cluster can be considered as an altered dimer of two aluminum-substituted and Si(O)4-centered tungstate clusters in the R-Keggin structure.215 One trimeric group of the four in each R-W13 Keggin structure is replaced by a group of three Al(O)6. The two dimers then link across a shared edge of two Al(O)6 groups.

Similarly, there are several classes of heteropolyanions that contain aluminum substituents, either as the tetrahedral core of a molecule with one of the Keggin isomer structures [e.g., ref 91] or as an Al(O)6 heteroatom substituent in the outer part of a tungsten or molybdate Keggin isomer structure [e.g., refs 92-95]. Son et al.87 demonstrated that aluminumrich tungstate clusters could be synthesized based upon the aluminum polyoxocations of the Keggin class, such as the Al30.

6 Chemical Reviews, 2006, Vol. 106, No. 1

Casey

similar Al(III), Ga(III), and Fe(III) clusters can be stabilized using aminocarboxylate ligands [e.g., refs 18, 19, 22, and 99-101]. For aluminum clusters, two ion stoichiometries have been synthesized, Al13(µ3-OH)6(µ2-OH)12(H2O)6(heidi)63+ (heidi ) N-(2-hydroxyethyl)iminodiacetic acid, see refs

Figure 8. A new series of aqueous aluminum clusters is based upon a repeated set of cubane-like moieties (top) arrayed in a lattice of edge-shared Al(OH)6 octahedra similar to the brucite lattice.96 Three such molecules have been synthesized, having either 8, 13, or 15 aluminums, but several other molecules have been made from other trivalent metals (i.e., Fe(III) and Ga(III), see ref 96). The flatAl13 has the core structure Al13(µ3-OH)6(µ2-OH)18(H2O)615+ and is shown at bottom in polyhedral representation. The flat-Al13 has been synthesized using chloride salts to give the Al13(µ3-OH)6(µ2-OH)18(H2O)615+ ion20,21 and employing heidi as a tetradentate chelating ligand,18 which bonds at the six corner-shared Al(O)6 groups, giving Al13(µ3-OH)6(µ2-OH)12(H2O)6(heidi)63+ ion.

18, 19, 22, 99, and 101) and Al15, Al15(µ3-O)4(µ3-OH)6(µ2-OH)14(hdpta)3- (H5hpdta ) 2-hydroxypropane-1,3-diamine-N,N,N′,N′′-tetraacetic acid.101 To distinguish these molecules from the -Al13 molecule, I refer to the inorganic chloride salt as the flat-Al13 and the heidi-ligated structure as the flat-Al13-heidi. Goodwin et al.96 argued that a wide range of metal hydroxide clusters are built upon this hypothetical brucitelike lattice, including clusters of Fe(III), Al(III), Mn(III), and Ga(III) with 7, 8, 13, 15, 17, 19, and 21 metals. The relation between this hypothetical Al(OH)3 brucite lattice and the cluster structure is illustrated in Figure 9, where we relate

2.2. Molecular Clusters Based upon Brucite-like Al3(OH)45+ Cores Other aluminum clusters have been isolated in the past decade that are based upon a periodic array of alternating Al(III) and hydroxyl linkages in cubane-like moieties [Figure 8] arrayed in edge-shared Al(O)6 similar to brucite [see ref 96]. These molecules have no central Al(O)4 site and thus are difficult to detect in an 27Al NMR spectrum because the Al(O)6 sites yield broad peaks. The purely inorganic salts are also difficult to crystallize, and the only ion stoichiometries that have yet been reported are the octamer Al8, Al8(µ3-OH)2(µ2-OH)12(H2O)1210+ 97 and the flat-Al13, Al13(µ3-OH)6(µ2-OH)18(H2O)615+.20,21 The latter ion probably forms one of the salts in the early study by Breuil [ref 98; F. Taulelle, personal communication] and can be synthesized in large quantities by hydrothermal reaction. In their core structure, these clusters resemble somewhat the Andersontype structures, which are familiar in molybdate chemistry (e.g., Al(OH)6Mo6O183-; see refs 92 and 93) and consist of linked M(O)6 octahedra surrounding a central Al(O)6. In the aluminomolybdates, however, there are no corner-shared Al(O)6, as in these brucite-like molecules. Although the purely inorganic salts are difficult, the synthesis is much easier if the molecules are ligated at the edges by organic anions that reduce the molecule charge. Much work on this subject has been done by the research groups of Profs. Powell and Heath at Universities of Karlsruhe and Manchester, respectively, who showed that

Figure 9. Aluminum and many other metal hydroxide clusters96 can be built as fragments of a hypothetical brucite-like lattice made of trivalent metals linked by µ3-OH. An Al(OH)3 solid built on the brucite (Mg(OH)2) lattice is shown at the top with the Al(III) as gold, the oxygens red, and the hydrogens as white spheres. To generate the flat-Al13 clusters (bottom right), a hexagonal array of Al(OH)6 groups (light-blue atoms) has one bond to a µ3-OH replaced with a water molecule and three bound waters truncating them. The Al(III) atoms in the core (dark-blue) retain their brucitelike structure, while the light-blue atoms bond to the next row of Al(O)6 via µ2-OH. In the clusters synthesized with aminocarboxylates (e.g., the flat-Al13-heidi and Al15).

the flat-Al13 structure of Seichter et al.20 to a brucite basal plane. The metal atoms are arranged on a hexagonal array with six of the outermost Al(OH)6 groups having a single bond to the µ3-OH cleaved and replaced with a bound water molecule to retain the octahedral coordination. The result is a molecule with a central cubane-like core with corner-shared Al(O)6 at the edges. Although the Al8 is the smallest cluster of this series, the Al(III)-citrate trimer isolated and crystal-

Large Aqueous Aluminum Hydroxide Molecules

lized by Feng et al.102 bears some similarity to the structure in that one Al(O)6 site is corner-shared to the other two Al(O)6, which share edges via two µ2-OH [see also refs 81 and 99]. Ligation to the aminocarboxylates does not change the core structure of linked Al(O)6 in a brucite-like arrangement. The aminocarboxylate ligands are tetradentate and replace the outer four water molecules in each of the peripheral cornershared Al(O)6 groups. Thus, the flat-Al13 molecule in chloride salt has the same core structure as the flat-Al13heidi, but the tetradentate heidi replaces bound waters. In the flat-Al13, bound water molecules exist that are both cis and trans to a µ2-OH and are hence expected to be kinetically distinct. In the Al8 molecule, a second set of water molecules is bound to edge-shared Al(O)6 in the core. These water molecules are trans to a µ3-OH and cis to four µ2-OH and are hence expected to be anomalously labile. In the Al15 structure, there are no bound waters, and this material is relatively insoluble. The Al8, flat-Al13, and Al15 molecules are ill-suited for standard solution 27Al NMR studies because there is no highly symmetric site in the cubane-like structures, such as the Al(O)4 in the -Al13 molecule, that yields a distinctive peak in the 27Al NMR spectrum. Furthermore, both the flatAl13 and Al8 molecules decompose in most aqueous solutions, rendering 17O NMR studies of their reactivity impossible. However, multiquantum 27Al-MAS NMR (MQMAS) has proven useful, and a key study was that of Allouche et al.,83 who related peak positions and quadrupolar-coupling constants to structural aluminum positions in salts of the -Al13, the Al30, and the flat-Al13 clusters. For the flat-Al13, there are six different aluminum sites in the core structure. The central Al(µ3-OH)6 site has a relatively sharp 27Al NMR resonance in MQMAS at δ ) 11.7 ppm. The surrounding ring of Al(µ3-OH)2(µ2-OH)3(H2O) yields peaks at δ ) 13.3 ppm and δ ) 13.9 ppm. Finally, the six peripheral Al(O)6 that share corners with adjacent Al(O)6 in polyhedral representation have peaks of δ ) 2.9 and 3.1 ppm. Casey et al.97 made a generally similar set of assignments for the Al8 at 11.7 T. They report that the core aluminums have chemical shifts near δ ) 13 ( 1 ppm and Cq values near 6 MHz, which are close to the values for similar sites in the flatAl13 molecule.83,93 Likewise, the remaining outer Al(O)6 sites closely resemble those in the periphery of the flat-Al13 that are linked to two µ2-OH and four η-H2O. They assigned these peripheral Al(O)6, sharing corners with adjacent Al(O)6, to δ ) 6-7 ppm.

2.3. Alumoxanes Alumoxanes form from the hydrolysis of organoaluminum compounds to yield cagelike aluminum structures with µ-oxo and µ-hydroxo bridges and terminal organic ligands [see refs 103 and 104]. Most alumoxanes decompose rapidly in water and are thus beyond the scope of this article. A few hydrolytically stable carboxylate-bridged alumoxanes have been synthesized. Barron’s group105-108 isolated alumoxanes by reaction of carboxylic acids with boehmite [γ-AlOOH, see also ref 109]. However, Callender et al.109 point out that the water-soluble alumoxanes are better thought of as carboxylate-stabilized colloids and not as aqueous clusters of uniform stoichiometry and structure. These molecules could be profoundly useful as kinetic models for aqueous reactions if metastable monospecific solutions could be prepared.

Chemical Reviews, 2006, Vol. 106, No. 1 7

2.4. New Methods of Aluminum and Heteroatom Cluster Isolation The supramolecular methods of synthesis hold great promise for isolating the large aluminum hydroxide clusters, and progress toward this goal was achieved by Drljaca et al.110 and Hardie and Raston111 who showed that the -Al13 molecule could be trapped in a calixarene/8-crown-6-ether combination. A new method of isolating Ga(III) clusters was reported by Gerasko et al.,63 who crystallized the flat-Ga13 molecule (structurally analogous to the flat-Al13) and discovered a new large Ga32 ion [Ga32(µ4-O)12(µ3-O)8(µ2-OH)39(H2O)2017+] that has a unique structure. The Ga32 contains two pairs of corner-shared Ga(O)4 sites surrounded by Ga(O)6 arranged in trimeric clusters like the -Al13 and -Ga13 molecules. This Ga32 structurally resembles a gallium version of the AlP2 molecule originally imagined by Fu et al.64 that has since been identified as the Al30. These new supramolecular clusters could potentially isolate many more aluminum molecules from a hydrolyzed solution. The Gerasko et al. group has also been able to isolate the Al30 and flat-Al13 molecule using their cucurbit supramolecule (V. Fedin, personal communication) but have not yet identified any new aluminum clusters. Rather et al.112 established a method of crystallizing the flat-Ga13 molecule [Ga13(µ3-OH)6(µ2-OH)18(H2O)615+] by slow hydrolysis of an organonitrate and have since also synthesized the flat-Al13 using this method (D.W. Johnson, personal communication).

3. Kinetics of Ligand-Exchange Reactions in Aluminum and Heteroatom Clusters 3.1. The Analogy to Aluminum Hydroxide Mineral Surfaces The similarity of at least some features of these molecules and aluminum hydroxide soil minerals can be restated. Consider the Brønsted acidities, which influence many of the surface properties of the minerals. When fully protonated, minerals such as gibbsite or bayerite [γ-Al(OH)3] have surface proton charge densities of 0.16-0.48 C/m2.113,114 The surface charge density of fully protonated -Al13 is 0.32 C/m2, and the flat-Al13 molecule has a charge density of ∼1.2 C/m2, which is, of course, reduced by ligation to the heidi ligand. It is also important that the cubane-like aluminum hydroxide clusters (Al8, flat-Al13) are broadly similar to minerals of the hydrotalcite class, which are catalysts and are environmentally important. The hydrotalcites are layered structures composed of positively charged, brucite-type metal hydroxide layers intercalated with anions and water molecules. These minerals are important to environmental chemistry because they form quickly in some metal-contaminated environments as aluminum released by dissolving soil minerals combines with the pollutant metal.

3.2. Oxygen-Isotope Exchange Rates In this section, I review some recent kinetic data using the aluminum hydroxide clusters described above. All work to date has employed the small -isomers in the MAl12 series (-Al13, GaAl12, and GeAl12) and the Al30 molecule, which is less useful because of its structural complexity. In 17O NMR studies, sulfate or selenate salts of the -Al13, GaAl12, and GeAl12 are dissolved metathetically in the presence of BaCl2 to release the polyoxocation ions into an 17O-enriched

8 Chemical Reviews, 2006, Vol. 106, No. 1

Figure 10. (top) 17O NMR spectra at 313 K as a function of time for a ∼0.010 M solution of GaAl13 with 0.25 M Mn(II) added to remove the bulk water peak.119 Vertical scaling is normalized to the integrated intensity of the peak near -100 ppm, which corresponds to an external, coaxial TbCl3(aq) insert that was used as an intensity standard. The peak near +22 ppm corresponds to bound water molecules in the GaAl12 complex. The broader downfield peak that increases in intensity with time arises from the two µ2-OH sites in the molecule, which react at different rates. Missing is the peak that would correspond to the four µ4-O, indicating that these sites are inert to exchange. (bottom) The reduced intensity (R(t) ) Iδ)35ppm/Iδ)22ppm) grows in a biexponential manner and ultimately reaches R(t) ) 2, corresponding to the intensity of two sets of 12 fully exchanged µ2-OH (δ ) 35 ppm) divided by the intensity in one set of 12 fully exchanged η-OH2 (δ ) 22 ppm). The ratio corresponds to the stoichiometry GaO4Al12(µ2-OH)24(η-OH2)127+. The biexponential growth indicates that µ2-OHa and µ2-OHb react at different rates, and the inset shows the earliest times, as R(t) f 1, corresponding to exchange of the first set of µ2-OH. Casey and Phillips119 used eq 1 and a value of τc ) 130 ps from ref 116 to calculate Cq values of 8.7 and 6.5 MHz for the two µ2-OH sites in the dissolved GaAl12 molecule.

solution. Changes in 17O NMR peak intensities [Figure 10] then provide information about rates of exchange of µ2-OHa and µ2-OHb sites, and rate parameters for exchange of bound waters can be determined using the 17O-line-broadening method.115 The µ4-O sites in the center of the -Al13, GaAl12, and GeAl12 would appear at +55 ppm116 in a 17O NMR spectrum and yet are missing when the MAl12 salts are dissolved in 17O-enriched water, indicating that these µ4-O are inert to isotopic exchange. This peak at +55 ppm only

Casey

appears in the 17O NMR spectra when the -Al13, GaAl12, and GeAl12 oligomers are completely dissociated into monomers in 17O-enriched water by acidification and then reassembled back into the polyoxocation by base addition. Because the µ4-O sites are inert to exchange, the rates of steady exchange of all other oxygens in the molecules can be measured and these exchanges can be distinguished from dissolution and reformation of the molecule. In fact, virtually all other oxygen sites exchange several times with bulk solution before the molecule dissolves. The 17O NMR spectra exhibit a peak near +20 ppm that is assignable to the 12 bound water molecules. This peak appears instantaneously in the spectra and has a constant intensity. A second peak near +35 ppm grows with time [Figure 10], and the peak exhibits a distinctly biexponential rate of growth. This +35 ppm peak is assigned to two resonances that cannot be resolved from one another and that correspond to the two sets of 12 µ2-OH in the -Al13, GaAl12, and GeAl12 oligomers.117-120 Although the two resonances that contribute to the +35 ppm peak cannot be resolved from one another, they are kinetically distinct and are identified as µ2-OHfast and µ2-OHslow so as not to confuse them with the structural positions, µ2-OHa and µ2-OHb. Although the µ2-OHfast and µ2-OHslow sites cannot yet be unequivocally assigned to the µ2-OHa and µ2-OHb sites within each of the molecules, it is reasonable to assume that the µ2-OHslow site corresponds to the µ2-OHb sites and the µ2-OHfast sites correspond µ2-OHa. There are considerable differences in the reactivities of the hydroxyl bridges both within a single molecule and between the same site in the set of isostructural molecules (-Al13, GaAl12, and GeAl12). Characteristic times for exchange of the µ2-OHfast and µ2-OHslow sites in the GaAl12 molecule at 298 K are τ298 ≈ 15.5 h and τ298 ≈ 680 h, respectively,119 whereas the corresponding times are τ298 ≈ 1 min and τ298 ≈ 17 h for the -Al13 molecule.117,118 The µ2-OHfast bridge in the GeAl12 molecule exchanges at rates too fast to measure,120 but the other has τ298 ≈ 25 min. For slow 3 within the -Al13 complex, τfast 298/τ298 ≈ 10 , but for the fast slow GaAl12 molecule, τ298/τ298 ≈ 44. Thus, the difference in reactivities of these two structural sites is profound. Extrapolated to 298 K, the observed rates of exchange of µ2-OH sites in these molecules spans a factor of ∼105 and is probably much larger if the more reactive set of µ2-OH sites in the GeAl12 molecule could be included. The lifetimes of the bound water molecules, however, fall close to one -3 s) in the series -Al , GaAl , and another (τ298 13 12 ex ≈ 10 GeAl12 and are near values measured for simple aluminum monomer complexes [Table 2]. With the exception of the datum for the µ2-OHfast in the -Al13 molecule [Table 1], the activation enthalpies fall into a range similar to those corresponding to the dissociation of a hydroxyl bridge in simple inert-metal dimers; that is, ∆Hq ≈ 100 kJ mol-1, and ∆Sq is near zero.121 The values of ∆Hq and ∆Sq for the µ2-OHfast in the -Al13 molecule are extraordinarily high (∆Hq ≈ 204 kJ mol-1; ∆Sq ≈ 400 J mol-1 K-1) and may account for the large differences in slow τfast 298/τ298 for this molecule relative to the GaAl12. High-pressure data are available for the GaAl12, and the 17O NMR line widths of bound waters decrease from 0.1 to 350 MPa, yielding an activation volume of ∆Vq ) +3 ( 1 cm3 mol-1,122 which is smaller than the value for the Al(H2O)63+ complex (+5.7 cm3 mol-1),123 even though the average charge density on an aluminum in the GaAl12 is lower than that in the Al(H2O)63+. This deviation is striking.

Large Aqueous Aluminum Hydroxide Molecules

Chemical Reviews, 2006, Vol. 106, No. 1 9

Table 1. Rates of Exchange of Oxygens between Bulk Solution and µ2-OH Bridges in the E-Al13, GaAl12, and GeAl12 Molecules117-120,122 a molecule

k298 ex (s-1)

∆Hq ∆Sq ∆Vq (kJ mol-1) (kJ mol-1 K-1) (cm3 mol-1)

E-Al13 µ2-OHfast (1.6 ( 0.4) × 10-2 204 ( 12 µ2-OHslow (1.6 ( 0.1) × 10-5 104 ( 20

403 ( 43 5(4

GaAl12 µ2-OHfast (1.8 ( 0.1) × 10-5 98 ( 3 µ2-OHslow (4.1 ( 0.2) × 10-7 125 ( 4

-8 ( 9 54 ( 12

+7 ( 1

GeAl12 µ2-OHfast b µ2-OHslow (6.6 ( 0.2) × 10-4

82 ( 2

-29 ( 7

Al30 rates are similar to GaAl12 cluster a At the pH conditions, the molecules are fully protonated and there is no strong pH dependence for exchange rates. A small pH dependence is detectable for the GeAl12 molecule, which is the strongest acid in the set. b Complete in minutes or less.

Table 2. Rate Parameters for Exchange of Water Molecules from the Inner Coordination Sphere of Al(III) Complexes to the Bulk Solution, as Determined from 17O NMRa speciesb

kH298 2O ((1σ)

(s-1)

∆Hq ∆Sq ∆Vq (kJ mol-1) (J K-1 mol-1) (cm3 mol-1)

Monomeric Complexes Al(H2O)6+3 1.29 ( 0.03 85 ( 3 42 ( 9 Al(H2O)5OH2+ 31000 ( 7750 36 ( 5 -36 ( 15 AlF(H2O)52+ 240 ( 34 79 ( 3 17 ( 10 AlF2(H2O)4+ 16500 ( 980 65 ( 2 53 ( 6 Al(ssal)+ 3000 ( 240 37 ( 3 -54 ( 9 Al(sal)+ 4900 ( 340 35 ( 3 -57 ( 11 Al(mMal)+ 660 ( 120 66 ( 1 31 ( 2 Al(mMal)26900 ( 140 55 ( 3 13 ( 11 + Al(ox) 109 ( 14 69 ( 2 25 ( 7 E-Al13 GaAl12 GeAl12

Multimeric Complexes 1100 ( 100 53 ( 12 -7 ( 25 227 ( 43 63 ( 7 29 ( 21 190 ( 43 56 ( 7 20 ( 21

+5.7

+3 ( 1

a The original sources can be found in refs 118-120 and 122. The ∆Vq value for Al(H2O)5OH2+ is not reported but is consistent with a change in coordination.136 b abbreviations: ox ) oxalate; ssal ) sulfosalicylate; sal ) salicylate; mMal ) methylmalonate.

Usually, reduced charge density on a trivalent metal results in a more dissociative character for solvent exchange, such as is observed when a fully protonated metal ion (e.g., Fe(H2O)63+, ∆Vq ) -5.4 ( 0.4 cm3 mol-1, or Ga(H2O)63+, ∆Vq ) +5.0 ( 0.5 cm3 mol-1) is compared to its first conjugate base (Fe(H2O)5OH2+, ∆Vq ) +7.0 ( 0.5 cm3 mol-1, or Ga(H2O)5OH2+, ∆Vq ) +6.2 cm3 mol-1).123-128 The difference suggests that water exchange on the larger GaAl12 complex has less dissociative character than the fully hydrated ion, although the average charge density is lower. The paradox of reactivity for these -Keggin aluminum molecules is that a single atom substitution in the central M(O)4 site has profound effect on the rates of exchange of µ2-OH but not bound waters. For isotopic exchange into the bridges, the reactivity trend is GeAl12 > -Al13 > GaAl12, and the rates differ by at least a factor of ∼105. For the -Al13 and GaAl12 molecules, the data indicate that oxygen-isotope exchange from the bulk solution to the µ2-OH sites is independent of pH, albeit over a narrow range (4.5 < pH < 5.5). Yet, for all hydroxyl bridges and bound waters, the exchanging oxygens are three bonds removed from the central metal.

Figure 11. The rates of oxygen-isotope exchange into µ2-OHa and µ2-OHb in the -Al13, GaAl12, and GeAl12 vary by a factor of at least 105 (GeAl12 > -Al13 > GaAl12), yet the µ4-O sites are inert, rates are largely independent of pH, and the bound waters exchange at virtually identical rates. This paradox was explained by Rustad et al.,129 who hypothesized from computer simulations that a metastable intermediate forms that involves partial detachment of two Al(O)6 from the µ4-O to release part of the linked structures. The resulting partly detached moiety undergoes isotopic exchange via pathways that are similar to those affecting an inert-metal dimer, which parts of it resemble. Dissociation of one µ2-OH and hydration by a bulk water molecule forms an H3O2- bridge where oxygens can exchange rapidly. After the oxygens exchange positions in the H3O2- bridge, the reaction reverses and the partly detached structures collapses back into the stable MAl12 structure. In the Rustad et al. model,129 this partly detached structure forms constantly at a low concentration. The steady concentration depends critically on the strength of the bond from the central metal to the µ4-O, which explains the enormous range in reactivities. The µ4O, of course, remains unexchanged.

A reasonable explanation of the paradox was provided by Rustad et al.,129 who employed molecular-dynamics simulations and ab initio calculations to come up with a pathway for oxygen-isotopic exchange that involved a partly dissociated form of each molecule. They hypothesized that two Al(O)6 in the outer part of the structure detach from the µ4-O [Figure 11], creating an Al(µ2-OH)2Al dimer-like structure on the molecule that interacts strongly with bulk waters and partly dissociates. As the µ2-OH in this dimer dissociates, a H3O2- bridge forms [Figure 11] that can easily exchange for a bulk water, much like the transient H3O2- bridge that forms in dimers of trivalent metals [e.g., ref 123]. The extent to which this metastable moiety forms from the stable structure depends on the central metal and varies inversely with the strength of the M-(µ4-O) bond. Hence the intermediate forms in the order GeAl12 > -Al13 > GaAl12. The reaction is not proton catalyzed, and the rates are thus

10 Chemical Reviews, 2006, Vol. 106, No. 1

independent of pH. The exchange of bound waters takes place at the stable structures so the atom in the central M(O)4 site has little effect on the Al-(η-OH2) bond strength. Thus, the rates of exchange of bound waters are largely independent of the central metal, whereas the rates of exchange of hydroxyl bridges are enormously sensitive. The Rustad et al.129 hypothesis, although unproven, is consistent with measurements of the ∆Vq value for oxygenisotope exchange into the µ2-OHslow site in the GaAl12 molecule,122 which is the only µ2-OH for which ∆Vq values exist. If an expanded structure forms from the stable -Keggin structure at steady state, then one expects a large positive contribution of reaction volume to the experimental activation volume. Correspondingly, near 322 K the rates of exchange for the less labile set of bridging hydroxyls in the GaAl12 decrease by a factor of about two with increasing pressure from 0.1 to 350 MPa, consistent with ∆Vq ) +7 ( 1 cm3 mol-1. This ∆Vq parameter would include any expansion of the molecule to form a metastable intermediate and does suggest significant bond lengthening in the reaction. A similar set of 17O NMR studies were conducted on the Al30 molecule at conditions limited to pH ≈ 4.7 and 32-40 °C.80 Although it was impossible to determine rates for individual oxygen sites in the Al30, the 17O NMR peak positions fall into the same range as those for the MAl12 molecules. With the growth in intensities of these peaks used as a guide, rates of isotopic equilibration of the µ2-OH, µ3-OH, and bound water molecules fall broadly within the same range as the -Al13 and GaAl12 molecules. The µ2-OH and µ3-OH equilibrate within a couple of weeks in this temperature range, and the peak in the 17O NMR spectra that is assigned to bound water molecules varies in width with temperature in a similar fashion as that for other aluminum solutes (τ298 ex ≈ 0.01-0.0001 s). Thus, most of the bound waters on the Al30 probably exchange with bulk solution at rates that fall within the range observed for other aluminum complexes. However, signal from one anomalous group of four η-OH2 sites is not observed, indicating that these sites exchange at least a factor of 10 more rapidly than the other bound waters on the Al30.80 Phillips et al.80 speculated that these four waters were those cis to the µ3-OH formed by capping the δ-Al13 units [Figure 5]. Subsequent work has shown that these caps are particularly reactive (see below).

3.3. Kinetic Data for Other Ligand Exchanges on the MAl12 3.3.1. Reactions at Bound Water Molecules The most extensive data on ligand exchanges involving the MAl12 molecules are from Forde and Hynes,130 who employed stop-flow methods to determine the rate coefficients for reaction of a series of phenolic compounds [Scheme 1] with the -Al13. (In Scheme 1, the set of phenols are shown above a typical reaction written with catechol.) This study is particularly important because, along with Yu et al.,131 it establishes that the reaction rates at the large MAl12 ions proceed via familiar Id pathways and that one can estimate rate coefficients. Some of the phenolic compounds studied by Forde and Hynes130 are common in natural waters as breakdown products of lignin, and they bind strongly to aluminum hydroxide colloids. The authors presented a pathway that involves rapid formation of an electrostatic ion pair, followed by slow loss of a bound water and formation of a strong inner-sphere bond. The closure of the chelate

Casey Scheme 1

ring [Scheme 1] is relatively rapid, and the final step is rapid decomposition of the modified molecule to form aluminumphenolate monomer complexes. This rapid decomposition is consistent with the documented instability of -Al13 in phenol-rich soil solutions (see below). Yu et al.131 examined pathways whereby fluoride ion replaces bound waters and hydroxyl bridges in the GaAl12 molecule. The first and most conspicuous reaction was replacement of a bound water molecule on the GaAl12 molecule: kF278 -

GaAl127+ + F- 98 GaAlF6+ + H2O

(2)

The experimental conditions (4.0 < pH < 5.5) were chosen so that virtually all of the dissolved fluoride was present as the F- ion, the GaAl12 molecule was fully protonated, and the replaceable functional groups were usually in excess of the fluoride concentration. The potentiometry indicated a first-order reaction in both fluoride and GaAl12, as expected, and no pH dependence. The reaction at low temperatures (278 K) was sufficiently slow to follow using a fluoridespecific electrode [Figure 12]. Enough data exist to establish the reasonableness of the Eigen-Wilkins-Tamm mechanism for ligand exchange by comparing the measured rate coefficients with one estimated from the product of an equilibrium constant (Kos) for

Large Aqueous Aluminum Hydroxide Molecules

Chemical Reviews, 2006, Vol. 106, No. 1 11

Figure 12. Changes in the concentration of dissolved fluoride in a 5 mL solution of ∑[F] ) 6 × 10-5 M before and after injection of 0.100 mL of 1.7 × 10-3 M GaAl12 at pH ) 4.5 and 278 K.131 Bound waters on the GaAl12 molecule are replaced by fluoride ions with nearly the same rate law as that on the Al(H2O)63+(aq) ion; that is, the rates are first order in [F-] and GaAl12 concentrations and largely controlled by the rates of solvolysis. Thus, rates of ligand substitution at the bound waters on the MAl12 molecules are similar to monomeric aqueous complexes [Table 3]. A slower subsequent reaction is fluoride replacement of µ2-OH, which can be followed by 19F NMR.131,137

formation of the outer-sphere precursor and the rate of water exchange reported for the MAl12: kcalcd ) KoskHT 2O. The constant Kos is calculated from a version of the Fuoss equation:128,133

[

] [

]

-z+z-e2 -z+z-e2κ 4000Nπa3 exp Kos ) exp 3 4π0kTa 4π0kT(1 + κa) (3) where κ ) x20002NI/(0kT), e is the elementary charge in coulombs, k is Boltzmann’s constant in J K-1, N is Avogadro’s number (mol-1), 0 is the vacuum permittivity (8.854 × 10-12 J-1 C2 m-1),  is the dielectric constant of water at temperature, the parameter a is the distance of closest approach of the ions (5 × 10-10 m), T is temperature in kelvin, I is ionic strength (mol L-1), and z+ and z- are the charges of the ions. The value of kHT 2O is given or can be estimated from the data in Table 2. All of the phenolic ligands used by Forde and Hynes130 are uncharged at the experimental pH and have Kos ) 0.32 M-1. With use of kH298 ) 1100 s-1, the second-order rate 2O coefficient for ligand exchange is estimated to be k ≈ 350 M-1 s-1. The ratios of kexpt/kcalcd vary around unity [Table 3], as expected. The experiments of Yu et al.131 were carried Table 3. A Summary of Kinetic Data for Reaction of Phenol Ligands with E-Al13 in Aqueous Solution at 25 °C and I ) 1.0 M from ref 130 and for Reaction of Fluoride Ion with the GaAl12 Molecule at 278 K and I ) 0.6 M from ref 131 k, M-1 s-1

ligand catechol L-dopa salicylic acid maltol gallic-acid ester chlorogenic acid

-Al13, kH298 ) 1100 s-1 2O 23.0 ( 0.08 33.3 ( 0.7 119 ( 3.4 96.6 ( 3.4 192 ( 6.8 489 ( 16.2

F-

GaAl12, kH278 ) 36 s-1 2O 7.05 ( 1.2

F-

Al3+,

kH298 2O

) 1.29

Kos ) 0.15 1.55 ( 0.35 s-1;

kexpt/kcalcd 0.07 0.10 0.34 0.28 0.55 1.40 0.5 M-1 8.0

out at lower temperature (278 K) and ionic strength (I ) ≈ 36 s-1 is estimated from 0.6). At these conditions, kH278 2O the data in Table 2, and the average rate coefficient for eq

2 can be calculated from Table 2 in Yu et al.,131 yielding -1 s-1. To calculate the ion-pairing kF278 - ) 7.05 ( 1.2 M constant, we use the average charge on an Al(III) metal in the GaAl12 molecule (+0.54), rather than the +7 total ion charge, to get Kos ≈ 0.5 M-1 or kexpt/kcalcd ≈ 0.4, which is fully consistent with the data for the phenols130 and consistent with an Eigen-Wilkins-Tamm mechanism. The loss of a bound water molecule controls the rate of the overall ligandexchange reaction, and this rate is relatively unaffected by the incoming ligand. For comparison, also included in Table 3 is the ratio kexpt/kcalcd for the reaction kF298 -

Al3+ + F- 98 AlF+2

(4)

using data from refs 133 and 134 (see also ref 216) but recalculated to be consistent with the Al13 and GaAl12 results reported here. This recalculation follows the lead of Forde and Hynes130 and ignores the 3/4 statistical factor that is typically employed and sets a ) 5 × 10-10 m for all molecules and ligands. Given these approximations, we calculate Kos ) 0.15 M-1 at I ) 0.1 and 298 K, and kexpt/ kcalcd ≈ 8.0. This fluoridation reaction is well accepted as proceeding via an Eigen-Wilkins-Tamm mechanism and indicates that the reactions for the MAl12 molecules are probably also Id. By extrapolation, ligand-exchange reactions on larger molecules (see below), and perhaps even aluminum colloids, will be controlled by rates of water dissociation, which fall into the familiar millisecond range for aluminum monomers. Oxygen exchange will involve concerted motions of the entering and leaving groups, and it is unsurprising that ligand exchanges on Al(III) in these large molecules have a considerable dissociative character. As pointed out by Swaddle,135 the ionic radius [53.5 pm] of Al(III) is smaller than the hexagonal interstices [57 pm] in an array of cubicclose-packed spheres with radii of 138 pm, that of water in ice, so aluminum is much more likely to decrease coordination to oxygens than to tolerate an increase in coordination number [e.g., ref 136]. In summary, the data for exchange of a ligand for a bound water molecule on the MAl12 series of molecules appear to be consistent with the results for monomers and Id mechanisms, although high-pressure data only exist for a single molecule.

3.3.2. Reactions at Bridging Hydroxyls A striking result is the extraordinary difference in reactivity of the µ2-OHfast and µ2-OHslow sites within the -Al13, GaAl12, and GeAl12 molecules [Table 1]. Clear evidence for the different reaction rates is particularly evident in the fluorideuptake experiments of Yu et al.131 who followed the reaction using 19F NMR and potentiometry and saw peaks in the 19F NMR spectra that were assigned to both bridging (-131 to -138 ppm) and nonbridging fluorides (-148 ppm) on the GaAl12 molecule. The results were interpreted to indicate parallel and possibly even reversible transfer of fluoride from nonbridging sites to the two bridging sites. The essential feature of the rate law was that fluoride replaces bound water molecules within seconds at 278 K and that the two µ2-OH sites form more slowly and at different rates. The nonbridging fluoride peaks decrease in intensity with time as peaks assigned to the bridging fluorides increase. Notably, rates of fluoride replacement of µ2-OH sites are 101-103 times more rapid than the rates of oxygen

12 Chemical Reviews, 2006, Vol. 106, No. 1

exchange with bulk waters into the bridging hydroxyls. Allouche and Taulelle137 found that fluoridation of the Al13 molecule accelerated conversion to other isomers (see below). Most importantly, they found that the rate of fluoridation of the Al13 is greater than that of the GaAl12, as one expects from the rates of oxygen-exchange into the respective µ2-OH bridges [Table 1]. Allouche and Taulelle137 also studied fluoridated Al13 crystals and found peaks assignable to bridging fluorides at -132 and -134 ppm, consistent with the earlier results. Phillips et al.138 conducted 19F{27Al}-transfer of populations in double resonance (TRAPDOR) experiments in fluoridesubstituted GaAl12-selenate crystals and found an additional peak near -124 ppm, as well as three peaks in the range -130 to -139 ppm that are assignable to the fluoridated bridges, suggesting that a richer array of 19F environments can be detected with advanced spectroscopy. There is thus broad agreement between the results of fluoridation with these aluminum polyoxocations and aluminum hydroxide minerals. Site-specific reaction rates can be established that are probably relevant to environmental materials such as aluminous clays.139

3.3.3. Dissociation Rates and Pathways The fact that four central µ4-oxo sites remain isotopically normal during 17O injection experiments indicates that exchange rate laws for all other atom positions in the -Al13, GaAl12, and GeAl12 molecules are measured without complete dissociation. The rates of complete dissociation of the -Al13 alone have been measured140-142 and indicate that the molecule experiences many tens to hundreds of isotopic exchanges in the µ2-OH sites before it dissociates completely. Complete dissociation of the -Al13 molecule is first-order in proton concentration in acidic solutions and second-order in proton concentration at pH < 2.5. The activation energies for dissolution are also relatively small (13.3 ( 1.9 and 44.9 ( 4.9 kJ mol-1 140) and vary with solution pH, indicating that there are contributions of protonation enthalpy to the temperature dependence of the reaction rate even in pH condtions (2 < pH < 3.5) that are well away from the pKa of the bound waters (pH ≈ 6.5). Furrer et al.140 interpreted the dominant mechanism at pH > 2.5 to involve a protonated µ2-OH site and the mechanism at pH < 2.5, which is secondorder in protons, to involve two adjacent protonated bridges. The newer 17O NMR data indicate that both the µ2-OHslow and µ2-OHfast sites exchange many times before the -Al13 molecule dissociates and that these exchanges are independent of pH (however at slightly higher pH conditions than the Furrer et al.140 study). Besides the mismatch between the pH dependence of the dissolution and the pH independence of the isotope-exchange reactions, the activation energies for dissolution are much smaller than those for oxygen-isotope exchange at the hydroxyl bridges (∆Hq ) 104 and 204 kJ mol-1, Table 1). Furthermore, because the µ4-O remains inert even as all other oxygen sites exchange isotopes with the aqueous solution, dissociation of µ2-OH sites cannot be key to the dissolution mechanism. One model to account for the data is that protonation of the inert µ4-O sites, followed by hydration, causes the molecule to dissociate. Casey et al.118 reached this conclusion largely by elimination, since the time scales for dissociating all other oxygens are too small relative to dissolution and the rate parameters (activation energies and rate orders) are too different to account for the dissolution experiments.

Casey

The pH dependencies for dissociation and µ2-OH bridge cleavage are also wrong for these ruptures to control dissociation of the molecule. Experiments in the acid-pH direction are limited by dissolution of the molecule in lengthy experiments, and experiments in more basic solutions are difficult because the molecules irreversibly link as the pH exceeds the dissociation pH. Nevertheless, the rates of oxygen-isotope exchange between bulk solution and the Al13 and GaAl12 are independent of pH (over about 0.5 pH units) as long as pH < 6.5.117-119 The GeAl12 has a clear pH dependence to the exchange rates of both the bound waters and the one set of µ2-OH sites (µ2-OHslow) that can be studied. The pH dependence measured for oxygen-isotope exchange rates in the GeAl12 is because the experimentally accessible pH range is near to the pKa value for deprotonation of the bound water molecules.143,144 Thus, a bound hydroxyl labilizes both the bound waters and the µ2-OH, as is observed for the fluoride-substituted GaAl12, which is isoelectronic.131

3.3.4. Formation Pathways The conventional model for forming the -Al13, GaAl12, and GeAl12 molecules involves the initial formation of a set of Al3(µ3-OH)(µ2-OH)3(H2O)95+ trimers that then link together with a tetrahedral M(OH)4 ion and assemble into the -Al13 and MAl12 structures [see refs 29, 52, 61, 89, 145, and 146]. Direct data are few, but Akitt and Farthing147 concluded that the assembling molecules were (H2O)4Al(µ2-OH)2(H2O)44+ dimers, based largely on their existence in solution with the -Al13. A stable tetramer was postulated from X-ray scattering studies by Michot et al.62 for the -Ga13. Such a tetramer must certainly form in some stage of the assembly, but neither the planar Al(O)6 trimers nor the tetramer has been crystallized, and these molecules probably comprise some of the missing 27Al NMR intensity in a concentrated solution [e.g., refs 67 and 147]. The trimer, in contrast, is inferred from potentiometric studies148,149 and is assigned an appropriate stoichiometry, Al3(µ3-OH)(µ2-OH)35+. Evidence from these potentiometric titrations indicate that the Al3(µ3-OH)(µ2-OH)35+ trimer exists in solutions at 4 < pH < 5 but never reaches a large concentration, even in concentrated aluminum solutions, probably because the trimers condense to form larger clusters. There is some spectroscopic hint of the trimers because there is sometimes a relatively narrow peak in concentrated solutions at +11 ppm in the 27Al NMR spectra, particularly for solutions in which -Al 13 is being formed or decomposed [ref 81 and unpublished data]. Assignment of this peak to the planar trimer is suggested by the similarity of the chemical shift to that of the Al(O)6 of the -Al13 complex, which is composed of four of these trimers linked together, and the small peak width, consistent with a relatively small complex. In any case, it is clear that the -Al13 forms via a rapid pathway during a titration, as the yield depends on the rate of base addition30,146 and does not reach equilibrium in a typical synthesis. Allouche and Taullele79 and Shafran et al.68,69 conducted a detailed study of the -Al13 f Al30 conversion and argue that capping the -Al13 by an Al(O)6 stabilizes the structure, which matches some new evidence for metal exchanges. The slow step involves the rotation of the cap by 60° by simultaneous rupture of the more-reactive µ2-OH bridges, followed by monomer addition and dimerization of the δ-Al13. The Allouche and Taullele79 supposition is consistent with the Son et al.87 observation that the Al(O)6 caps on the δ-Al13 are particularly reactive to ligand substitution.

Large Aqueous Aluminum Hydroxide Molecules

Remarkably, Fu¨rrer found that a solution of -Al13 is metastable for at least 12 years at ambient conditions but that over this time scale, the -Al13 converts to the Al30 plus some protons [see Figure 8 in ref 81]. Clearly the elevated temperatures that are most commonly used in synthesis64,68-70 accelerate the conversion of the -Al13 to the Al30, but the same pathways act at ambient conditions, albeit more slowly. It would involve a substantial set of experiments, but the pH dependence of the conversion could considerably advance the field.

4. Uses and Environmental Significance of the Aqueous Aluminum Hydroxide Clusters The -Al13 (and to a much lesser extent the Al30) is a major constituent of the industrial chemical aluminum chlorohydrate (also referred to as polyaluminum chloride). The soluble reagent that has a ratio of OH to Al of about 2.3-2.5 has the most-common stoichiometry, Al2(OH)5Cl [see refs 2, 150, and 151]. The ensemble of polymers in this chemical changes with aging in water [e.g., ref 152], just as in titrations [e.g., refs 68, 69, 84, and 147] and species ranging from 1000 to 10 000 Da can be inferred from size-exclusion chromatography2,82 and reaction with size-sensitive dyes. As mentioned above, none of these larger molecules, save for the Al30, can be assigned in 27Al NMR. Smaller aluminum hydroxide polymers, such as trimers, are inferred experimentally [e.g., ref 147] and from calculations153 and make up some of the broad unresolvable peak in 27Al NMR spectra. Although there are suspicions about toxicity from these products, aluminum in blood plasma is usually at a common low level even in heavy users,154,155 although antiperspirantinduced hyperaluminemia is reported [ref 15, see also refs 156 and 157]. In no recorded case is a polymeric aluminum cluster the toxicant, although Rao and Rao158 report -Al13 in excised synaptosomes from rat brains. Entry into the toxicology literature is found in the November 2001 issue of Inorganic Biochemistry and review articles14,159 dealing with aluminum biochemistry and speciation in body fluids [see also ref 160].

4.1. Water Treatment Polymeric aluminum hydroxide reagents are essential to industries, such as paper production, that release large amounts of tannins, phenols, and organic acids to wastewater [e.g., refs 13, 161, and 162]. Aluminum chlorohydrate is added to maintain large concentrations of -Al13 in solution, which then flocculates to form cationic sols that adsorb the organic solutes. As the flocs settle, the water is clarified, thereby improving color and taste [e.g., refs 163 and 164], and this treatment can even inactivate viruses [e.g., ref 165]. Alumino-silicate polymers are also used; although the polymer structures are unknown, they probably contain a silicate--Al13 ternary complex [e.g., refs 166-168].

4.2. Pillaring Agents The -Al13 molecule is among the most common pillaring agent for clays [see refs 169-173] and anionic layered solids, including titanates, manganates,174,175 and molybdates.176 Briefly, the -Al13 molecule props open the interlayer spacing of the material, which increases the microporosity [e.g., ref 177], and affects the Lewis and Brønsted acidity and proton conduction of the material [e.g., refs 178 and 179] and the ion-exchange properties [e.g., ref 179; see ref 172 for review).

Chemical Reviews, 2006, Vol. 106, No. 1 13

The pillared solid is often calcined to dehydrate the polyoxocation [e.g., ref 180], and the pillars also allow addition of interlayer dopants with particularly useful properties, such as catalytic transition-metal complexes [e.g., refs 181-184] or hydrophobic organics to pick up pollutants [e.g., ref 11]. There have been several attempts to expand the number of aluminum polyoxocations to include the flat-Al13-heidi,10,185 the GaAl12,6-8,186 and Ga13.6,8,187 There are literally hundreds of papers on this subject, and the interested reader is referred to refs 169-172 and 188-190 for entry into this literature.

4.3. Environmental Significance Environmental interest in the -Al13 molecule stems largely from its reported phytoxicity [e.g., refs 191-194] and the potential toxicity to fish [ref 195; see also refs 196-198]. It is also considered to provide a vector for transporting pollutants199,200 and for influencing pesticides.201 Natural conditions for forming the -Al13 molecule would be when acidic and low-organic-acid waters mix rapidly with a higher-pH solution. Such an environment could be found as dilution of acid rainfall percolating through soil into a higher-pH stream200 or over a limestone terrain.202 However, the critical pH window (4.5 < pH < 6.5) is difficult to sample during a mixing event and the -Al13 is suppressed by constituents common in natural waters. It dissociates rapidly by bonding to phenolic compounds130,203-205 and metals [e.g., ref 206] and is flocculated by anions such as sulfate166,207 or humic acids [e.g., refs 203 and 208-210]. Although there are doubts as to whether these polyoxocations exist in natural waters,211 the -Al13 molecule has been detected in soil solutions,212 and its presence is inferred from the 27Al NMR spectra of pollutant floc.200 As mentioned earlier, the cubane-like clusters are particularly interesting because they relate closely to a hypothetical brucite packing [see ref 96]. The thermodynamically stable structure of aluminum hydroxide is gibbsite, which has sixmembered rings of edge-shared Al(O)6 octahedra and no µ3-OH. This gibbsite structural arrangement is central to dioctahedral clays of trivalent metals [e.g., ref 213]. The fact that the flat-Al13 and Al8 molecules form quickly in concentrated aqueous solutions and that hydrotalcite phases form in dilute solutions suggests that the cubane-like structures may be stabilized in nature by substitution of divalent metals for aluminum. Synthesis of a model complex that is metastable in solution and could be used for kinetic studies would be a major advance.

5. Conclusions In the five decades since Johansson25-28 isolated and crystallized the -Al13 polyoxocation, hundreds of scientific studies have been conducted on aqueous aluminum hydroxide polyoxocations, largely because of their use as clay pillars. Interest in these aluminum clusters is resurging now because of their use as kinetic models for understanding aqueous environmental reactions. The field is reinvigorated by the persistent effort by a handful of research groups that have isolated and separated the large polymers that have long been suspected to exist in hydrolyzed aluminum solutions. The isolation and structural characterization of the δ-Al13, Al30, and cubane-like clusters (e.g., the flat-Al13) provide new kinetic models for experiments in aqueous solutions. Furthermore, the rate at which these molecules are being found is accelerating, as the new methods of supramolecular complexation and separation help uncover clusters such as the Ga32.

14 Chemical Reviews, 2006, Vol. 106, No. 1

This work is important because these 1-2 nm aqueous clusters provide windows into the complicated processes that affect minerals and the world around us. Using the aluminum hydroxide clusters and modern spectroscopies, scientists can now follow isotope-exchange and hydrolysis reactions in ways that are impossible with aluminum hydroxide colloids, and they are uncovering surprising reactivity trends. Furthermore, the experimental results are at an appropriate scale to test algorithms of computational chemistry. Thus, experimental probing of these 1-2 nm aluminum hydroxide molecules advances our ability to predict reactive properties for many materials in water. In this sense, the aluminum polyoxocations are distinct from other polyoxometalates (e.g., tungstates) that have limited similarity to natural materials. Aluminum hydroxide substrates are everywhere in our lives, and we have little quantitative understanding as to how the form, dissolve, and react with other solutes in natural waters.

6. Acknowledgments The author is particularly grateful to Profs. Brian Phillips and James Rustad for discussions and suggestions and to two referees who provided detailed suggestions to improve the text. Support for this research was from the Department of Energy (Grant Number DE-FG03-96ER 14629), from the American Chemical Society (PRF Grant 40412-AC2), and from the National Science Foundation (Grant EAR 0101246).

7. References (1) Rustad, J. R. ReV. Mineral. Geochem. 2001, 42, 169. (2) Fitzgerald, J. J.; Rosenberg, A. H. In Antiperspirants and Deodorants, 2nd ed.; Laden, K., Ed.; Marcel Dekker: New York, 1999; p 83. (3) Beer, R.; Steiger, R.; Banerjee, D.; Furrer, G.; Rentsch, D.; Studer, B. Recording sheets for ink-jet printing. U.S. Patent 20050053735 to Ilford Imaging, Switzerland GmbH, 2005; 18 pp. (4) Furrer, G.; Banerjee, D.; Rentsch, D.; Studer, B.; Beer, R. and Steiger, R. Recording sheet for ink jet printing. Eur. Pat. Appl. EP 1512544, 2005; 14 pp. (5) van Bruggen, M. P. B.; Donker, M.; Lekkerkerker, H. N. W.; Hughes, T. L. Colloids Surf., A 1999, 150, 115. (6) Bradley, S. M.; Kydd, R. A. Catal. Lett. 1991, 8, 185. (7) Bradley, S. M.; Kydd, R. A. J. Catal. 1993, 141, 239. (8) Bradley, S. M.; Kydd, R. A. J. Catal. 1993, 142, 448. (9) Bradley, S. M., Kydd, R. A.; Brandt, K. K. Stud. Surf. Sci. Catal. 1992, 73, 287. (10) Vicente, M. A.; Lambert, J.-F. Clays Clay Miner. 2003, 51, 168. (11) Montarge`s, E.; Moreau, A.; Michot, L. J. Appl. Clay Sci. 1998, 13, 165. (12) Bertram, R.; Gessner, W.; Mu¨ller, D.; Danner, M. Acta Hydrochim. Hydrobiol. 1994, 22b, 265. (13) Shen, Y.-H.; Dempsey, B. A. EnViron. Int. 1998, 24, 899. (14) Gupta, R. K. AdV. Drug DeliVery ReV. 1998, 32, 155. (15) Lindblad, E. B. Immunol. Cell Biol. 2004, 82, 497. (16) Pope, M. T. In ComprehensiVe Coordination Chemistry II: From Biology to Nanotechnology; Wedd, A. G., Ed.; Elsevier Ltd.: Oxford, U.K., 2004; Vol. 4, p 635. (17) Hill, C. L. In ComprehensiVe Coordination Chemistry-II: From Biology to Nanotechnology; Wedd, A. G., Ed.; Elsevier: Oxford, U.K. 2004; Vol. 4, p 679. (18) Heath, S. L.; Jordan, P. A.; Johnson, I. D.; Moore, J. R.; Powell, A. K.; Helliwell, M. J. Inorg. Biochem. 1995, 59, 785. (19) Jordan, P. A.; Clayden, N. J.; Heath, S. L.; Moore, G. R.; Powell, A. K.; Tapparo, A. Coord. Chem. Rev. 1996, 149, 281. (20) Seichter, W.; Mo¨gel, H.-J.; Brand, P.; Salah, D. Eur. J. Inorg. Chem. 1998, 1998, 795. (21) Karlsson, M. Ph.D. thesis in Chemistry, University of Umeå, Sweden, 1998. (22) Schmitt, W.; Jordan, P. A.; Henderson, R. K.; Moore, G. R.; Anson, C. E.; Powell, A. K. Coord. Chem. ReV. 2002, 228, 115-126. (23) Abragam, A. The Principles of Nuclear Magnetism; Oxford University Press: Oxford, U.K., 1961. (24) Akitt, J. W.; Mann, B. E. NMR and Chemistry, 4th ed.; Stanley Thornes: New York, 2000. (25) Johansson, G. Acta Chem. Scand. 1960, 14, 771-773.

Casey (26) Johansson, G.; Lundgren, G.; Sille´n, L. G.; So¨derquist, R. Acta Chem. Scand. 1960, 14, 769. (27) Johansson, G. Ark. Kemi 1963, 20, 305. (28) Johansson, G. Ark. Kemi 1963, 20, 321. (29) Akitt, J. W.; Farthing, A. J. Chem. Soc., Dalton Trans. 1981, 1981, 1624. (30) Bottero, J. Y.; Cases, J. M.; Fiessinger, F.; Poirier, J. E. J. Phys. Chem. 1980, 84, 2933. (31) Akitt, J. W. Prog. NMR Spectrosc. 1989, 21, 1. (32) Bertsch, P. M.; Layton, W. J.; Barnhisel, R. I.. Soil Sci. Soc. Am. J. 1986, 50, 1449. (33) Bertsch, P. M.; Thomas, G. W.; Barnhisel, R. I. Soil Sci. Soc. Am. J. 1986, 50, 825. (34) Kloprogge, J. T.; Seykens, D.; Jansen, J. B. H.; Geus, J. W. J. NonCryst. Solids 1992, 142, 94. (35) Kloprogge, J. T.; Seykens, D.; Geus, J. W.; Jansen, J. B. H. J. NonCryst. Solids 1992, 142, 87. (36) Simpson, S. L.; Powell, K. J.; Nilsson, Nils H. S. Anal. Chim. Acta 1997, 343, 19. (37) Bertsch, P. M.; Anderson, M. A.; Layton, W. J. Magn. Reson. Chem. 1989, 27, 283. (38) Ferguson, J. H.; Kustin, K.; Phipps, A. Inorg. Chim. Acta 1980, 43, 49. (39) Parker, D. R.; Bertsch, P. M. EnViron. Sci. Technol. 1992, 26, 908. (40) Parker, D. R.; Bertsch, P. M. EnViron. Sci. Technol. 1992, 26, 914. (41) Hils, A.; Grote, M.; Janssen, E.; Eichhorn, J. Fresenius’ J. Anal. Chem. 1999, 364, 457. (42) Bertsch, P. M.; Parker, D. R. In The EnVironmental Chemistry of Aluminum, 2nd ed.; Sposito, G., Ed.; CRC Press: Boca Raton, FL, 1996; p 117. (43) Akitt, J. W.; Elders, J. M.; Letellier, P. J. Chem. Soc., Faraday Trans. 1987, 83, 1725. (44) Baes, C. F.; Mesmer, R. E. The Hydrolysis of Cations; John Wiley: New York, 1976. (45) Plyasunov, A. Y.; Grenthe, I. Geochim. Cosmochim. Acta 1994, 58, 3561. (46) Scho¨nherr, S.; Go¨rz, H. Z. Anorg. Allg. Chem. 1983, 503, 37. (47) Thomas, B.; Go¨rz, H.; Scho¨nherr, S. Z. Chem. 1987, 27, 183. (48) Bradley, S. M.; Kydd, R. A.; Yamdagni, R. J. Chem. Soc., Dalton Trans. 1990, 1990, 413. (49) Bradley, S. M.; Kydd, R. A.; Yamdagni, R. Magn. Reson. Chem. 1990, 28, 746. (50) Bradley, S. M.; Kydd, R. A.; Yamdagni, R. J. Chem. Soc., Dalton Trans. 1990, 1990, 2653. (51) Bradley, S. M.; Kydd, R. A.; Fyfe, C. A. Inorg. Chem. 1992, 31, 1181. (52) Bradley, S. M.; Kydd, R. A. J. Chem. Soc., Dalton Trans. 1993, 1993, 2407. (53) Bradley, S. M.; Lehr, C. R.; Kydd, R. A. J. Chem. Soc., Dalton Trans. 1993, 1993, 2415. (54) Bradley, S. M.; Kydd, R. A.; Howe, R. F. J. Colloid Interface Sci. 1993, 159, 405. (55) Bertram, R.; Scho¨nherr, S.; Go¨rz, H. Z. Chem. 1984, 24, 225. (56) Go¨rz, H.; Scho¨nherr, S.; Pertlik, F. Monatsh. Chem. 1991, 122, 759. (57) Lee, A. P.; Phillips, B. L.; Olmstead, M. M.; Casey, W. H. Inorg. Chem. 2001, 40, 4485. (58) Kudynska, J.; Buckmaster, H. A.; Kawano, K.; Bradley, S. M.; Kydd, R. A. J. Chem. Phys. 1993, 99, 3329. (59) Nagy, J. B.; Bertrand, J.-C.; Palinko, I.; Kirisci, I. J. Chem. Soc., Chem. Commun. 1995, 1995, 2269. (60) Oszko´, A.; Kiss, J.; Kiricsi, I. Phys. Chem. Chem. Phys. 1999, 1, 2565. (61) Parker, W. O., Jr.; Millini, R.; Kiricsi, I. Inorg. Chem. 1997, 36, 571. (62) Michot, L. J.; Montarge`s-Pelletier, E.; Lartiges, B. S.; d’Espinose de la Caillerie, B.; Briois, V. J. Am. Chem. Soc. 2000, 122, 6048. (63) Gerasko, O. A.; Mainicheva, E. A.; Naumov, D. Yu.; Kuratieva, N. V.; Sokolov, M. N.; Fedin, V. P. Inorg. Chem. 2005, 44, 4133. (64) Fu, G.; Nazar, L. F.; Bain, A. D. Chem. Mater. 1991, 3, 602. (65) Nazar, L. F.; Fu, G.; Bain, A. D. J. Chem. Soc., Chem. Commun. 1992, 1992, 251. (66) Akitt, J. W.; Mann, B. E. J. Magnet. Reson. 1981, 44, 584. (67) Akitt, J. W.; Elders, J. M.; Fontaine, X. L. R.; Kundu, A. K. J. Chem. Soc., Dalton Trans. 1989, 1989, 1889. (68) Shafran, K. L.; Perry, C. C. Dalton Trans. 2005, 2005, 2098. (69) Shafran, K. L.; Deschaume, O.; Perry, C. C. AdV. Eng. Mater. 2004, 6, 836. (70) Rowsell, J.; Nazar, L. F. J. Am. Chem. Soc. 2000, 122, 3777. (71) Allouche, L.; Ge´rardin, C.; Loiseau, T.; Fe´rey, G.; Taulelle, F. Angew. Chem., Int. Ed. 2000, 39, 511. (72) Turco, G. Compt. Rend. 1964, 258, 3331. (73) Pauling, L. Z. Kristallogr. 1933, 84, 442. (74) Bartl, H. Neues Jahrb. Mineral., Monatsh. 1970, 12, 552.

Large Aqueous Aluminum Hydroxide Molecules (75) Von Lampe, F.; Mueller, D.; Gessner, W.; Grimmer, A. R.; Scheler, G. Z. Anorg. Allg. Chem. 1982, 489, 16. (76) Kloprogge, J. T.; Frost, R. L. Spectrochim. Acta, Part A 1999, 55A, 1505. (77) Kunwar, A. C.; Thompson, A. R.; Gutowsky, H. S.; Oldfield, E. J. Magn. Reson. 1984, 60, 467. (78) Zhou, B.; Sherriff, B. L.; Taulelle, F.; Wu, G. Can. Mineral. 2003, 41, 891. (79) Allouche, L.; Taulelle, F. Inorg. Chem. Commun. 2003, 6, 1167. (80) Phillips, B. L.; Lee, A. P.; Casey, W. H. Geochim. Cosmochim. Acta 2003, 67, 2725. (81) Casey, W. H.; Phillips, B. L.; Furrer, G. F. ReV. Mineral. Geochem. 2001, 44, 167. (82) Shafran, K. L.; Deschaume, O.; Perry, C. C. J. Mater. Chem. 2005, 15, 3415. (83) Allouche, L.; Huguenard, C.; Taulelle, F. J. Phys. Chem. Solids 2001, 62, 1525. (84) Son, J.-H.; Choi, H.; Kwon, Y.-U. J. Am. Chem. Soc. 2000, 122, 7432. (85) Son, J.-H.; Kwon, Y.-U. Inorg. Chem. 2004, 43, 1929. (86) Son, J.-H.; Kwon, Y.-U. Inorg. Chim. Acta 2005, 358, 310. (87) Son, J.-H.; Kwon, Y.-U.; Han, O. H. Inorg. Chem. 2003, 42, 4153. (88) Son, Jung-Ho; Kwon, Young-Uk. Bull. Korean Chem. Soc. 2001, 22, 1224. (89) Jolivet, J.-P. Metal Oxide Chemistry and Synthesis; Wiley: New York, 2000. (90) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. AdVanced Inorganic Chemistry; Wiley-Interscience: New York, 1999. (91) Akitt, J. W.; Farthing, A. J. Chem. Soc., Dalton Trans. 1981, 1981, 1624. (92) O ¨ hman, L.-O. Inorg. Chem. 1989, 28, 3629. (93) Carrier, X. J.-B.; Lambert, J.-F.; Che, M. J. Am. Chem. Soc. 1997, 119, 10137. (94) Cowan, J. J.; Bailey, A. J.; Heintz, R. A.; Do, B. T.; Hardcastle, K. I.; Hill, C. L.; Weinstock, I. A. Inorg. Chem. 2001, 40, 6666. (95) Carrier, X.; d’Espinose de la Caillerie, J.-B.; Lambert, J.-F.; Che, M. J. Am. Chem. Soc. 1999, 121, 3377. (96) Goodwin, J. C.; Teat, S. J.; Heath, S. C. Angew. Chem., Int. Ed. 2004, 43, 4037. (97) Casey, W. H.; Olmstead, M. M.; Phillips, B. L. Inorg. Chem. 2005, 44, 4888. (98) Breuil, H. Ann. Chim. (Paris) 1965, 10, 467. (99) Powell, A. K.; Heath, S. L. Coord. Chem. ReV. 1996, 149, 59. (100) Heath, S. L.; Jordan, P. A.; Johnson, I. D.; Moore, G. R.; Powell, A. K.; Helliwell, M. J. Inorg. Biochem. 1995, 59, 785. (101) Schmitt, W.; Baissa, E.; Mandel, A.; Anson, C. E.; Powell, A. K. Angew. Chem. 2001, 40, 3578. (102) Feng, T. L.; Gurian, P. L.; Healy, M. D.; Barron, A. R. Inorg. Chem. 1990, 29, 408. (103) Pasynkiewicz, S. Polyhedron 1990, 9, 429. (104) Roesky, H. W.; Walawalkar, M. G.; Murugavel, R. Acc. Chem. Res. 2001, 34, 201. (105) Landry, C. C.; Pappe´, N.; Mason, M. R.; Apblett, A. W.; Tyler, A. N.; MacInnes, A. N.; Barron, A. R. Mater. Chem. 1995, 5, 331. (106) Bethley, C. E.; Aitken, C. L.; Harlan, C. J.; Koide, Y.; Bott, S. G.; Barron, A. R. Organometallics 1997, 16, 329. (107) Koide, Y.; Barron, A. R. Organometallics 1995, 14, 4026. (108) Callender, R. L.; Harlan, C. J.; Shapiro, N. M.; Jones, C. D.; Callahan, D. L.; Wiesner, M. R.; MacQueen, D. B.; Cook, R.; Barron, A. R. Chem. Mater. 1997, 9, 2418. (109) Narayanan, R.; Laine, R. M. J. Mater. Chem. 2000, 10, 2097. (110) Drljaca, A.; Hardie, M. J.; Raston, C. L. J. Chem. Soc., Dalton Trans. 1999, 1999, 3639. (111) Hardie, M. J.; Raston, C. L. J. Chem. Soc., Dalton. Trans. 2000, 2000, 2483. (112) Rather, E.; Gatlin, J. T.; Nixon, P. G.; Tsukamoto, T.; Kravtsov, V.; Johnson, D. W. J. Am. Chem. Soc. 2005, 127, 3242. (113) Hiemstra, T.; Yong, H.; Van Riemsdijk, W. H. Langmuir 1999, 15, 5942. (114) Sposito, G. The Surface Chemistry of Natural Particles; Oxford University: Oxford, U.K., 2004. (115) Swift, T. J.; Connick, R. E. J. Chem. Phys. 1962, 37, 307. (116) Thompson, A. R.; Kunwar, A. C.; Gutowsky, H. S.; Oldfield, E. J. Chem. Soc., Dalton Trans. 1987, 1987, 2317. (117) Phillips, B. L.; Casey, W. H.; Karlsson, M. Nature 2000, 404, 379. (118) Casey, W. H.; Phillips, B. L.; Karlsson, M.; Nordin, S.; Nordin, J. P.; Sullivan, D. J.; Neugebauer-Crawford, S. Geochim. Cosmochim. Acta 2000, 64, 2951. (119) Casey, W. H.; Phillips, B. L. Geochim. Cosmochim. Acta 2001, 65, 705. (120) Lee, A. P.; Phillips, B. L.; Casey, W. H. Geochim. Cosmochim. Acta 2002, 66, 577.

Chemical Reviews, 2006, Vol. 106, No. 1 15 (121) Springborg, J. AdV. Inorg. Chem. 1988, 32, 55. (122) Loring, J. S.; Yu, P.; Phillips, B. L.; Casey, W. H. Geochim. Cosmochim. Acta 2004, 68, 2791. (123) Helm, L.; Merbach, A. E. Chem. ReV. 2005, 105, 1923. (124) Hugi-Cleary, D.; Helm, L.; Merbach, A. E. HelV. Chim. Acta 1985, 68, 545. (125) Hugi-Cleary, D.; Helm, L.; Merbach, A. E. J. Am. Chem. Soc. 1987, 109, 4444. (126) Swaddle, T. W.; Merbach, A. E. Inorg. Chem. 1981, 20, 4212. (127) Grant, M.; Jordan, R. B. Inorg. Chem. 1981, 20, 55. (128) Richens, D. T. The Chemistry of Aqua Ions; John-Wiley: New York, 1997. (129) Rustad, J. R.; Loring, J. S.; Casey, W. H. Geochim. Cosmochim. Acta 2004, 68, 3011. (130) Forde, S.; Hynes, M. J. New J. Chem. 2002, 26, 1029. (131) Yu, P.; Lee, A. P.; Phillips, B. L.; Casey, W. H. Geochim. Cosmochim. Acta 2003, 67, 1065. (132) Margerum, D. W.; Cayley, G. R.; Weatherburn, D. C.; Pagenkopf. G. K. In Coordination Chemistry; Martell, A., Ed.; ACS Monograph 174, American Chemical Society: Washington, DC, 1978; Vol. 2, p 1. (133) Plankey, B. J.; Patterson, H. H.; Cronan, C. S. EnViron. Sci. Technol. 1986, 20, 160. (134) Plankey, B. J.; Patterson, H. H. Inorg. Chem. 1989, 28, 4331. (135) Swaddle, T. W. Inorg. Chem. 1980, 19, 3203. (136) Swaddle, T. S.; Rosenqvist, J. R.; Yu, P.; Bylaska, E.; Phillips, B. L.; Casey, W. H. Science 2005, 308, 1450. (137) Allouche, L.; Taulelle, F. Chem. Commun. 2003, 2003, 2084. (138) Phillips, B. L.; Lee, A. P.; Casey, W. H. In Proceedings of the 11th Annual Water-Rock Interaction Conference; Wanty, R. B., Seal, R. S., Eds.; A. A. Balkema Publishers: Leiden, The Netherlands, 2004; p 629. (139) Nordin, J. P.; Sullivan, D. J.; Phillips, B. L.; Casey, W. H. Geochim. Cosmochim. Acta 1999, 63, 3513. (140) Furrer, G.; Gfeller, M.; Wehrli, B. Geochim. Cosmochim. Acta 1999, 63, 3069. (141) Amirbahman, A.; Gfeller, M.; Furrer, G. Geochim. Cosmochim. Acta 2000, 64, 911. (142) Wehrli, B.; Wieland, E.; Furrer, G. Aquat. Sci. 1990, 52, 3. (143) Lee, A. P.; Furrer, G.; Casey, W. H. J. Colloid Interface Sci. 2002, 250, 269. (144) Furrer, G.; Ludwig, Chr.; Schindler, P. W. J. Colloid Interface Sci. 1992, 149, 56. (145) Bi, S.; Wang, C.; Qing, C.; Zhang, C. Coord. Chem. ReV. 2004, 248, 441. (146) Bertsch, P. M. Soil Sci. Soc. Am. J. 1987, 51, 825. (147) Akitt, J. W.; Farthing, A. J. Chem. Soc., Dalton Trans. 1981, 1981, 1617. (148) O ¨ hman, L.-O.; Forsling, W. Acta Chem. Scand. 1981, A35, 795. (149) Hedlund, T.; Sjo¨berg, S.; O ¨ hman, L.-O. Acta Chem. Scand. 1987, A41, 197. (150) Teagarden, D. L.; Kozlowski, J. F.; White, J. L.; Hem, S. L. Aluminum chlorohydrate 1: structure studies. Pharm. Sci. 1981, 70, 758-761. (151) Teagarden, D. L.; Radavich, J. F.; White, J. L.; Hem. S. L. Pharm. Sci. 1981, 70, 762. (152) Wang, W.-Z.; Hsu, P. H. Clays Clay Miner. 1994, 42, 356. (153) Pophristic, V.; Klein, M. L.; Holerca, M. N. J. Phys. Chem. A 2004, 108, 113. (154) Flarend, R. E.; Bin, T.; Elmore, D.; Hem. S. L. Food Chem. Toxicol. 2001, 39, 163. (155) Flarend, R. E.; Hem, S. L.; White, J. L.; Elmore, D.; Suckow, M. A.; Rudy, A. C.; Dandashli, E. A. Vaccine 1997, 15, 1314. (156) Guillard, O.; Fauconneau, B.; Olichon, D.; Dedieu, G.; Deloncle, R. Am. J. Med. 2004, 117, 956. (157) Exley, C. Am. J. Med. 2004, 117, 969. (158) Rao, G. V.; Rao, K. S. FEBS 1992, 331, 49. (159) Yokel, R. A.; McNamara, P. J. Pharmacol. Toxicol. 2001, 88, 159. (160) Harris, W. R.; Berthon, G.; Day, J. P.; Exley, C.; Flaten, T. P.; Forbes, W. F.; Kiss, T.; Orvig, C.; Zatta, P. F. J. Toxicol. EnViron. Health 1996, 48, 543. (161) Buffle, J.; Parthasarathy, N.; Haerdi, W. Water Res. 1985, 19, 7. (162) Annadurai, G.; Sung, S. S.; Lee, D.-J. AdV. EnViron. Res. 2004, 8, 713. (163) Snodgrass, W. J.; Clark, M. M.; O’Melia, C. R. Water Res. 1984, 18, 479. (164) O’Melia, C. R. Water Sci. Technol. 1998, 37, 129. (165) Matsui, Y.; Matsushita T.; Sakuma S.; Gojo T.; Mamiya T.; Suzuoki H.; Inoue T. EnViron. Sci. Technol. 2003, 37, 5175. (166) Boisvert, J.-P.; Jolicoeur, C. Colloids Surf., A 1999, 155, 161. (167) Kloprogge, J. T.; Ruan, H.; Frost, R. L. Spectrochim. Acta 2000, A56, 2405. (168) Gao, B. Y.; Hahn, H. H.; Hoffmann, E. Water Res. 2002, 36, 3573.

16 Chemical Reviews, 2006, Vol. 106, No. 1 (169) Schoonheydt, R. A.; Jacobs, K. Y. In Natural Microporous Materials in EnVironmental Technology; Misaelides, P., Maca´×f0ek, F., Pinnavaia, T. J., Colella, C., Eds.; NATO Science Series E, Vol. 362; Kluwer Academic Publishers: Boston, 1999; p 19. (170) Schoonheydt, R. A.; Pinnavaia, T., Lagaly; Gangas, N. Pure Appl. Chem. 1999, 71, 2367. (171) Schoonheydt, R. A.; Jacobs, K. Y. Stud. Surf. Sci. Catal. 2001, 137, 299. (172) Pinnavaia, T. J. Science 1983, 220, 365. (173) Pinnavaia, T. J.; Tzou, M.-S.; Landau, S. D.; Raythatha, R. H. J. Mol. Catal. 1984, 27, 195. (174) Wang, L.; Ebina, Y.; Takada, K.; Kurashima, K.; Sasaki, T. AdV. Mater. 2004, 16, 1412. (175) Wang, L.; Sakai, N.; Ebina, Y.; Takada, K.; Sasaki, T. Chem. Mater. 2005, 17, 1352. (176) Lalik, E.; Kolodziejski, W.; Lerf, A.; Klinowski, J. J. Phys. Chem. 1993, 87, 223. (177) Occelli, M. L.; Bertrand, J. A.; Gould, S. A. C.; Dominguez, J. M. Microporous Mesoporous Mater. 2000, 34, 195. (178) Jones, D. J.; Leloup, J.-M.; Ding, Y.; Rozie`re, J. Solid State Ionics 1993, 1993, 117. (179) Bergaoui, L.; Mrad, I.; Lambert, J.-F.; Ghorbel, A. J. Phys. Chem. B 1999, 103, 2897. (180) Ge, Z.; Li, D.; Pinnavaia, T. J. Microporous Mater. 1994, 3, 165. (181) Yamada, H.; Azuma, N.; Kevan, L. J. Phys. Chem. 1995, 99, 2110. (182) Butruille, J.-R.; Michot, L. J.; Barre`s, O.; Pinnavaia, T. J. J. Catal. 1993, 139, 664. (183) Thomas, S. M.; Betrand, J. A.; Occelli, M. L.; Huggins, F.; Gould, S. A. C. Inorg. Chem. 1999, 38, 2098. (184) Thomas, S. M.; Betrand, J. A.; Occelli, M. A. Chem. Mater. 1999, 11, 184. (185) d’Espinose de la Caillerie, J.-B.; Man, P. P.; Vicente, M. A.; Lambert, J.-F. J. Phys. Chem. B 2002, 106, 4133. (186) Benito, I.; del Riego, A.; Martı´nez, M.; Blanco, C.; Pesquera, C.; Gonza´lez, C. Appl. Catal. 1999, 180, 175. (187) Montarge`s, E.; Michot, L. J.; Ildefonse, P. Microporous Mesoporous Mater. 1999, 28, 83. (188) Clearfield, A. AdV. Catal. Nanostruct. Mater. 1996, 1996, 345. (189) Cheng, S. Catal. Today 1999, 49, 303. (190) Schoonheydt, R. A.; Leeman, H.; Scorpion, A.; Lenotte, I.; Grobet, P. Clays Clay Miner. 1994, 42, 518. (191) Shann, J. R.; Bertsch, P. M. Soil Sci. Soc. Am. J. 1993, 57, 116. (192) Masion, A.; Bertsch, P. M. Plant, Cell EnViron. 2004, 20, 504.

Casey (193) Parker, D. R.; Kinraide, T. B.; Zelasny, L. W. Soil Sci. Soc. Am. J. 1988, 52, 438. (194) Parker, D. R.; Kinraide, T. B.; Zelasny, L. W. Soil Sci. Soc. Am. J. 1989, 53, 789. (195) Flaten, T. P.; Garruto, R. M. J. Theor. Biol. 1992, 156, 417. (196) Exley, C.; Chappel, J. S.; Birchall, J. D. J. Theor. Biol. 1991, 151, 417. (197) Exley, C.; Wicks, A. J.; Hubert, R. B.; Birchall, D. J. J. Theor. Biol. 1994, 167, 415. (198) Exley, C.; Wicks, A. J.; Hubert, R. B.; Birchall, D. J. J. Theor. Biol. 1996, 179, 25. (199) Lothenbach, B.; Furrer, G.; Schulin, R. EnViron. Sci. Technol. 1997, 31, 1452. (200) Furrer, G.; Phillips, B. L.; Ulrich, K.-U.; Po¨thig, R.; Casey, W. H. Science 2002, 297, 2245. (201) Dubbin, W. E.; Sposito, G. EnViron. Sci. Technol. 2005, 39, 2509. (202) Furrer, G.; Trusch, B.; Mueller, Chr. Geochim. Cosmochim. Acta 1992, 56, 3831. (203) Hiradate, S.; Yamaguchi, N. U. J. Inorg. Biochem. 2003, 97, 26. (204) Masion, A.; Thomas, F.; Tchoubar, D.; Bottero, J. Y.; Tekely, P. Langmuir 1994, 10, 4353. (205) Krishnamurti, G. S. R.; Wang, M. K.; Huang, P. M. Clays Clay Miner. 2004, 52, 734. (206) Zhang, P.; Hahn, H. H.; Hoffmann, E.; Zeng, G. Chemosphere 2004, 57, 1489. (207) Kerven, G. L.; Larsen, P. L.; Blamey, F. P. C. Soil Sci. Soc. Am. J. 1995, 59, 765. (208) Masion, A.; Thomas, F.; Tchoubar, D.; Bottero, J. Y.; Tekely, P. J. Non-Cryst. Solids 1994, 171, 191. (209) Masion, A.; Vilge-Ritter, A.; Rose, J.; Stone, W. E. E.; Teppen, B. J.; Tybacki, D.; Bottero, J.-Y. EnViron. Sci. Technol. 2000, 34, 3242. (210) Yamaguchi, N.; Hiradate, S.; Mizoguchi, M.; Miyazaki, T. Soil Sci. Soc. Am. J. 2004, 68, 1838. (211) Gerard, F.; Boudot, J.-P.; Ranger, J. Appl. Geochem. 2001, 16, 513. (212) Hunter, D.; Ross, D. S. Science 1991, 251, 1056. (213) Casey, W. H.; Swaddle, T. W. ReV. Geophys. 2003, 41, 1. (214) Baker, L. C. W.; Figgis, J. S. J. Am. Chem. Soc. 1970, 92, 3794. (215) Chen, Y.-G.; Qu, L.-Y.; Peng, J.; Yu, M.; Lin, Y.-H.; Yu, Z.-L. Jiegon Huaxue 1993, 12, 338. (216) Bodor, A.; To´th, T.; Ba´nyai, I.; Szabo´, Z.; Hefter, G. T. Inorg. Chem. 2000, 39, 2530.

CR040095D