Copper−Glyphosate Sorption to Microcrystalline Gibbsite in the

Environ. Sci. Technol. 2005, 39, 2509-2514

Copper-Glyphosate Sorption to Microcrystalline Gibbsite in the Presence of Soluble Keggin Al13 Polymers W I L L I A M E . D U B B I N * ,† A N D GARRISON SPOSITO‡ Department of Mineralogy, Natural History Museum, London, SW7 5BD, U.K., and Division of Ecosystem Sciences, University of California, Berkeley, California 94720-3110

Among the most reactive yet largely neglected adsorbents of toxicant species occurring in acidic aquatic environments are the -Keggin Al13 polyoxocations [AlO4Al12(OH)24(H2O)127+], known generally as Al13 polymers. Here, we report on the sorption of Cu(II), a common ingredient of pesticides, and glyphosate {N-[phosphonomethyl]glycine (PMG)}, a widely applied herbicide, to microcrystalline gibbsite [γ-Al(OH)3] in the presence of soluble Al13 polymers over the pH range 4-7. In the presence of gibbsite and soluble Al13 polymers, dissolved Cu(II) decreased gradually with pH, achieving a minimum at pH 5.5. Between pH 5.5 and 6.0, however, soluble Cu increased markedly, with approximately 80% of the added metal remaining in solution at pH 5.86. At pH > 6.0, soluble Cu once again decreased, becoming undetectable at pH 7. The anomalous Cu solubilization was attributed to a concomitant deprotonation of soluble Al13 polymers, yielding surface OH groups possessing high affinity for Cu(II). Removal of Cu from solution at pH > 6.0 is facilitated by flocculation of the Al13 polymers to which Cu had sorbed. The sorption behavior of the zwitterionic PMG in the presence of gibbsite and Al13 polymers was consistent with this interpretation, there being a dramatic increase in sorbed PMG at pH > 6.0 as the Al13 polymers deprotonated and flocculated. Copper and PMG loss from solution with increasing pH when both adsorptives were added to the gibbsite-Al13 polymer system was broadly similar to what was observed in the PMG-free systems, although small differences were detected in response to varying the order of adsorptive addition. The inclusion of soluble Al polymers in our experiments exposes a fundamental limitation of models based on but a single inorganic adsorbent as a means to predict the behavior of trace metals and xenobiotic organic compounds in natural systems.

Introduction Sorption of toxicant metals and ligands by colloids is perhaps the most important reaction affecting their transport and bioavailability in soils, sediments, and aquatic environments (1). Among the toxic metals of greatest interest, Cu figures prominently due to its ubiquity as a constituent of fungicides, * Corresponding author phone: +44 (0)207 942 5616; fax: +44 (0)207 942 5537; e-mail: [email protected]. † Natural History Museum. ‡ University of California. 10.1021/es048199t CCC: $30.25 Published on Web 02/23/2005

 2005 American Chemical Society

bactericides, and algicides (2). Similarly, the organophosphonate ligand, glyphosate (N-[phosphonomethyl]glycine), a widely used herbicide (3) which chemisorbs to metal oxide surfaces via its phosphonate moiety (4-6) and complexes Cu at the amine group, remains of significant environmental concern, abetted by recent evidence that its residues may undergo facile biodegradation according to a pathway in which chemisorption to a metal oxide phase is the first step (4). Experimental protocols incorporating a single, wellcharacterized inorganic adsorbent have for decades been the accepted means through which to derive toxicant sorption models valid over a range of predetermined physicochemical conditions (1). Despite considerable progress deducing sorption reactions in these monosorbent model systems, extrapolation of sorption behavior to multicomponent natural systems remains a continuing challenge. As a consequence, the model systems investigated have become ever more complex by incorporating, for example, sorptive ligands such as humic substances in an attempt to mimic more closely the complexity found in natural environments (7, 8). Among the most reactive inorganic sorbents occurring in both field (9-11) and laboratory (12-14) settings, yet largely neglected as components of model sorbent systems, are the -Keggin Al polyoxocations [AlO4Al12(OH)24(H2O)127+ “tridecameric Al” or simply “Al13 polymers”] (15-17). The Al13 polymer consists of a central, tetrahedrally coordinated Al surrounded by four groups of three AlO6 octahedra (Figure 1). At pH 5, the Al13 polymer has a +7 valence, which can be modified through deprotonation reactions that begin near pH 6 and go to completion over a narrow pH range, with protons believed to be released in pairs, yielding the discrete species Al135+, Al133+, and Al13+ without distinct plateaus in the titration curve (18). The precise pH range over which Al13 deprotonates varies principally with the concentration of Al13 (19, 20), the polymer being considerably more acidic at higher concentrations (e.g., 1 mM) because the aggregation induced at these concentrations promotes H+ release. The titration curve of Al13 is remarkably steep, with 3 protons per polymer lost over the small interval of 0.2 pH units in dilute solutions and over 1.5 pH units at higher concentrations (18, 21). As a consequence of this deprotonation, the Al13 polymer exposes surface OH groups having a high affinity for Cu and other metals (2224). In light of this affinity and its rapid flocculation at pH > 6.5 (18, 19, 25), the Al13 polymer has evident potential to influence Cu(II) solubility, and the same may be true for zwitterions such as glyphosate. The primary objective of the present study was therefore to examine the sorption behavior of Cu(II) and glyphosate in a model system comprised of Al13 polycations and microcrystalline gibbsite [γ-Al(OH)3], a prototypical metal oxide adsorbent. Our experiments included studies of each of the two toxicants alone with the two adsorbents as well as combined together, including an examination of the effect of their order of addition.

Experimental Section Synthesis and Characterization of the Keggin Polymers. Hydroxy-Al polymers were synthesized by titrating a freshly prepared 0.1 M AlCl3 (GFS, ACS grade) solution with 0.1 M NaOH (GFS, ACS grade) at 1 mL min-1 until a OH/Al molar ratio (rOH/Al) of 3.0 was reached (26). Excess Cl- was not removed prior to aging to restrict crystal growth and thus maximize surface area. After aging for 10 days, the solid phase was isolated by centrifugation (20 min at 7000 rpm), washed VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2509

FIGURE 1. Polyhedral depiction of the AlO4Al12(OH)24(H2O)127+(aq) complex showing the single AlO4 tetrahedron surrounded by four groups of three AlO6 octahedra. Two distinct sets of 12 shared OH are indicated by µ2-OH and µ2-OH′. The terminal ligand of each of the 12 octahedra is a water molecule (η-OH2).

FIGURE 2. Structural representation of glyphosate (N-[phosphonomethyl]glycine). The acid dissociation constants are pKa1 ) 0.80 (first phosphonic); pKa2 ) 2.22 (carboxylate); pKa3 ) 5.44 (second phosphonic); and pKa4 ) 10.13 (amine) (6). three times with high-purity 18 MΩ cm-1 water (Milli-Q Plus, Millipore), then identified as γ-Al(OH)3 by X-ray diffraction. Nitrogen multipoint BET analysis (27) indicated a specific surface area of 94 m2 g-1. Mean diameter of the suspended polynuclear Al colloids was estimated by dynamic light scattering using a 35 mW He-Ne laser (λο ) 632.8 nm) (model 127, Spectra-Physics Inc.), with the beam polarized vertically with respect to the plane of the goniometer (BI-200SM, Brookhaven Instruments Corp.). Following calibration with polystyrene spheres (92.0 ( 3.7 nm; 560 ( 6 nm, Duke Scientific), supernatant subsamples, suitably diluted, were transferred to the Burchard cells and then placed in the sample holder. The scattered light was collected by a photomultiplier tube placed in the horizontal plane and positioned at an angle θ relative to the incident beam. Data were processed with dedicated particlesizing software reporting Z-average diameters (1). Supernatant solution Al was quantified by ICP-AES. Aqueous hydroxy-Al species were identified using the ferron method (28-30), a kinetically based spectrophotometric technique which distinguishes soluble Al fractions according to their reaction rate with a complexing agent, ferron (C9H6INO4S, [7-iodo-8-hydroxyquinoline-5-sulfonic acid]; Sigma). Briefly, volumes of prepared ferron reagent and stock suspension supernatant, diluted to 0.1 mM with 18 MΩ cm-1 water, were chosen to give a ferrontotal/Altotal molar ratio ) 50. Immediately after combining the ferron reagent and sample, absorbance increases at 363 nm were measured at 10 s intervals for 60 min. Confirmation of Al13 presence can be achieved following derivation of the rate coefficient from a first-order plot of log [Alunreacted] versus time as described previously (28). Temperature was maintained at 25.0 ( 0.5 °C throughout the reaction period. Adsorption Experiments. Three milliliter aliquots of hydroxy-Al stock suspension (25 mM Al) were transferred to 50 mL Nalgene PPCO centrifuge tubes. To each centrifuge tube was added, dropwise while stirring, 3 mL of 10-4 M CuCl2 (GFS, ACS Grade) or 10-4 M glyphosate (N-[phosphonomethyl]glycine, “PMG”; Sigma), which occurs as a zwitterionic species (6) over the pH range studied [pKa1 ) 0.80 (first phosphonic); pKa2 ) 2.22 (carboxylate); pKa3 ) 5.44 (second phosphonic); pKa4 ) 10.13 (amine)] (Figure 2). An additional 3 mL of high-purity 18 MΩ cm-1 water was added 2510

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

FIGURE 3. Supernatant [Al] as a function of pH and centrifugation speed at 25 °C and I ) 0.025. At 7000 rpm, supernatant [Al] remains largely constant until about pH 6 at which point [Al] markedly decreases until it is no longer detected at pH 7. At 15 000 rpm, colloidal Al is removed from the suspension gradually until pH > 6, where the remaining Al settles. to each tube, bringing all solutions to 9 mL. A series of identical suspensions was prepared in this way, adjusting the pH of each with the dropwise addition of either 0.1 M HCl (GFS, ACS Grade) or 0.1 M NaOH (GFS, ACS Grade) to produce a set of suspensions with a pH range from 4 to 7. All suspensions were then agitated gently by mixing endover-end for 24 h. Following the 24 h reaction, pH was measured and the supernatant solution was separated from the solid phase by centrifuging for 20 min at 7000 rpm. Where both PMG and Cu were added to the hydroxy-Al suspension, variable amounts of 10-4 M PMG (either 3, 1.5, or 1 mL) were chosen to give one of three Cu/PMG molar ratios (i.e., 1, 2, or 3). All suspensions contained 3 mL of 10-4 M CuCl2. High-purity 18 MΩ cm-1 water was added where necessary to bring the total volume of all suspensions to 9 mL. Following the addition of the first reagent, either Cu or PMG, the hydroxy-Al suspensions were agitated as described above for 24 h before addition of the second reagent. The second reagent was then introduced dropwise with stirring, followed by readjustment of pH if required. The suspensions were once again agitated for a further 24 h, measured for pH, and then centrifuged for 20 min at 7000 rpm to obtain the supernatant solution. Copper in supernatant solution was quantitated by ICP-AES. Supernatant solution glyphosate was first oxidized with 30% H2O2 (GFS, ACS Grade) to orthophosphate, which then could be quantitated spectrophotometrically as the heteropoly blue complex at 830 nm (31).

Results and Discussion Characterization of Supernatant Solution Al. Total [Al] in each 50 mL centrifuge tube was 8.3 mM (i.e., one-third the stock solution concentration). Most of this Al was present as a white precipitate, identified as microcrystalline γ-Al(OH)3(S), which rapidly settled. Following centrifugation at 7000 rpm for 20 min, approximately 0.65 mM Al remained in solution over the pH range 4-6 (Figure 3), yet the supernatant solution was visibly clear. With longer and faster centrifugation (30 min at 15 000 rpm), the remaining soluble Al decreased gradually as pH was increased from 4 to 6, evidently because of polymer growth due to flocculation arising primarily from a reduction in Al137+ charge. Soluble Al decreased rapidly with pH for all supernatant solutions at pH > 6, becoming negligible near pH 7. Speciation of Supernatant Al. Reaction of stock suspension supernatant solution with ferron gave absorbance

FIGURE 4. Plot of absorbance at 363 nm versus time, representing the ferron-Al reaction kinetics at 25 ( 0.5 °C for: (a) an acidified (pH ≈ 2.5), mononuclear Al standard ([Al]tot ) 0.1 mM); and (b) the hydroxy-Al stock suspension supernatant ([Al]tot diluted to 0.1 mM). increases at 363 nm from 0 to 20 min (Figure 4). Similar absorbance increases were observed for supernatant solutions obtained following centrifugation of stock suspension at both 7000 and 15 000 rpm. Rate coefficients derived from first-order plots of log [Alunreacted] versus time indicate the presence of reactive soluble Al fractions (28). Based on the reaction kinetics of an acidified (pH ≈ 2.5) Al standard, the reaction of mononuclear Al with ferron is complete within about 3 min (Figure 4, curve a). Therefore, for the stock suspension supernatant solution (Figure 4, curve b), absorbance increases at reaction times >3 min can be attributed solely to polynuclear species (i.e., Al13), whereas absorbance increases at reaction times 2.3, the Al13 units form linear aggregates, with Cl- serving as a bridging ion (19, 25). With increasing [Al]tot, the Al13 units aggregate further by forming either additional Cl- bridged outer-sphere associations or inner-sphere condensation products. Rates of Al13 decomposition have been determined as a function of ionic strength, temperature, and pH (20, 44). Decomposition rate is positively correlated with H+, with the Al13 half-life at pH 5 and 25 °C being approximately 24 d. Given the synthesis conditions used in the present study [i.e., 0.025 M Al; rOH/Al ) 3.0; I ) 0.075], the 10 d aging period was not sufficient for complete Al13 transformation to γ-Al(OH)3, and, therefore, significant amounts of Al must have remained in solution as the tridecamer species.

FIGURE 5. Mean diameter (Z-average) of Al13 clusters as a function of pH at 25 °C and I ) 0.025. Cluster mean diameter increased steadily from pH 4.35 to pH 6.29, with a dramatic increase at pH > 6.29. If all supernatant solution Al (i.e., 0.65 mM) were present as Al13, its concentration would be approximately 0.05 mM. However, it is apparent from the ferron analyses that the supernatant solution contained a highly reactive Al fraction (i.e., ferron reaction < 3 min), comprised of mononuclear Al species (15). A more realistic estimate therefore assumes Al13 comprised about two-thirds of the total supernatant solution Al, yielding [Al13] ≈ 0.03 mM. As this latter concentration falls within the range examined by Furrer et al. (18), the acidbase behavior of Al13 in the current study is assumed to be intermediate to that in the Al13 solutions described by them. Colloid Size. Mean colloid diameter increased gradually from 570 nm at pH 4.35 to 693 nm at pH 5.92 (Figure 5). The size of the Al13 polycation as determined by small-angle X-ray scattering is approximately 1.2 nm (32), and, therefore, the mean diameters measured by light scattering in the present study must represent clusters of Cl--bridged Al13 monomers, such as those proposed previously (19, 25). These earlier studies also reported that individual Al13 polymers flocculate to form large clusters whose structure is fractal, with dimension increasing from 1.4 to 1.8 as pH increases, implying a denser cluster (1). There was a marked increase in mean colloid diameter at pH > 6, consistent with the expected polymer growth following Al13 deprotonation. Fully protonated aqueous Al137+ molecules at pH 5 remain in solution due to electrostatic repulsion. In freshly neutralized AlCl3 solutions, the Al13 polycations are bridged loosely by Cl- to yield branching aggregates (19, 25, 32). With aging, or an increase in pH above 6, the Cl- -linked Al13 molecules polymerize more extensively as Cl- is displaced and octahedral hydroxyls from adjacent Al13 units coalesce to form oxo bridges. Continued polymer growth leads to precipitation of the tridecamers (19, 20). Sorption of Cu. Copper sorption by hydrous Al oxides is highly pH dependent, with plots of Cu adsorbed versus pH (adsorption edges) typically yielding sigmoidal curves (1, 45). Over the pH range 4-5.5, and in the absence of PMG, Cu(II) adsorption to γ-Al(OH)3 is similar to that reported previously (45, 46), with approximately 40% of [Cu]tot adsorbed at pH 5.5 (Figure 6a). However, between pH 5.5 and 6, the sorptive behavior of Cu departed sharply from that observed in previous studies: soluble Cu increased, with nearly 80% of [Cu]tot remaining in solution at pH 5.86. At pH > 6, soluble Cu once again decreased, with no soluble Cu detected at pH 7. Copper loss from solution in the presence of PMG (Figure 6b-d) followed the same general trends as that for the PMGfree system (Figure 6a). That is, in all cases, there was a marked VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2511

FIGURE 6. Cu removal from solution (%) as a function of pH in four hydroxy-Al systems, showing the influence of reagent addition sequence and variable Cu/PMG molar ratio at 25 °C and I ) 0.025. increase in soluble Cu at pH > 5.5, followed by Cu removal from solution within a narrow pH range. Where many data points cluster near pH 6 (e.g., Figure 6d), the “solubilization cliff” is found to be remarkably steep, which is expected, as it mirrors the deprotonation curves for Al13 and -Keggin molecules generally (18, 21). When PMG and Cu were combined, then later introduced as a Cu-PMG solution to the OH-Al suspension (Figure 6d), more Cu remained in solution at pH 5 (about 80%), than when PMG and Cu were added separately (60-70%) (Figure 6b,c). The greater soluble Cu observed at pH 5 when Cu and PMG were added together is unsurprising, given the high stability of the Cu(II)-glyphosate complex (47). Enhanced Cu sorption at pH < 4.5 is evident where PMG was the first adsorptive added to the OH-Al suspension (Figure 6b,c). The phosphonate moiety of PMG sorbs strongly to γ-Al(OH)3 at low pH, thus leaving the amine N and carboxyl O free to complex Cu2+ (4). The effect of Cu/PMG molar ratio on soluble Cu was not readily apparent and consistent trends could not be observed, although it appears that slightly more Cu moves into solution at pH 6 when Cu/PMG ) 3 (Figure 6c,d). The increase in soluble Cu beginning near pH 5.5 coincides with the deprotonation of Al13 and the concomitant increase in Al13 affinity for Cu(II). Given that [Cu] ) 0.033 mM in each centrifuge tube, and that supernatant [Al13] ≈ 0.03 mM, [Cu]/ [Al13] ≈ 1. There was, therefore, sufficient Al13 to complex all Cu present, particularly as each Al13 possesses 12 surface H2O, any one of which can deprotonate to form a reactive site for Cu complexation. Unlike most polyprotic acids, which show increasing pKa values with increasing deprotonation, the 12 terminal H2O ligands of Al13 (η-OH2, Figure 1) deprotonate largely independently. The model that best fits experimental deprotonation data describes Al137+ proton loss occurring via six two-proton steps, which, although discrete, are largely indistinguishable (18). Autonomy of the 12 terminal η-OH2 sites was examined further in a 17O NMR study, which revealed 2512

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

exchange rate coefficients at 298 K for the 12 η-OH2 sites of kex298 ) 1100 ((300) s-1, and also confirmed that H2O exchange is essentially independent of -Keggin structure and composition (17, 48). A question arises as to what factors control the distribution of Cu between solid-phase γ-Al(OH)3 and soluble Al13. The 20% Cu which remains on the solid phase at pH 5.86 after 24 h reaction (Figure 6a) may be the result of slow kinetics of Cu desorption from gibbsite, or perhaps this residual 20 % Cu reflects a thermodynamically stable partitioning of Cu resulting from the relative stabilities of the Cu-Al13 and Cugibbsite adsorption complexes. The loss of Cu from solution between pH 6 and 7 mimics the sigmoidal adsorption edges observed for other Cugibbsite systems (45, 46), except that Cu removal in the present case was complete within about 0.5 pH unit. In the current study, Cu lost from solution cannot strictly be described as an adsorption edge because a significant component of this removal was from precipitation of the Al13 polymers to which Cu had sorbed. The onset of Al13 precipitation is shifted to higher pH values in the presence of PMG (compare the initiation of precipitation in Figure 6a with that in Figure 6c and d). This observation is consistent with the long-established idea that strongly complexing ligands, such as PMG, increase the pH of maximum precipitation of Al hydroxides (49). Sorption of PMG. As with Cu, glyphosate in the gibbsiteAl13 polymer system will be distributed among three fractions: (i) adsorbed to gibbsite, (ii) sorbed by soluble Al13, or (iii) present as a soluble species unassociated with OH-Al polymers. Fractions (ii) and (iii) were indistinguishable using the method of PMG quantitation employed in our study. Consequently, it was not possible to determine the coordination environment of soluble PMG; only the coordination environment of PMG in the solid phase is known [i.e., sorbed via the phosphonate moiety (4)]. In the system comprised of only PMG and OH-Al sorbents, PMG desorbed from gibbsite as pH increased from

FIGURE 7. PMG removal from solution (%) as a function of pH in four hydroxy-Al systems, showing the influence of reagent addition sequence and variable Cu/PMG molar ratio at 25 °C and I ) 0.025. 4 to 6 (Figure 7a). This trend is consistent with the pH-induced desorption of organic ligands from hydrous oxides known generally (1, 50). Near pH 6, there was an abrupt increase in soluble PMG, most evident in Figure 7d, that deviates from the expected gradual desorption of PMG from gibbsite with increasing pH. This increase in soluble PMG coincides with the pH range for Al13 deprotonation and the consequent increased solubility of Cu that we observed. At pH > 6, solution PMG decreased and became undetectable by pH 7, also as observed for Cu. Partitioning of PMG between solid and solution was largely unaffected by Cu (compare Figure 7a and b-d). However, the partitioning of PMG is subtly influenced by the sequence of adsorptive addition. When PMG was added before Cu, about 60-75% of total PMG was bound to the solid phase at pH < 5.5 (Figure 7c). This compares with 40-60% of total PMG bound to gibbsite when Cu was added first (Figure 7b), or when PMG and Cu were added together (Figure 7d). Because PMG is an effective chelator of Cu, the potential for Cu removal from solution at pH < 5.5 is therefore enhanced when PMG is the first to be added. This PMG-facilitated removal of Cu from solution at low pH is evident in Figure 6c. The complete removal of PMG from solution over the pH range 6-7 is superficially similar to the removal of Cu, both coinciding with the flocculation and precipitation of Al13 polymers. The coincidence of PMG removal and Al13 precipitation suggests that a major portion of the soluble PMG was in fact complexed with Al13 prior to the onset of precipitation (i.e., at pH values < 6). The properties and identity of the coprecipitated PMG-Al13 polymer were not detemined, although the initial short-range ordered materials would likely nucleate on the gibbsite surface, thereby facilitating Cu sorption to the solid. The incorporation of Al13 tridecamers into our model CuPMG-gibbsite systems has induced Cu sorption behavior, which, to our knowledge, has never been reported previously. The anomalous and abrupt solubilization of Cu near pH 6,

coincident with Al13 deprotonation, has considerable environmental relevance, as it informs model predictions of Cu transport and bioavailability. More importantly, the inclusion of soluble Al polymers in our model systems exposes the potential limitations of models based on but a single inorganic adsorbent as a means to mimic the complex behavior of trace metals and xenobiotic organic compounds in multicomponent natural systems.

Acknowledgments W.E.D. thanks the Natural Sciences and Engineering Research Council of Canada for financial support in the form of a postdoctoral fellowship. We thank Prof. William Casey, University of California, for providing Figure 1.

Literature Cited (1) Sposito, G. The Surface Chemistry of Natural Particles; Oxford University Press: Oxford, 2004. (2) Kabata-Pendias, A. Soil-plant transfer of trace elements-an environmental issue. Geoderma 2004, 122, 143-149. (3) Castle, L. A.; Siehl, D. L.; Gorton, R.; Patten, P. A.; Chen, Y. H.; Bertain, S.; Cho, H.-J.; Duck, N.; Wong, J.; Liu, D.; Lassner, M. W. Discovery and directed evolution of a glyphosate tolerance gene. Science 2004, 304, 1151-1154. (4) Dubbin, W. E.; Sposito, G.; Zavarin, M. X-ray absorption spectroscopic study of Cu-glyphosate adsorbed by microcrystalline gibbsite. Soil Sci. 2000, 165, 699-707. (5) Sheals, J.; Sjoberg, S.; Persson, P. Adsorption of glyphosate on goethite: Molecular characterization of surface complexes. Environ. Sci. Technol. 2002, 36, 3090-3095. (6) Sheals, J.; Granstrom, M.; Sjoberg, S.; Persson, P. Coadsorption of Cu(II) and glyphosate at the water-goethite (R-FeOOH) interface: molecular structures from FTIR and EXAFS measurements. J. Colloid Interface Sci. 2003, 262, 38-47. (7) Alcacio, T. E.; Hesterberg, D.; Chou, J. W.; Martin, J. D.; Beauchemin, S.; Sayers, D. E. Molecular scale characteristics of Cu(II) bonding in goethite-humate complexes. Geochim. Cosmochim. Acta 2001, 65, 1355-1366. (8) Christl, I.; Kretzschmar, R. Interaction of copper and fulvic acid at the hematite-water interface. Geochim. Cosmochim. Acta 2001, 65, 3435-3442. VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2513

(9) Hunter, D.; Ross, D. S. Evidence for a phytotoxic hydroxyaluminum polymer in organic soil horizons. Science 1991, 251, 1056-1058. (10) Furrer, G. New aspects on the chemistry of aluminum in soils. Aquat. Sci. 1993, 55, 281-290. (11) 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. (12) Furrer, G.; Trusch, B.; Muller, C. The formation of polynuclear Al13 under simulated natural conditions. Geochim. Cosmochim. Acta 1992, 56, 3831-3838. (13) Parker, D. R.; Bertsch, P. M. Formation of the “Al13” tridecameric polycation under diverse synthesis conditions. Environ. Sci. Technol. 1992, 26, 914-921. (14) Pernitsky, D. J.; Edzwald, J. K. Solubility of polyaluminium coagulants. J. Water Supply: Res. Technol.-AQUA 2003, 52, 395406. (15) Bertsch, P. M.; Parker, D. R. In The Environmental Chemistry of Aluminum; Sposito, G., Ed.; CRC Press: Boca Raton, FL, 1996; pp 117-168. (16) Amirbahman, A.; Gfeller, M.; Furrer, G. Kinetics and mechanisms of ligand-promoted decomposition of the Keggin Al13 polymer. Geochim. Cosmochim. Acta 2000, 64, 911-919. (17) Casey, W. H.; Swaddle, T. W. Why small? The use of small inorganic clusters to understand mineral surface and dissolution reactions in geochemistry. Rev. Geophys. 2003, 41, art. no. 1008. (18) Furrer, G.; Ludwig, C.; Schindler, P. W. On the chemistry of the Keggin Al13 polymer. I. Acid-base properties. J. Colloid Interface Sci. 1992, 149, 56-67. (19) Bottero, J. Y.; Cases, J. M. In Studies in Surface Science and Catalysis, Vol. 99: Adsorption on New and Modified Inorganic Sorbents; Dabrowski, A., Tertykh, V. A., Eds.; Elsevier: London, 1996; pp 319-331. (20) Furrer, G.; Gfeller, M.; Wehrli, B. On the chemistry of the Keggin Al13 polymer: Kinetics of proton-promoted decomposition. Geochim. Cosmochim. Acta 1999, 63, 3069-3076. (21) Lee, A. P.; Furrer, G.; Casey, W. H. On the acid-base chemistry of the Keggin polymers: GaAl12 and GeAl12. J. Coll. Interface Sci. 2002, 250, 269-270. (22) Furrer, G.; Lothenbach, B.; Scharli, H.; Ludwig, C. Interactions between polynuclear aluminium compounds and heavy metal ions. J. Ecol. Chem. 1994, 3, 303-312. (23) Lothenbach, B.; Furrer, G.; Schulin, R. Immobilization of heavy metals by polynuclear aluminum and montmorillonite compounds. Environ. Sci. Technol. 1997, 31, 1452-1462. (24) Badora, A.; Furrer, G.; Grunwald, A.; Schulin, R. Immobilization of zinc and cadmium in polluted soils by polynuclear Al13 and Al-montmorillonite. J. Soil Contam. 1998, 7, 573-588. (25) Bottero, J. Y.; Axelos, M.; Tchoubar, D.; Cases, J. M.; Fripiat, J. J.; Fiessinger, F. Mechanism of formation of aluminum trihydroxide from Keggin Al13 polymers. J. Colloid Interface Sci. 1987, 117, 47-57. (26) Hsu, P. H. In Minerals in Soil Environments; Dixon, J. B., Weed, S. B., Eds.; Soil Science Society of America: Madison, WI, 1989; pp 331-378. (27) Brunauer, S.; Emmett, P. H.; Teller, E. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 1938, 60, 309-319. (28) Parker, D. R.; Bertsch, P. M. Identification and quantification of the “Al13” tridecameric polycation using ferron. Environ. Sci. Technol. 1992, 26, 908-914. (29) Luster, J.; Yang, A.; Sposito, G. On the interpretation of labile aluminum as determined by reaction with 8-hydroxyquinoline. Soil Sci. Soc. Am. J. 1993, 57, 976-980. (30) Bertsch, P. M.; Bloom, P. R. In Methods of Soil Analysis, Part 3 - Chemical Methods; Sparks, D. L., Ed.; American Society of Agronomy: Madison, WI, 1996; pp 517-550. (31) Glass, R. L. Colorimetric determination of glyphosate in water after oxidation to orthophosphate. Anal. Chem. 1981, 53, 921923. (32) Bottero, J. Y.; Tchoubar, D.; Cases, J. M.; Fiessinger, F. Investigation of the hydrolysis of aqueous solutions of aluminum

2514

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

(33) (34)

(35) (36) (37)

(38)

(39)

(40) (41)

(42) (43) (44) (45) (46) (47)

(48)

(49) (50)

chloride. 2. Nature and structure by small-angle X-ray scattering. J. Phys. Chem. 1982, 86, 3667-3673. Bertsch, P. M. Conditions for Al13 polymer formation in partially neutralized aluminum solutions. Soil Sci. Soc. Am. J. 1987, 51, 825-828. Akitt, J. W.; Elders, J. M. Multinuclear magnetic resonance studies of the hydrolysis of aluminum(III). 8. Base hydrolysis monitored at very high magnetic field. J. Chem. Soc., Dalton Trans. 1988, 5, 1347-1355. Ross, D. S.; Bartlett, R. J.; Zhang, H. Photochemically induced formation of the “Al13” tridecameric polycation in the presence of Fe(III) and organic acids. Chemosphere 2001, 44, 827-832. Liu, H. J.; Qu, J. H.; Hu, C. Z.; Zhang, S. J. Characteristics of nanosized polyaluminum chloride coagulant prepared by electrolysis process. Colloids Surf., A 2003, 216, 139-147. Masion, A.; Thomas, F.; Tchoubar, D.; Bottero, J. Y.; Tekely, P. Chemistry and structure of Al(OH)/organic precipitates - A small-angle X-ray scattering study. 3. Depolymerization of the Al13 polycation by organic ligands. Langmuir 1994, 10, 43534356. Molis, E.; Thomas, F.; Bottero, J. Y.; Barres, O.; Masion, A. Chemical and structural transformation of aggregated Al13 polycations, promoted by salicylate ligand. Langmuir 1996, 12, 3195-3200. Boisvert, J. P.; Jolicoeur, C. Influences of sulfate and/or silicate present in partially prehydrolyzed Al(III) flocculants on Al(III) speciation in diluted solutions. Colloids Surf., A 1999, 155, 161170. Kerven, G. L.; Larsen, P. L.; Blamey, F. P. C. Detrimental sulfate effects on formation of Al-13 tridecameric polycation in synthetic soil solutions. Soil Sci. Soc. Am. J. 1995, 59, 765-771. Huang, P. M.; Wang, M. K.; Kampf, N.; Schulze, D. G. In Soil Mineralogy with Environmental Applications; Dixon, J. B., Schulze, D. G., Eds.; American Society of Agronomy: Madison, WI, 2002; pp 261-289. Violante, A.; Huang, P. M. Influence of inorganic and organic ligands on the formation of aluminum hydroxides and oxyhydroxides. Clays Clay Miner. 1985, 33, 181-191. Bradley, S. M.; Kydd, R. A.; Howe, R. F. The structure of Al gels formed through the base hydrolysis of Al3+ aqueous solutions. J. Colloid Interface Sci. 1993, 159, 405-412. Hsu, P. H. Effect of temperature on the degradation of Al13 complex. Clays Clay Miner. 1997, 45, 286-289. McBride, M. B. Cu2+ adsorption characteristics of aluminum hydroxide and oxyhydroxides. Clays Clay Miner. 1982, 30, 2128. McBride, M. B.; Fraser, A. R.; McHardy, W. J. Cu2+ interaction with microcrystalline gibbsite. Evidence for oriented chemisorbed copper ions. Clays Clay Miner. 1984, 32, 12-18. Daniele, P. G.; DeStefano, C.; Prenesti, E.; Sammartano, S. Copper(II) complexes of N-(phosphonomethyl)glycine in aqueous solution: A thermodynamic and spectrophotometric study. Talanta 1997, 45, 425-431. Casey, W. H.; Phillips, B. L.; Karlsson, M.; Nordin, S.; Nordin, J. P.; Sullivan, D. J.; Neugebauer-Crawford, S. Rates and mechanisms of oxygen exchanges between sites in the AlO4Al12(OH)24(H2O)127+ (aq) complex and water: Implications for mineral surface chemistry. Geochim. Cosmochim. Acta 2000, 64, 2951-2964. Marion, S. P.; Thomas, A. W. Effects of diverse anions on the pH of maximum precipitation of aluminum hydroxide. J. Colloid Sci. 1946, 1, 221-228. Schindler, P. W.; Stumm, W. In Aquatic Surface Chemistry; Stumm, W., Ed.; John Wiley and Sons: New York, 1987; pp 83-110.

Received for review November 17, 2004. Revised manuscript received January 17, 2005. Accepted January 18, 2005. ES048199T