7054
Langmuir 2008, 24, 7054-7057
Highly Mobile Iron Pool from a Dissolution-Readsorption Process John S. Loring,* Anna A. Simanova, and Per Persson Department of Chemistry, Umeå UniVersity, SE-901 87 Umeå, Sweden ReceiVed March 12, 2008. ReVised Manuscript ReceiVed April 25, 2008 Oxalate (C2O42-) acts synergistically on the dissolution of goethite (R-FeOOH) in the presence of siderophores that are secreted by plants and microorganisms to sequester iron. We report here the first in situ molecular-scale observations of synergistic ligand-promoted dissolution processes. We show that there are conditions under which oxalate promotes goethite dissolution, but dissolved Fe(III) concentrations do not increase because Fe(III)-oxalate complexes readsorb to the mineral surface. We demonstrate that these readsorbed Fe(III)-oxalate complexes are highly mobile, extremely reactive in the presence of uncomplexed siderophores, and responsible for the synergistic effects on the dissolution of goethite.
Fe(III) is an essential nutrient for life on our planet, yet the solubilities of Fe(III)-bearing minerals are typically parts-perbillion or less at circumneutral pH.1 Geochemists are trying to understand the processes that sustain bioavailable Fe(III), despite its extreme insolubility.2–4 Small carboxylate ligands are ubiquitous in the environment and can promote the dissolution of Fe(III)-(hydr)oxide minerals.5 Minerals dissolve via a progressive ligand exchange reaction that results in the detachment of a metal from the mineral surface;6 and the dissolution of (hydr)oxide minerals by the small carboxylate ligand, oxalate (C2O42-), is the textbook example for ligand-promoted mineral dissolution.5,7–10 Yet, it has recently been shown that oxalate acts synergistically on the dissolution of goethite (R-FeOOH) in the presence of siderophores, even under conditions where oxalate by itself does not increase dissolved iron concentrations.11,12 While several macroscopic studies concerning this synergistic effect have appeared in the literature,11–17 a molecular-level understanding of this phenomenon is missing, and the roles of small ligands on mineral dissolution are still not completely understood. * Corresponding author. (1) Cornell, R. M.; Schwertmann, U. The iron oxides: structure, properties, reactions, occurences and uses, 2nd ed.; Wiley: Cambridge, 2003; p 664. (2) Kraemer, S. M.; Crowley, D. E.; Kretzschmar, R. AdV. Agron. 2006, 91, 1–46. (3) Kraemer, S. M.; Butler, A.; Borer, P.; Cervini-Silva, J. ReV. Miner. Geochem. 2005, 59, 53–84. (4) Kappler, A.; Straub, K. L. ReV. Miner. Geochem. 2005, 59, 85–108. (5) Stumm, W.; Sigg L.; Sulzberger, B. Chemistry of the solid-water interface: processes at the mineral-water and particle-water interface in natural systems; Wiley: New York, 1992; p 428. (6) Casey, W. H.; Ludwig, C. Nature 1996, 381, 506–509. (7) Sposito, G. The surface chemistry of natural particles; Oxford University Press: Oxford, 2004; p 242. (8) Stumm, W.; Morgan, J. J. Aquatic chemistry: chemical equilibria and rates in natural waters, 3rd ed.; Wiley: New York, 1996; p 1022. (9) Morel, F. M.; Hering, J. G. Principles and applications of aquatic chemistry; Wiley: New York, 1993; p 588. (10) Sparks, D. L. EnVironmental soil chemistry, 2nd ed.; Academic Press: London, 2003; p 352. (11) Cervini-Silva, J.; Sposito, G. EnViron. Sci. Technol. 2002, 36, 337–342. (12) Reichard, P. U.; Kraemer, S. M.; Frazier, S. W.; Kretzschmar, R. Plant Soil 2005, 276, 115–132. (13) Wolff-Boenisch, D.; Traina, S. J. Chem. Geol. 2007, 243, 357–368. (14) Reichard, P. U.; Kretzschmar, R.; Kraerner, S. A. Colloids Surf., A: Physicochem. Eng. Aspects 2007, 306, 22–28. (15) Reichard, P. U.; Kretzschmar, R.; Kraemer, S. M. Geochim. Cosmochim. Acta 2007, 71, 5635–5650. (16) Kraemer, S. M. Aquat. Sci. 2004, 66, 3–18. (17) Cheah, S. F.; Kraemer, S. M.; Cervini-Silva, J.; Sposito, G. Chem. Geol. 2003, 198, 63–75.
Cervini-Silva and Sposito were the first to show that the steadystate dissolution rate of goethite in the presence of both oxalate and the siderophore, desferrioxamine-B (DFO-B), is typically greater than the sum of the dissolution rates in the presence of the two ligands alone.11 Later, Reichard et al. demonstrated that prereaction of oxalate with a goethite suspension produces a reactive iron pool that is consumed in a fast non-steady-state dissolution reaction after the addition of DFO-B.12 Subsequently, Reichard et al. presented a conceptual model with corresponding rate laws that accounts for these steady-state and non-steadystate dissolution processes.14,15 All these previous investigations were based on macroscopic experimental methods. We show a combination of infrared spectroscopic, ligand adsorption, and dissolved iron results and present a molecularlevel interpretation of the dissolution of goethite in the presence of oxalate and DFO-B that considerably expands the findings of Cervini-Silva, Reichard, et al.11–17 The key data are infrared spectra of adsorbed oxalate because they reveal the manner in which the ligand is coordinated. To obtain these spectroscopic results, we used an innovative method developed in our laboratory for simultaneous infrared and potentiometric titrations of mineral surfaces. Our experiments were performed at 25.0 ( 0.5 °C, in the absence of visible light, in 100 mM NaCl ionic medium, in ∼10 g/L goethite suspensions, using either 90.1 or 104.5 m2/g goethite (N2 BET method), and at pH of either 4.0 or 6.0 ( 0.1 (see Supporting Information). We identify the coordination geometry of oxalate using the spectra of model solution complexes (see Supporting InformationFigure 1) as a guide. Oxalate in an inner-sphere coordination complex is bound directly to Fe(III) in a bidentate mononuclear configuration, while oxalate in an outer-sphere coordination complex is separated from Fe(III) by one or more waters. Reichard et al.14,15 observed a correlation between the time that a goethite suspension is prereacted with oxalate in the absence of DFO-B and the concentration of Fe(III) that is released immediately after the addition of the siderophore. We designed a simple yet informative infrared spectroscopic experiment to investigate this phenomenon. Spectra were measured at the water-goethite interface for 5 days after the addition of a 1.1 µmol/m2 total concentration of oxalate to a suspension of goethite at pH 6 (Figure 1A). The time dependence of the spectra show a gradual transition from predominantly outer- to mostly innersphere coordinated oxalate after the system was nearly equilibrated, at the end of the fifth day. Next, we added a 1.2 µmol/m2
10.1021/la800785u CCC: $40.75 2008 American Chemical Society Published on Web 06/13/2008
Letters
Langmuir, Vol. 24, No. 14, 2008 7055
Figure 2. Infrared spectra were collected as a function of time for two days after the addition of a 1.3 µmol/m2 concentration of oxalate to a goethite suspension (surface area ) 104.9 m2/g) at pH 6. After a given prereaction time with oxalate, an aliquot of the suspension was collected, a 1.3 µmol/m2 concentration of DFO-B was added, and the mixture was reacted for 1.5 h. Shown in this figure is the concentration of dissolved iron released after the addition of the siderophore versus the integrated absorbance of the νCsOb + νCsC band of inner-sphere coordinated oxalate, measured just before the addition. The R2 value is from the least-squares fit of a line through the data between 0.19 and 0.31 arbitrary units.
total concentration of DFO-B, and spectra were collected for an additional day. Most dramatic were the changes observed during the first 1.5 h after the addition of the siderophore (Figure 1B): the signals due to inner-sphere coordinated oxalate were rapidly diminished and followed by a correlated increase in intensities due to outer-sphere coordinated oxalate. These observations led us to formulate the following hypotheses. The goethite dissolves when it is reacted with oxalate, but the Fe(III)-oxalate complexes that are released into solution readsorb to the mineral surface. The concentration of readsorbed Fe(III)-oxalate surface complexes increases with increasing time, and this is manifested in the infrared spectrum as an increase in the intensities of peaks due to inner-sphere coordinated oxalate. These readsorbed complexes either outcompete or consume outersphere coordinated oxalate, and thus the spectral intensity due
to outer-sphere complexes decreases. When DFO-B is added, the intensity due to inner-sphere coordinated oxalate decreases because the readsorbed complexes are highly mobile, and the Fe(III) in these complexes is quickly scavenged by uncomplexed siderophores.19 The intensity due to outer-sphere coordinated oxalate promptly increases because the ligands that were originally in the readsorbed Fe(III)-oxalate complexes immediately adsorb to the goethite surface as outer-sphere coordination complexes. These ideas are supported by results from additional experiments, presented below. It follows that there should exist a linear correlation between the concentration of Fe(III) that is in the form of the readsorbed Fe(III)-oxalate complex and the intensities of the peaks in the infrared spectrum due to this complex. To test this, we collected spectra while a goethite suspension was prereacted with 1.3 µmol/ m2 oxalate for increasing periods of time at pH 6. After a given prereaction time, a 1.3 µmol/m2 concentration of DFO-B was added to an aliquot of the suspension and reacted for 1.5 h. In Figure 2, we plot the concentration of dissolved iron released after the addition of the siderophore versus the integrated absorbance of the νbCsO + νCsC band of inner-sphere coordinated oxalate measured before the addition. This plot shows two regions: a relatively flat portion up to an integrated absorbance of about 0.19, and a linear portion between values of 0.19 and 0.31. We interpret the integrated absorbance in the flat region as being due to oxalate in inner-sphere complexes that are relatively immobile, i.e., oxalate bound to surface metals that have most of their coordination shell filled with atoms of the mineral lattice. The integrated absorbance in the linear region is due to oxalate that is inner-sphere coordinated to highly mobile Fe(III) that is accessible to uncomplexed DFO-B in solution, i.e., oxalate in readsorbed Fe(III)-oxalate complexes. Accordingly, these results indicate that 39% of the inner-sphere coordinated oxalate is in the form of readsorbed Fe(III)-oxalate complexes after a two-
(18) Edwards, D. C.; Nielsen, S. B.; Jarzecki, A. A.; Spiro, T. G.; Myneni, S. C. B. Geochim. Cosmochim. Acta 2005, 69, 3237–3248.
(19) This depletion of inner-sphere oxalate complexes must also be partly due to ligand competition, since DFO-B clearly adsorbs to the surface.
Figure 1. (A) Infrared spectra at the water-goethite interface collected as a function of time after the addition of a 1.1 µmol/m2 concentration of oxalate to a goethite suspension (surface area ) 90.1 m2/g) at pH 6. Dissolved iron concentrations were below our detection limit (0.0005 µmol/m2 Fe). (B) Infrared spectra collected as a function of time after the addition of a 1.2 µmol/m2 concentration of DFO-B to the goethite suspension at the end of the fifth day. The spectrum collected in an independent experiment of only DFO-B adsorbed at the goethite surface is also shown and offset for clarity; the tentative assignments of this spectrum are based primarily on ref.18
7056 Langmuir, Vol. 24, No. 14, 2008
Letters
Figure 3. The integrated absorbance between 1495 and 1345 cm-1 and the concentrations of uncomplexed and adsorbed DFO-B measured as a function of time after the simultaneous addition of 0.96 µmol/m2 and 0.23 µmol/m2 total concentrations of oxalate and DFO-B, respectively, to a goethite suspension (surface area ) 104.9 m2/g) at pH 6.
µmol/m2
day reaction period. Furthermore, 0.055 is the concentration of dissolved iron that accounts for the linear region of the data, and this corresponds to the concentration of Fe(III) in the readsorbed complexes. An important corollary to our dissolution-readsorption hypothesis is the prediction that as long as there is uncomplexed DFO-B present in the bulk solution, then the concentration of readsorbed Fe(III)-oxalate surface complexes will be small. In a second experiment, we tested this by measuring concentrations of adsorbed and uncomplexed DFO-B and collecting infrared spectra at the water-goethite interface as a function of time after the simultaneous addition of 0.96 µmol/m2 and 0.23 µmol/m2 total concentrations of oxalate and DFO-B, respectively, to a goethite suspension at pH 6. The results are shown in Figure 3. Here, we plot the integrated absorbance between 1495 and 1345 b cm-1, where the νCsO + νCsC vibration of inner-sphere coordinated oxalate and various vibrations related to the skeletal backbone of the siderophore absorb.18 Because the concentration of adsorbed DFO-B remains relatively constant, we assume that the changes in the integrated absorbance within this spectral region primarily indicate changes in the concentration of innersphere coordinated oxalate. Uncomplexed DFO-B exists in solution for the first 1500 min, during which time the integrated absorbance increases and plateaus. After 1500 min, the concentration of uncomplexed DFO-B is zero, and the integrated absorbance sharply increases. These results are consistent with the slow formation of highly mobile readsorbed Fe(III)-oxalate complexes. When uncomplexed DFO-B is present in solution, the steady-state concentration of readsorbed complexes is small because the rate at which these complexes form is slower than both the rate at which they desorb and the rate at which their Fe(III) is sequestered by uncomplexed siderophore in solution. Therefore, the integrated absorbance primarily reflects the adsorption of DFO-B and the formation of surface-metal innersphere oxalate complexes. However, when the concentration of uncomplexed DFO-B is zero, the integrated absorbance increases because of the buildup of readsorbed complexes at the mineral surface. We find indirect evidence for a dissolution-readsorption process from a comparison of two adsorption experiments: one
Figure 4. Infrared spectra collected as a function of time after the addition of a 0.33 µmol/m2 concentration of trioxalatoiron(III) (spectra offset of clarity) or a 1.0 µmol/m2 concentration of oxalate to a goethite suspension (surface area ) 90.1 m2/g) at pH 4.
where infrared spectra were collected as a function of time after a 1.0 µmol/m2 total concentration of oxalate was added to a goethite suspension and another where the same concentration of oxalate was added but in the form of the trioxalatoiron(III) anion. These experiments were performed at pH 4 because the trioxalatoiron(III) anion is predominant at this pH and not expected to precipitate as amorphous Fe(III)-hydroxide.20 The spectra (Figure 4) from the oxalate-adsorption experiment indicate a transition from mostly outer- to inner-sphere coordinated ligands similar to what was observed at pH 6 (see Figure 1A), while the spectra from the trioxalatoiron(III)-adsorption experiment show an increase in inner-sphere coordinated oxalate, but the concentration of outer-sphere coordinated oxalate remains relatively small and constant. At equilibrium for both experiments, the oxalate was completely adsorbed to the surface, no dissolved iron was detected in solution, and the infrared spectra were practically identical. However, it took two hours to reach equilibrium in the trioxalatoiron(III)-adsorption experiment, while it took two days in the oxalate experiment. Two important conclusions can be drawn. First, Fe(III)-oxalate solution complexes clearly adsorb, although the mechanism is undoubtedly complicated. Lifetimes for ligand exchange are on the order of micro- to milliseconds for aqueous Fe(III) complexes,21 and we do not necessarily presume that the trioxalatoiron(III) anion adsorbs intact. Second, the main difference between these adsorption experiments is the iron source, and this likely explains the strikingly different times required to reach equilibrium. In the oxalate-adsorption experiment, the only source of iron is from the goethite, and the longer time for equilibration is consistent with a relatively slow dissolution process. However, in the trioxalatoiron(III) experiment, the complex itself is an iron source, and the shorter time for equilibration is consistent with predissolved Fe(III) adsorbing directly to the mineral surface. The key to the synergism between DFO-B and oxalate on goethite dissolution is the readsorbed Fe(III)-oxalate complex. This complex is not an intermediate14,15 to the ligand-promoted dissolution reaction; it is an end result. In fact, the thermodynamic stability of the readsorbed complex drives the dissolution in the (20) Martell, A. E.; Smith, R. M. NIST Critically Selected Stability Constants of Metal Complexes, 4.0; NIST Standard Reference Data; 1997. (21) Casey, W. H.; Swaddle, T. W. ReV. Geophys. 2003, 41, 1–20.
Letters
absence of uncomplexed DFO-B. Hence, the extent of dissolution must be controlled by parameters such as pH, ionic strength, mineral surface area, and concentrations of competing ligands. No surface-complexation model has yet been published that accounts for a readsorbed Fe(III)-oxalate complex,22,23 but geochemists need models that can predict the stabilities of these types of highly reactive species. This presents a significant challenge to the field, since spectroscopists must first determine the structures and concentrations of these complexes before modelers can constrain their efforts to produce realistic models. Similar changes in surface speciation with time, as seen in Figure 1A, have been observed for oxalate adsorbed on boehmite (γ-AlOOH),24,25 and small ligands other than oxalate have been shown to cause synergistic effects on the dissolution of goethite.15 Hence, it is likely that dissolution-readsorption is not exclusive (22) Sverjensky, D. A.; Fukushi, K. EnViron. Sci. Technol. 2006, 40, 263–271. (23) Filius, J. D.; Hiemstra, T.; Van Riemsdijk, W. H. J. Colloid Interface Sci. 1997, 195, 368–380. (24) Axe, K.; Persson, P. Geochim. Cosmochim. Acta 2001, 65, 4481–4492. (25) Yoon, T. H.; Johnson, S. B.; Musgrave, C. B.; Brown, G. E. Geochim. Cosmochim. Acta 2004, 68, 4505–4518.
Langmuir, Vol. 24, No. 14, 2008 7057
to the oxalate-goethite system, but is a general process that commonly occurs in ligand-mineral systems. Conventional thinking is that metals adsorbed at mineral surfaces are not accessible to biota. However, we have demonstrated a case where readsorbed metal-ligand complexes are highly mobile and, in all likelihood, readily bioavailable. This has important implications because it suggests that standard measurements of dissolved metal concentrations in solution could lead to significant underestimates of metal bioavailability. Acknowledgment. William H. Casey is gratefully acknowledged for his valuable comments. Funding for this research was from the Swedish Research Council and the Kempe Foundation. Supporting Information Available: Materials and methods. Infrared spectra of model solution species for inner- and outer-sphere coordinated oxalate (the aqueous trioxalatoiron(III) and oxalate anions, respectively). This material is available free of charge via the Internet at http://pubs.acs.org. LA800785U