Article
The depletion and regeneration of dissolution-active sites at the mineral-water interface:: I. Fe, Al, and In sesquioxides

https://doi.org/10.1016/S0016-7037(00)00462-2Get rights and content

Abstract

The responses of α-Al2O3 and In2O3 dissolution rates to pH-jumps in flow-through reactors are compared to previous results for α-Fe2O3 in order to test the following hypothesis: because the same fundamental processes of ligand exchange and detachment of metal centers govern both steady state and nonsteady state dissolution, both the steady state and nonsteady state dissolution rates of these sesquioxides correlate with the water exchange rates of the corresponding aqueous ions. Our results show that steady state dissolution rates at pH 1 for α-Fe2O3, α-Al2O3, and In2O3 correlate with the rates of water exchange.

Nonsteady state dissolution rates, in response to jumps to pH 1 from higher initial pH values at which the oxides had been aged, also correlate with water exchange rates although in a complicated way. The amount of solute dissolved during nonsteady state dissolution increases with increasing initial pH, and is consistent with the depletion at low pH and regeneration at higher pH of a reservoir of surface sites active for dissolution at pH 1. This suggests that the oxide surfaces must be sufficiently kinetically dynamic at circumneutral pH for processes to occur that result in the creation of dissolution-active sites. The pH dependence of the amount of solute released can be related to rates of active site production and dissolution, using the idea that hydroxylation of the mineral surface with increasing solution pH has a labilizing effect on the rates of ligand exchange at a surface metal center similar to the labilizing effect of hydrolysis on the rates of water exchange for aquaions.

Introduction

Mineral dissolution has been the subject of extensive study because of its importance in a variety of geochemical processes ranging from global elemental cycling and climate control to acid rain effects and biochemical strategies for nutrient acquisition. Our understanding of the effects of pH, solution composition, and adsorbed species on dissolution rate is based largely on steady state dissolution measurements. For example, dissolution rates are often related to the concentration of protons or organic (or inorganic) ligands adsorbed to the mineral surface (e.g., Stumm and Wieland, 1990). Such surface complexation approaches to dissolution kinetics are based explicitly on the assumption that the surface is in a steady state configuration; that is, the surface concentration of metal centers participating in dissolution, and their collective overall reactivity toward dissolution, are time independent. Here, in contrast, we utilize pH-jump-induced dissolution transients, caused by temporally changing concentrations and reactivities of dissolution-active surface metal centers, to obtain information about dissolution processes that cannot be obtained from steady state experiments.

Initially elevated dissolution rates that decay to a steady state are a virtually universal observation in mineral dissolution experiments Holt and King 1955, Holdren and Berner 1979, Schott et al 1981, Chou and Wollast 1985, Carroll-Webb and Walther 1988, Knauss et al 1993, Hellmann 1995, Maurice et al 1995, Kraemer and Hering 1997, Malmstrom and Banwart 1997. Occasionally dissolution rates increase to a steady state, e.g., quartz at pH > 10 (Knauss and Wolery, 1988). Such initial transients show that the approach of dissolution rate to steady state takes time, and have been attributed to artifacts of sample preparation Petrovic 1981a, Petrovic 1981b, Schott et al 1981, the formation of altered layers on mineral surfaces Casey and Bunker 1990, Hellmann 1995, or transient adjustments of surface topography or particle size distribution toward a steady state configuration Holdren and Berner 1979, Dibble and Tiller 1981. Rate transients, however, are not limited to the onset of dissolution, but have been reported in response to cycles in pH Chou and Wollast 1984, Spokes and Jickells 1996, Samson and Eggleston 1998, increases in ligand concentration Mast and Drever 1987, Wieland and Stumm 1992, Kraemer and Hering 1997, and changes in electrolyte composition (Sjöberg, 1989). Furthermore, such transients have been shown to contain useful information about dissolution mechanisms. For example, Holt and King (1955) related rapid initial dissolution rates to the adsorption of silicic acid to quartz surfaces, and Chou and Wollast (1984) used pH-jump-induced albite dissolution transients to obtain information about the formation, thickness, and composition of altered or leached layers on the feldspar surface. Similarly, we follow these authors in using nonsteady state phenomena as a window on oxide dissolution mechanisms.

A previous study of nonsteady state hematite dissolution (Samson and Eggleston, 1998) showed that following downward pH-jumps to pH 1, a consistent, reproducible, and regenerable nonsteady state period of elevated dissolution rate occurs and lasts for over 36 h. The results were consistent with the depletion at low pH and regeneration at higher pH of a reservoir of Fe surface sites active for dissolution at pH 1 in response to cycles in pH. Such active sites can be thought of in different ways. We interpreted the data with regard to long-standing crystal growth and dissolution models (e.g., Burton-Cabrera-Frank, BCF; Burton et al., 1951) that assume the existence of “adsorbed nutrient”, e.g., Fe3+ on hematite, that is structurally distinct from metal centers in the solid surface structure. We attributed the transients to the release of this adsorbed Fe.

Here, we report results for nonsteady state dissolution of corundum (α-Al2O3) and cubic indium sesquioxide (In2O3) in response to pH-jumps, compare them to results for hematite (α-Fe2O3), and use the data to test a simple hypothesis: because the same fundamental processes of ligand exchange and detachment of metal centers govern both steady state and nonsteady state dissolution, both the steady state and nonsteady state dissolution rates of these trivalent metal sesquioxides correlate with the water exchange rates of the corresponding aqueous ions. The model originally developed for hematite (Samson and Eggleston, 1998) is refined in order to describe the wider range of nonsteady state behavior defined by the new data. Finally, we reconsider the idea that the amount of solute released in the transients is directly related to the adsorption of nutrient to the oxide surface (e.g., FeIII to hematite, or Al to corundum), and consider whether the amount of adsorbed nutrient involved in transient dissolution is, at the initial higher pH, controlled by equilibrium or by steady state kinetics.

Section snippets

Experimental methods

We used the isostructural minerals α-Al2O3 and α-Fe2O3 (corundum structure, space group R3c), but a third suitable analog (e.g., corundum structure of a metal with a stable oxidation state and known water exchange rate that differs significantly from those of Al3+ and Fe3+) was not readily obtainable. In2O3, which has a cubic structure (space group Ia3), but the same stoichiometry, was taken as the best alternative.

Results

The steady state dissolution rates at pH 1 for the three sesquioxides correlate with the water exchange rates for the hexaaqua ions (Fig. 3). With respect to nonsteady state dissolution following downward pH-jumps to pH 1, there are consistent, reproducible, pH-dependent transients for all three sesquioxides (Fig. 4) as discussed individually below.

The results of downward pH-jump experiments with hematite (Fig. 4a) are described elsewhere (Samson and Eggleston, 1998). Briefly, following a

Discussion

Our hypothesis that steady state dissolution rates correlate with the rates of water exchange for the corresponding aqueous ions is supported by the data (Fig. 3). We have also shown that pH-jumps to pH 1 from higher pH consistently produce transients that are initial-pH-dependent, reproducible, and regenerable (Fig. 4).

Because the hematite and corundum powders were cycled between pH 1 and higher pH, the elevated dissolution rates that occur in response to downward pH-jumps might be interpreted

Conclusions and summary

  • 1.

    The steady state dissolution rates of α-Fe2O3, α-Al2O3, and In2O3 at pH 1 correlate with the water exchange rates of the corresponding aqueous ions. The range in the dissolution rates for the three phases, however, is considerably less than the 7 orders of magnitude difference in the water exchange rates for the aqueous ions, suggesting that factors other than the inherent properties of the metal may limit the extent to which dissolution rates vary for isostructural (or nearly isostructural)

Acknowledgements

We are grateful to Pradeep Agarwal and Steve Higgins for their valuable assistance with the reactor equations. We also thank Steve Boese for analytical assistance, and three anonymous reviewers for their thorough reviews and helpful comments. Financial support was provided by The National Science Foundation (Grant #EAR-9527031 to CME).

References (55)

  • S.M. Kraemer et al.

    Influence of solution saturation state on the kinetics of ligand-controlled dissolution of oxide phases

    Geochim. Cosmochim. Acta

    (1997)
  • D.H. Lee et al.

    An FTIR spectral investigation of the structural species found on alumina surfaces

    Mater. Lett.

    (1995)
  • S.F. Lincoln et al.

    Substitution reactions of solvated metal ions

  • P. Liu et al.

    Reaction of water vapor with α-Al2O3 (0001) and α-Fe2O3 (0001) surfacesSynchrotron X-ray photoemission studies and thermodynamic calculations

    Surface Sci

    (1998)
  • C. Ludwig et al.

    The effect of different functional groups on the ligand-promoted dissolution of NiO and other oxide minerals

    Geochim. Cosmochim. Acta

    (1996)
  • M. Malmström et al.

    Biotite dissolution at 25°CThe pH dependence of dissolution rate and stoichoimetry

    Geochim. Cosmochim. Acta

    (1997)
  • B.L. Phillips et al.

    Solvent exchange in AlFx(H2O)6−x3−x(aq) complexesLigand-directed labilization of water as an analogue for ligand-induced dissolution of oxide minerals

    Geochim. Cosmochim. Acta

    (1997)
  • N. Sahai et al.

    Evaluation of internally consistent parameters for the triple-layer model by the systematic analysis of oxide surface titration data

    Geochim. Cosmochim. Acta

    (1997)
  • J. Schott et al.

    Mechanism of pyroxene and amphibole weatheringI. Experimental studies of iron-free minerals

    Geochim. Cosmochim. Acta

    (1981)
  • T. Sugimoto et al.

    Characterization of hematite particles of different shapes

    J. Colloid Interface Sci.

    (1993)
  • E. Wieland et al.

    The coordination chemistry of weatheringIII. A generalization on the dissolution rates of minerals

    Geochim. Cosmochim. Acta

    (1988)
  • S. Ambe et al.

    Mossbauer study of ferric ions adsorbed at α-ferric oxide/aqueous solution interface

    Langmuir

    (1990)
  • C.F. Baes et al.

    The Hydrolysis of Cations

    (1976)
  • W.K. Burton et al.

    The growth of crystals and the equilibrium structure of their surfaces

    Phil. Trans. Roy. Soc. London Ser. A.

    (1951)
  • W.H. Casey et al.

    Leaching of mineral and glass surfaces during dissolution

  • W.H. Casey et al.

    Silicate mineral dissolution as a ligand-exchange reaction

  • W.H. Casey et al.

    Control of dissolution rates of orthosilicate minerals by divalent metal-oxygen bonds

    Nature

    (1992)
  • Cited by (0)

    Present address: Division of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138 USA.

    Present address: United States Geological Survey, University of NV-Reno, Reno, NV 89557 USA.

    View full text