Environ. Sci. Technol. 2005, 39, 2645-2650
Reduction of Organically Complexed Ferric Iron by Superoxide in a Simulated Natural Water ANDREW L. ROSE AND T. DAVID WAITE* School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia
Superoxide (and potentially its conjugate acid hydroperoxyl) is unique among the reactive oxygen species in that its standard redox potential in circumneutral natural waters potentially allows it to reduce ferric iron to the more soluble ferrous state. Here we have observed the superoxide/ hydroperoxyl-mediated reduction of ferric complexes with a variety of synthetic organic ligands and several complexes with natural organic matter (NOM), as well as freshly precipitated amorphous ferric oxyhydroxide, in bicarbonate buffered solutions at pH 8.1. From measurements of superoxide decay in the presence of the complexes, we calculated second-order rate constants for superoxide/ hydroperoxyl-mediated reduction that vary from (9.3 ( 0.2) × 103 M-1 s-1 for the complex between Fe(III) and desferrioxamine B up to (1.9 ( 0.2) × 105 M-1 s-1 for Fe(III)-salicylate and (2.3 ( 0.1) × 105 M-1 s-1 for one of the Fe(III)-NOM complexes. We also verified that ferrous iron was produced from superoxide/hydroperoxylmediated Fe(III) reduction using ferrozine to trap free Fe(II). Low yields of the ferrozine complex when compared to the measured rates of superoxide decay suggest that ferric complexes are reduced directly to corresponding ferrous complexes, with much of the ferrous complex reoxidizing before it is able to release free ferrous iron. This is an important consideration for microorganisms, as the kinetics of trace metal uptake is typically governed by free ion activity.
Introduction In natural waters, iron exists primarily in one of two redox states: the reduced ferrous (+II), or oxidized ferric (+III) state. Iron can undergo redox reactions with dissolved dioxygen as well as so-called reactive oxygen species (ROS) that occur when oxygen is present in intermediate redox states between its fully oxidized form, dioxygen, and fully reduced form, water. On the basis of the redox potentials of these species and iron, the ferric state is strongly thermodynamically favored in oxygenated waters around neutral pH. Ferric iron is highly insoluble at circumneutral pH (except when solubilized by organic ligands), which typically leads to extremely low concentrations of dissolved iron in most marine waters and many freshwaters. As iron, which is an essential micronutrient for nearly all microorganisms (1), is taken up only in dissolved form, processes which reduce iron to the more soluble ferrous state are therefore biologically important. The superoxide free radical (O2•-) and potentially its conjugate acid hydroperoxyl radical are unique among the * Corresponding author phone: +61-2-9385 5060; fax: +61-29385 6139; e-mail:
[email protected]. 10.1021/es048765k CCC: $30.25 Published on Web 03/09/2005
2005 American Chemical Society
ROS in that the potentials of their redox couples with H2O2 and O2, respectively, theoretically permit not only oxidation of ferrous iron, but also reduction of many forms of ferric iron to the ferrous state under standard conditions at circumneutral pH (2). However, superoxide is inherently unstable due to its propensity for self-reaction (disproportionation) (3). Superoxide undergoes protonation to form its conjugate acid, hydroperoxyl radical, with pKa ) 4.8
O2•- + H+ h HO2•
(1)
Disproportionation reactions occur between either two hydroperoxyl radicals, or hydroperoxyl and superoxide as follows:
HO2• + HO2• f O2 + HOOH
(2)
HO2• + O2•- f O2 + HOO-
(3)
Because of different rate constants for the reactions in eqs 2 and 3, the apparent rate constant for superoxide disproportionation varies as a function of pH. The third possibility of self-reaction between two superoxide anion radicals is negligibly slow compared to these reactions (3). As we consider only a fixed pH value in this work, the relative importance of the hydroperoxyl and superoxide species is not considered further and the disproportionation reaction is denoted for simplicity as follows: 2H+
O2•- + O2•- 98 O2 + HOOH
(4)
Also for simplicity, the term “superoxide” will be used hereafter to describe the superoxide/hydroperoxyl conjugate pair, unless specifically stated otherwise. However, it should be borne in mind that the relative contribution of each species in the conjugate pair to the chemistry investigated is not evaluated in this work. Superoxide disproportionation can be catalyzed by a variety of substances that exist in dual redox states (including iron (4), copper (5, 6) and natural organic matter(NOM) (7)), with a net reaction given by eq 4, according to the individual reactions
O2•- + A(n+1)+ f O2 + An+
(5)
O2•- + An+ f O22- + A(n+1)+
(6)
where A is the catalyst. Superoxide may be generated biotically or abiotically, and is involved in iron redox cycling under a variety of conditions. In both fresh and marine waters, superoxide formation has been observed from the photochemical reactions of dissolved NOM (8) and has been subsequently implicated in the photochemical formation of ferrous iron (9, 10). Superoxide is also generated both extra- and intracellularly in microorganisms (11, 12), where it may be involved (among other things) with iron acquisition and iron metabolism, respectively. However, the reduction of ferric iron by superoxide has actually been demonstrated only for dissolved inorganic iron (4) and complexes with a few simple organic ligands, such as EDTA (13, 14). Although it is thermodynamically expected that superoxide will reduce a wide range of ferric-organic complexes with suitable redox potentials (2, 15), this has not been shown directly. In theory, superoxide-mediated reducVOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2645
FIGURE 1. Mechanism for superoxide-mediated reduction of Fe(III) complexes illustrating two alternative pathways. In pathway (1), Fe(III) complexes are reduced directly to the corresponding Fe(II) complex. In pathway (2), Fe(III) complexes first dissociate, with the resulting inorganic Fe(III) then reduced by superoxide to inorganic Fe(II). Additional reactions that may also occur under the experimental conditions used in this work are shown with dashed lines. tion of ferric iron may proceed via one of two pathways, as shown in Figure 1. In the first pathway, superoxide reacts directly with the ferric-organic complex to produce the corresponding ferrous-organic complex according to the reaction
O2•- + Fe(III)L f O2 + Fe(II)L
(7)
The ferrous-organic complex may then be re-oxidized in natural waters by a variety of oxidants including dissolved oxygen, superoxide, hydrogen peroxide, hydroxyl radicals, and organic radicals without dissociating to form free (inorganic) ferrous iron. In the second pathway, the ferricorganic complex first dissociates to give free (dissolved inorganic) ferric iron, which reacts with superoxide to produce free ferrous iron
Fe(III)L f Fe(III) + L
(8)
Fe(III) + O2•- f Fe(II) + O2
(9)
In this case the net reaction is similar to that shown in eq 7, except that the product of iron reduction is free ferrous iron, rather than a ferrous-organic complex. In this work, we have investigated the reduction of ferric iron by superoxide in bicarbonate buffered solutions at pH 8.1. We examine the superoxide-mediated reduction of several ferric complexes with simple organic ligands exhibiting a wide range of conditional stability constants. In addition, we determine the apparent rates of superoxide-mediated reduction of complexes of ferric iron with a variety of samples of terrigenous NOM, the conditional stability constants of which have been determined in a previous study (16). Finally, we consider the mechanism for superoxide-mediated reduction of iron in terms of the two possible pathways described above.
Experimental Section Reagents. Chemicals were used as received. All solutions were prepared using 18 MΩ‚cm Milli-Q water (MQ). All pH 2646
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005
measurements were made using a Hanna HI9025 pH meter, calibrated on the free hydrogen scale with WTW pH 7.00 and pH 10.01 buffers, and pH adjustments performed using high purity 30% w/v HCl and 32% w/v NaOH (Fluka puriss p.a. plus). A solution of high-purity 0.2 M HCl for reagent preparation was made by dilution of 30% w/v HCl (Fluka puriss p.a. plus). A 4.0 mM Fe(II) stock in 0.2 M HCl was prepared from ferrous ammonium sulfate hexahydrate and replaced yearly. A working stock of 2.0 µM Fe(II) in 0.2 mM HCl was prepared weekly from the 4.0 mM Fe(II) stock with addition of 0.2 M HCl. A 3 mM stock solution of Fe(III) from FeCl3‚6H2O (Ajax chemicals) was prepared in 0.1 M HCl. A 500 µM Fe(III) stock solution was prepared by dissolution of FeCl3‚6H2O (Ajax chemicals) in 2 mM HCl at pH ∼2.7 (the pH at which iron does not precipitate but which is sufficiently high to prevent significant pH change upon addition to experimental solutions). All iron stock solutions were stored in the dark at 4 °C when not in use. Methyl cypridina luciferin analogue (MCLA, Fluka) reagent for use in chemiluminescence measurement of superoxide was prepared as 1 µM MLCA in 0.05 M acetate buffer at pH 6.0, stored in a dark glass bottle, and refrigerated at 4 °C when not in use. A stock solution of 0.1 M ferrozine (3-(2pyridyl)-5,6-diphenyl-1,2,4-triazine-p,pV-disulfonic acid, monosodium salt hydrate, Fluka) at pH 8 ( 1 was prepared in MQ with dropwise addition of 32% w/v NaOH for pH adjustment. A stock solution of 13.4 mM xanthine (Sigma) was prepared in MQ at pH 9 using 32% w/v NaOH for pH adjustment. Xanthine oxidase (Sigma) was dissolved in MQ to a final concentration of 1 unit mL-1 on receipt, and 1-mL aliquots were individually frozen at -86 °C until use. A stock solution of 15 mM diethylenetriaminepentaacetate (DTPA) (Fluka) was also prepared for use in superoxide generation. All experiments were conducted in 2 mM NaHCO3 and 10 mM NaCl at pH 8.1. Ferric Iron Complexes. Complexes of ferric iron at pH 8.1 ( 0.2 with 5-sulfosalicylate (Sigma), salicylate (Sigma), citrate (Sigma), ethylenediaminetetraacetate (EDTA, Ajax chemicals), desferrioxamine B (DFB, Sigma) and DTPA (Fluka) were prepared at a final concentration of 500 µM iron in 10 mM ligand, and a complex of ferric iron with Suwannee River fulvic acid (SRFA) was prepared at a final concentration of 500 µM iron in 400 mg L-1 SRFA, as described in the Supporting Information. A suspension of amorphous ferric oxyhydroxide (AFO) was prepared by spiking 500 µM Fe(III) stock directly into NaHCO3/NaCl solution and used within 2 min of preparation. In addition, samples of several types of dissolved NOM were provided by the Queensland Department of Natural Resources and Mines. These samples were used in this study to relate the rates of reduction of iron complexes by superoxide with previously determined kinetic and thermodynamic properties of their ferrous and ferric complexes (16). Briefly, the solutions were produced by leaching with rainwater the coarse (>2 mm) and fine (
