Environ. Sci. Technol. 2010, 44, 9337–9342
Oxygen and Superoxide-Mediated Redox Kinetics of Iron Complexed by Humic Substances in Coastal Seawater M A N A B U F U J I I , * ,†,| A N D R E W L . R O S E , ‡ T. DAVID WAITE,† AND TATSUO OMURA§ School of Civil and Environmental Engineering, The University of New South Wales, Sydney, New South Wales 2052, Australia, Southern Cross GeoScience, Southern Cross University, Lismore 2480, Australia, and Department of Civil and Environmental Engineering, Graduate School of Engineering, Tohoku University, 6-6-06 Aoba Aramaki Aobaku Sendai Miyagi 980-0859, Japan
Received July 29, 2010. Revised manuscript received October 22, 2010. Accepted October 27, 2010.
Complexes with terrestrially derived humic substances represent one of the most reactive pools of dissolved Fe in natural waters. In this work, redox kinetics of Fe-humic substance complexes (FeL) in simulated coastal seawater were investigated using chemiluminescence techniques with particular attention given to interactions with dioxygen (O2) II and superoxide (O•2 ). Although rate constants of Fe L oxidation by O2 (5.6-52 M-1 · s-1) were 4-5 orders of magnitude less 5 -1 -1 than those for O•2 (6.9-23 × 10 M · s ), O2 is likely to outcompete •II O2 for Fe L oxidation in coastal seawaters where steadystate O•2 concentrations are generally subnanomolar. Rate 4 -1 -1 constants for FeIIIL reduction by O•2 of 1.8-5.6 × 10 M · s II were also determined. From the balance of Fe L oxidation III II rates and O•2 -mediated Fe L reduction rates, steady-state Fe L concentrations were estimated to be in the subpicomolar to picomolar range, which is generally lower than measured in situ Fe(II) concentrations under relevant conditions. This suggests that (i) processes other than O•2 -mediated reduction (such as photochemical ligand-to-metal charge transfer) may be responsible for Fe(II) formation, (ii) the in situ ligands differ significantly from the humic substances used in this work, and/ or (iii) the influence of other environmental factors such as pH and temperature on Fe redox kinetics may have to be considered.
Introduction Iron (Fe) is one of the most interesting trace metals in marine science as Fe availability controls primary productivity in many open ocean (1) and coastal upwelling regimes (2). In natural waters, Fe bioavailability is strongly influenced by its redox kinetics (3). Compared to the ferric (Fe(III)) state, the reduced ferrous (Fe(II)) state of Fe is more soluble and more weakly complexed by natural organic matter (NOM) at seawater pH such that reduction of Fe(III) via photochemical * Corresponding author e-mail:
[email protected]. † The University of New South Wales. ‡ Southern Cross University. § Tohoku University. | Department of Civil Engineering Tokyo Institute of Technology 2-12-1-M1-4 Ookayama, Tokyo 152-8552, Japan. 10.1021/es102583c
2010 American Chemical Society
Published on Web 11/15/2010
or biological processes is critical for Fe acquisition by microorganisms in oxygenic surface seawaters (4). However, redox reactions involving Fe are dynamic and far from equilibrium with the result that quasi-steady-state Fe(II) concentrations, when measurable, are highly variable spatially and temporally depending on environmental parameters (5-7). Many aspects of Fe redox kinetics have been extensively examined, including interactions with dioxygen (O2) (8-12), hydrogen peroxide (H2O2) (13-16), superoxide (O•2 ) (17, 18), and other reactive species such as photochemical products (19). Reaction rates for these processes are typically influenced by the presence of Fe-binding ligands such as NOM. Coordination by NOM in many cases increases the rate of Fe(II) oxidation by O2 (8, 9, 11, 20, 21), while some organic ligands including in situ Fe-binding ligands may decrease the oxygenation rate (5). Although H2O2 (which is present at up to a few hundred nanomolar in surface seawater) has been recognized as a major oxidant for inorganic Fe(II) (13-15), recent findings have suggested that when complexed by humic substances (HS), the H2O2-mediated oxidation rate of Fe(II) at pH 8.4 decreases by at least a factor of ∼50, such that the contribution of H2O2 to Fe(II) oxidation was found to be unimportant compared to oxygenation under such conditions (15). The importance of H2O2 to the oxidation of dissolved Fe in the oceans, which is most likely bound to organic ligands (22), still remains unclear. O2•- is capable of reducing various types of Fe in seawater (including organically complexes and Fe oxyhydroxides) at different rates (17, 18), but its influence on rates of Fe(II) oxidation, particularly when Fe is complexed by NOM, is less well-known. Since the O2•concentration in seawater is typically subnanomolar (23-25), O2•--mediated Fe(II) oxidation has been considered to be relatively minor in many previous studies. However, to better constrain factors influencing Fe speciation and biological availability in seawater, understanding of contribution of O2•- to Fe redox reactions is crucial. In this work, the redox kinetics of Fe complexed by HS and model organic ligands are investigated with attention given to dioxygen- and superoxide-mediated redox processes in coastal seawaters, where Fe geochemistry is likely influenced to some extent by HS (26).
2. Materials and Methods 2.1. Glassware and Reagents. All glass and plasticware was acid-cleaned using 0.1 M HCl for a few days and then rinsed with ultrapure Milli-Q water (MQ; Millipore, 18 MΩ.cm resistivity) before use. All pH values are reported on the total hydrogen ion activity scale. Full descriptions of the acidcleaning procedure, chemical preparation and storage, and measurement of pH appear in the Supporting Information (SI). Coastal seawater (salinity 35‰) was collected in February 2005-5 km offshore from Matsushima Bay, Japan, in an acidwashed polyethylene container, vacuum filtered through a 0.22 µm membrane (Millipore) then stored in the acid-washed polyethylene container for at least several weeks before use to eliminate photochemical products such as H2O2. Seawater pHT ) 8.06 (i.e., on the total hydrogen ion concentration scale) was calculated from a measured value on the hydrogen ion activity scale at 25 °C of 8.20, which was corrected using an apparent ion activity coefficient of fH ) 0.73 as estimated by Millero et al. (27). Samples of HS (including soil fulvic acids, designated as SFA1-8, and aquatic humic substances, designated as AHS1-3) were isolated as lyophilized material according to VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9337
the IHSS isolation method. Soil fulvic acids were isolated from soil cores of eight representative vegetation types in the vicinity of Deception Bay, Queensland and were provided by the Queensland Department of Natural Resources, while aquatic humic substances were isolated from water samples collected at three sites in the Deception Bay catchment exhibiting markedly different values of salinity, pH, and dissolved organic carbon concentration. Abbreviations, origin, detailed isolation procedures, and characteristics of the samples are described in the SI. Hereafter, use of the term HS will include the isolated 11 humic substance samples and the reference fulvic acid, Dando fulvic acid (DFA, Japanese Humic Substance Society). Lyophilized HS was dissolved in MQ water to prepare 0.1-1.0 g · L-1 HS stock solutions. Citrate and ethylenediaminetetraacetic acid (EDTA) were also employed as model organic ligands to verify the methodology used in this work. In all experiments, pre-equilibrated solutions of organic Fe complexes (FeL) were used to minimize the participation of unchelated Fe in reactions. Stock solutions of organic Fe(II) complexes (FeIIL) were pre-equilibrated at a low Fe:HS ratio at pH ∼3, while stock solutions of organic Fe(III) complexes (FeIIIL) were pre-equilibrated at pH ∼8.2 as described in the SI. Chemiluminescence (CL) reagents luminol (5-amino-2,3dihydro-1,4-phthalazinedione (10)) and methyl cypridina luciferin analogue (2-methyl-6-(4-methoxyphenyl)-3,7-dihydroimid-azo[l,2- a]pyrazin-3(7H)-one, MCLA (28)), prepared as described in the SI, were used for the determination of Fe(II) and O2•-, respectively. A 1 mM borate buffered solution at pH 12 containing 41 mM acetone, 12 M absolute ethanol, and 15 µM diethylenetriaminepentaacetic acid was prepared for production of O2•- using the photochemical method (29). Stock solutions of xanthine (X), xanthine oxidase (XO) and ferrozine (FZ) were also prepared as described in the SI. All experiments were conducted at temperature of 25 °C under dark condition. 2.2. Experimental Procedures and Equipment. Nanomolar concentrations of Fe(II) and O2•- were measured using CL-based methods with a FeLume system (Waterville Analytical). Sample solution and CL reagent were continuously withdrawn via separate tubes by a peristaltic pump operating at 25 rpm and mixed in a spiral-shaped flow cell positioned in front of a Hamamatsu HC135 photomultiplier tube (PMT) operating at -1200 V. PMT signals were monitored every 1 s by personal computer. The time lag to reach steady flow conditions in the flow cell after the addition of sample solution was minimized to ∼15-20 s by configuring the system such that the waste and reagent lines were pumped, while the sample line was not. In the following experiments and for instrument calibration, CL signals before steady flow (including the initial signal at time zero) were estimated by extrapolating the signal collected after steady flow conditions were reached using a log-linear plot for Fe(II) measurement with luminol or a linear plot for O•2 measurement with MCLA. Potential interferences for Fe(II) and O2•- determination are discussed further in the SI. FeIIL oxygenation rate constants were determined by measuring the decay kinetics of FeIIL added to air-saturated seawater. Pre-equilibrated FeIIL stock solutions were added to seawater at final concentrations of 4 or 8 nM for Fe(II), resulting in total ligand concentrations of 1 mg · L-1 for HS and 10 µM for model organic ligands. The oxidation kinetics of FeIIL were monitored for several minutes using the FeLume system. As the presence of some HS resulted in quenching or enhancement of the CL signal, data were calibrated separately for each ligand. Since a linear relationship between the initial Fe(II) concentration and initial CL signal was obtained on a log-log plot for each 9338
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010
ligand (Figure S3), this was used to transform the CL signal data to Fe(II) concentration for all experiments using that ligand. Rate constants for reduction of FeIIIL and oxidation of II Fe L by O2•- were simultaneously determined by measuring the perturbation to steady-state O2•- concentrations when FeIIIL was added to air-saturated seawater containing the O2•--generating xanthine/xanthine oxidase (X/XO) system. FZ was also added to the seawater solution to prevent any inorganic Fe(II) that was formed during the experiment from reacting with O2•-. Initially, stock solutions of X, XO, and FZ were added to seawater at concentrations of 50 µM, 1 unit · L-1, and 1 mM, respectively. Subsequently, the O•2 concentration in the experimental solution was monitored by the FeLume system for ∼1 min until the O•2 concentration reached steadystate. The system was then perturbed by addition of FeIIIL stock solution at final concentrations of 200 nM for Fe(III) and either 2 mg · L-1 for HS or 10 µM for model organic ligands and the O2•- concentration measured for a further several minutes. Calibration of the instrument was performed by standard additions of O•2 that was photochemically generated by irradiation of a pH 12 aqueous ethanol solution with a 150 W Xe lamp. The concentration of O2•- in the borate buffer was quantified during irradiation by monitoring UV absorbance at 245 nm (ε240 ) 2350 M-1 · cm-1) (29) using an Ocean Optics USB2000 UV-vis spectrophotometer coupled to a DH-2000 deuterium tungsten halogen light source. For calibration, a linear relationship between the initial O2•concentration and initial CL signal was obtained and used to transform the CL signal data to O2•- concentration for all experiments. HS quenched the MCLA CL signal negligibly at HS concentrations examined in this work, as discussed in the SI. For all experiments, a magnetic stirrer was used to rapidly mix experimental solutions upon addition of chemicals but was turned off after ∼30 s to avoid over saturation with air. The pH of experimental solutions changed by
