with Superoxide Radical in Seawater and Simulated Freshwater

Mar 4, 2011 - Hillary D. Easter,. ‡ and Bettina M. Voelker. Department of Chemistry and Geochemistry, Colorado School of Mines, Golden, Colorado 804...
0 downloads 0 Views 860KB Size
Download PDF
ARTICLE pubs.acs.org/est

Rapid Reaction of Nanomolar Mn(II) with Superoxide Radical in Seawater and Simulated Freshwater S. Paul Hansard,† Hillary D. Easter,‡ and Bettina M. Voelker Department of Chemistry and Geochemistry, Colorado School of Mines, Golden, Colorado 80401, United States

bS Supporting Information ABSTRACT: Superoxide radical (O2-) has been proposed to be an important participant in oxidation-reduction reactions of metal ions in natural waters. Here, we studied the reaction of nanomolar Mn(II) with O2- in seawater and simulated freshwater, using chemiluminescence detection of O2- to quantify the effect of Mn(II) on the decay kinetics of O2-. With 3-24 nM added [Mn(II)] and 99.5%) and 1 mg/L Suwannee River fulvic acid (SRFA) obtained from the International Humic Substances Society (IHSS reference 1R101F). The pH of these solutions was 8.2 before addition of the alkaline O2- stock solution and 8.6 afterward. Simulated freshwater was used instead of actual water samples as a convenient way to ensure reproducible experimental conditions. Long superoxide lifetimes (at least several minutes) were observed in these solutions, indicating low concentrations of superoxide-reactive metals. Experiments were performed in the presence of SRFA because of the expectation that dissolved organic matter would influence Mn reaction rates. Photogenerated O2- stock solutions were prepared by the general techniques of McDowell et al.,9 with modifications detailed in Hansard et al.8 O2- stock solutions contained diethylenetriaminetetraacetic acid (DTPA) to prevent metalcatalyzed O2- decay. Concentrations of DTPA were minimized to prevent carryover into the experimental solutions, which could potentially affect the reaction kinetics of Mn(II) with O2- (see Supporting Information for details). Decay Experiments. To initiate a decay experiment, a continuously stirred reaction solution (either a seawater sample or simulated freshwater) was spiked with O2- working stock. Reaction solution was continuously pumped from the reaction vessel into the detection system to measure [O2-] versus time. All experimental solutions and reagents were pre-equilibrated to room temperature before the experiments were performed. Two types of Mn addition experiments were performed: either Mn(II) was added several minutes before O2-, or O2- was added first and Mn(II) was added 49-100 s later. All of the simulated freshwater experiments were performed with the latter strategy, primarily because observation of slow initial O2- decay (before addition of Mn) was helpful for confirming that reaction solutions were not contaminated with O2--reactive compounds during solution preparation. However, because of naturally present O2-reactive compounds in the seawater samples, O2- decayed too fast to apply this strategy (with the exception of GoA2), and the former strategy was used instead. Since inorganic complexation reactions affecting Mn(II) speciation are extremely fast, and organic complexation is usually assumed to be insignificant,10,11 Mn(II) reactivity with O2- should not be affected by which strategy was used. The analytical system for seawater experiments was configured with a 17 s delay between removal of the sample from the reaction solution and measurement; thus, the first 17 s of data

ARTICLE

collected after addition of O2- (and of Mn, if Mn was added after O2-) did not accurately reflect conditions in the reaction solution and were discarded. For the freshwater experiments, a slightly different configuration was used and the delay was only 8 s. [O2-] Quantification. Quantification of [O2-] was achieved by use of the chemiluminescence-generating reaction of O2with methyl Cypridina luciferin analogue (MCLA). Briefly, sample and MCLA reagent were mixed in a transparent flow cell fixed to the face of a photon-counting photomultiplier tube (PMT). The PMT was programmed to collect 20 consecutive counts (0.050 s counting period) at 3 s intervals. The average of the 20 counts, the standard deviation, and the time elapsed after spiking of the sample were digitally recorded until decay was no longer observed and the baseline appeared to be stable. The analytical methods are described elsewhere in more detail.8,12 [O2-] detection limits, defined as 3 the standard deviation of the baseline signal (converted to concentration), were typically on the order of 0.010 nM. Two slightly different approaches were taken to convert measured chemiluminescence signal to superoxide concentration, [O2-], depending on whether Mn was added before O2- or vice versa. In the former case, O2- decay curves obtained in the absence of added Mn were used to obtain the sensitivity. To calculate sensitivity, the analytical response as a function of time, Rt, was fitted to Rt ¼ RBL þ Rt ¼ 0 expð - kdecay tÞ

ð1Þ

where the fitting parameters RBL, Rt=0, and kdecay represent the baseline response, the baseline-corrected response at t = 0 (R0 RBL), and the pseudo-first-order decay rate constant, respectively. Best-fit values of RBL, Rt=0, and kdecay were obtained using Excel’s Solver function to minimize the sum of the squares of the differences between measured and calculated Rt. Fits to eq 1 were excellent (R2 > 0.999) for all decay data from experiments with no added Mn. Sensitivity was then calculated as Rt=0 divided by the nominal initial [O2-]. Typically, sensitivity was obtained from 2-3 decay curves and averaged. This average sensitivity was then used to convert signal to concentration for all of the decay curves, including those with added Mn, obtained for a given water sample on a given day. Separate baseline corrections were performed on each decay curve by subtracting the average of the final five points of the decay curve from the measured signal. Final baselines for the decay curves used to calculate sensitivity differed from the fitted RBL values by a signal equivalent to less than 0.003 nM, similar to the standard deviation of the baseline signal, in all cases. For solutions spiked first with O2- and then with Mn, sensitivity was calculated as described above, except that only the part of the data collected before the time of Mn addition was fitted to eq 1, and RBL was set equal to the average of the final five data points collected from the run, several minutes after addition of Mn. A new sensitivity was measured for each decay curve. For some of the data sets presented below, including all of the freshwater data, observed first-order rate coefficients of O2- are discussed. These coefficients were determined directly from the raw data by use of eq 1 as described in the previous paragraph. For the GoA2 data set, two kdecay values were obtained from each data set. The data collected between the time of O2- addition and the time of Mn addition were used to obtain kdecay for 0 nM added Mn. The data collected after Mn addition were then used to obtain kdecay for that concentration of added Mn. 2812

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817

Environmental Science & Technology

ARTICLE

Figure 1. Representative plots of the GoA1 data set, showing kinetics of [O2-] decay as a function of added Mn(II) (added before spiking with O2-). [O2-]0 = 0.61-0.65 nM. (—) “Full model” fits; ( 3 3 3 ) “simple model” fits. (Inset) Example of the deviation of model fits from data at small [O2-].

’ RESULTS AND DISCUSSION Experiments with Excess Mn(II) in Seawater Systems. When an excess of Mn(II) was added to a seawater sample (GoA1), the rate of decay of O2- was approximately pseudo-first order and increased with increasing [Mn(II)] (Mn additions were 0, 0, 3, 6, 6, 9, 12, 18, and 24 nM; representative plots are shown in Figure 1). Good fits to these data were obtained from a model considering only reactions R1/R-1 and R2, and an additional pseudofirst-order O2- decay process, R4, accounting for the decay that is observed in the absence of added Mn (solid lines in Figure 1; Table 1): - k4

O2 sf products

ðR4Þ

We used Gepasi13 to determine best-fit values of the rate constants of all reactions except R1. We assigned a literature-based value to k1, the rate constant of reaction R1. In NaClO4 solutions, with Mn(II) present mostly as the aquo complex, measured values of this parameter range from 7  107 M-1 3 s-1 at ionic strength 0.5 M to 1.5  108 M-1 3 s-1 at ionic strength 0.04 M.1,4 A combination of ionic strength effects and possibly complexation by SO42- decreases the value to 5.4  107 M-1 3 s-1 in 0.1 M sodium sulfate.3 Because of the sulfate content (0.03 M) and high ionic strength of seawater (0.7 M), we estimated the value of k1 at 5  107 M-1 3 s-1. This value implies that at [Mn(II)] > 3 nM, the reversible complex formation reaction R1/R-1 should approach pseudoequilibrium before our system begins to collect data, so that an actual measurement of the rate constant k1 was not possible. The kinetics of MnO2þ reactions are known to be strongly dependent on the ligands present.2,14 Since we have no information regarding which ligands in our experimental solutions complex MnO2þ, and how this affects MnO2þ reaction rates, we treated k-1 and k2 as unconstrained fitting parameters. We also included an uncertainty for [O2-]0 of (10% of the nominal value in the model, a reasonable estimate of our experimental limitations. Uncatalyzed dismutation of O2- was insignificant at the pH and [O2-] examined,15 and we did not include this reaction in the model. The success of our model indicates that [O2-] decay kinetics are not greatly affected by any accumulating products of the reaction, especially Mn(III). However, in the experiments with added Mn,

the model systematically underpredicted measured [O2-] once [O2-] decayed to less than about 0.02 nM (e.g., 2 in Figure 1). This deviation might suggest that the data in the latter parts of the experiments are affected by a source of O2- not considered in the model, such as oxidation of H2O2 by Mn(III) (the back-reaction of R2). However, a rigorous kinetic analysis of this possibility, or others, is not feasible with our data, because we have observed baseline drifts of similar magnitude as the deviations at [O2-] < 0.02 nM. Furthermore, we did not measure [H2O2] in the experimental solutions (which could have differed slightly from experiment to experiment due to contamination from air), and we do not know the kinetics of further reactions of the initially formed Mn(III) species [e.g., to form complexes and Mn(III/IV) colloids and solids that may have different reactivities with H2O2]. To ensure that our uncertainties regarding the system’s behavior at low [O2-] did not have a great effect on the model presented in Table 1, we repeated the Gepasi modeling using only data points with [O2-] above 0.05 nM. The kinetic parameters obtained this way differed from the original parameters by less than 10% (Table 1). Because reaction R4 may include loss of O2- due to reaction with ambient Mn(II) present in the water sample, we also repeated the Gepasi modeling with the assumption that 1 nM or 2 nM Mn(II) is initially present, consistent with published concentrations for our study area.16 Other than the expected decrease in the fitted rate constant of reaction R4, these assumptions had little effect (at most 11%) on the kinetic parameters obtained (Table 1). Our model predicts fast initial consumption of O2- via complexation with Mn(II) (reaction R1) until [MnO2þ] reaches a steady-state value given by ½MnO2 þ ss ¼

k1 ½MnðIIÞ½O2 -  k-1 þ k2

ð2Þ

Once this steady state has been reached, the rate-limiting step is the rate of Mn oxidation via reaction R2, given by k2[MnO2þ]ss, and the overall rate of consumption of superoxide is then given by -

d½O2 -  ¼ k2 ½MnO2 þ ss þ k4 ½O2 -  dt ¼

k2 k1 ½MnðIIÞ½O2 -  þ k4 ½O2 -  k-1 þ k2

ð3Þ

where k4 (s-1) represents the pseudo-first-order loss of O2- observed in the absence of Mn(II). Equation 3 implies that if [MnO2þ]ss remains small compared to [O2-], so that [MnO2]ss remains proportional to [O2-], the overall rate of the reaction of O2- with Mn(II) to form H2O2 and Mn(III) should be well described by an “effective” second-order rate constant keff (M-1 3 s-1), given by k2 k1 ð4Þ keff ¼ k-1 þ k2 For the model parameters shown in Table 1, keff is equal to 6.5  106 M-1 3 s-1 and [MnO2þ]ss remains small under the conditions of our experiments. For example, for [Mn(II)] = 10 nM, the ratio of [MnO2þ]ss to [O2-] calculated from eq 2 is 0.145. Thus, we tested whether the kinetic behavior of this system could also be modeled by neglecting [MnO2þ] and using only keff and k4 as fitting parameters [essentially assuming that oxidation of Mn(II) (reaction R1 plus R2) takes place in a single step]. As before, we constrained fitted [O2-]0 values to within 10% of the nominal value. The resulting “simple” model fits were almost indistinguishable from those of the full model (dotted lines in 2813

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817

Environmental Science & Technology

ARTICLE

Table 1. Reactions and Rate Constants Used for the Kinetic Model of the GoA1 Data Set rate constants (results of Gepasi fits) all data

data >0.05 nM

initial [Mn] = 1 nM

initial [Mn] = 2 nM

5  107 M-1 3 s-1

Full Model R1: Mn2þ þ O2- f MnO2þ R-1: MnO2þ f Mn2þ þ O22H þ R2: MnO2þ s Mn3þ þ H2O2 R4: O2- f products

f

R4:

5  107 M-1 3 s-1

5  107 M-1 3 s-1

3.00 s

2.96 s

2.89 s

-1

-1

-1

0.446 s

0.483 s

-1

-1

0.0167 s

0.0169 s

6.1  106 M-1 3 s-1

0.419 s

-1 -1

0.0115 s

3.01 s-1

0.401 s-1 0.0068 s-1

6.4  106 M-1 3 s-1

-1

f products

0.0169 s-1

0.0167 s

Figure 1, Table 1). We conclude that our determination of the value of keff is robust but that the values of rate constants k-1 and k2 are much less well-constrained by our data than the uncertainties given by the Gepasi output (which were less than 1.5%). An alternative kinetic model leading to catalyzed dismutation of O2- with no net oxidation of Mn(II) (reaction sequence R1 followed by R3a or R3b) was not able to fit these data with k1 fixed at 5  107 M-1 3 s-1. This is expected, because unless reaction R1 is the rate-limiting step and reaction R-1 is insignificant, these reaction sequences will produce second-order, not first-order, O2- decays. A nearly 10-fold lower value of k1 would be required to fit our data with the reaction sequence R1 plus R3a or R3b. We consider this alternative unlikely, since this would require that the speciation (and hence reactivity) of Mn(II) is significantly altered by the seawater matrix. We conducted similar measurements, but with less extensive series of Mn additions, on three other seawater samples (GoA2, Mn additions 0/4.1 and 0/7.9 nM, with Mn added after O2-; GoA3, Mn additions 0, 0, 3.0, and 6.0 nM; and GoA4, Mn additions 0, 3.0, 6.0, and 12.0 nM). Using the simple model to analyze these data sets, we obtained good fits (R2 g 0.986) with values of keff of 1.1  107 M-1 3 s-1 (GoA2), 9.8  106 M-1 3 s-1 (GoA3), and 6.9  106 M-1 3 s-1 (GoA4). Experiments with Excess Mn(II) in Simulated Freshwater. In the simulated freshwater experiments with Mn(II) in excess of O2-, observed kdecay values were a linear function of added [Mn(II)] (Figure 2). Data sets were well described by first-order kinetics: R2 values of exponential fits were 0.965-0.998 except for the highest Mn(II) addition, 30 nM (R2 = 0.916). The linear relationship between added Mn(II) and kdecay is expected from eq 3, which, if it is assumed that [Mn(II)] remains approximately constant and that [MnO2þ]ss , [O2-], simplifies to the following: -

-1

-1

Simple Model

2H þ

Reff: Mn2þ þ O2- sf Mn3þ þ H2O2 O2-

5  107 M-1 3 s-1

d½O2 -  ¼ keff ½MnðIIÞ½O2 -  þ k4 ½O2 -  dt ¼ kdecay ½O2 - 

ð5Þ

Because Mn was added after O2- in these experiments, we have a direct measurement of [O2-]0 at t = 0 (the time of Mn addition). The exponential fits passed through these [O2-]0 values, indicating that there was no rapid initial decrease in O2- upon addition of Mn, which means that [MnO2þ] remained small compared to [O2-]. In addition, added [Mn(II)] in these experiments was at least a 10-fold excess over [O2-]0, so the assumption that [Mn(II)] remains approximately constant is also valid.

Figure 2. Pseudo-first-order O2- decay rate coefficients (kdecay) as a function of added Mn(II) in simulated freshwater. [O2-]0 = 0.4-0.5 nM.

The value of keff, calculated from the slope of the line in Figure 2, was 1.6  106 M-1 3 s-1, somewhat lower than the keff values obtained in the seawater samples. This difference is probably mostly attributable to a difference in the extent and nature of complexation of MnO2þ. This value is much higher than the keff value of 5.7  103 M-1 3 s-1 derived from the rate constants in the model of Nico et al.7 of Mn(II) photo-oxidation in the presence of Aldrich humic acid in pH 8.1 borate buffer. While the differences in solution composition might account for the difference between the keff values, we also note that the system of Nico et al. was much more complicated than ours. If they did not correctly account for all side reactions in their photochemical experiments, for example, reactions with Fe, which is present in large quantities in Aldrich humic acid,17 their model could have extracted erroneous rate constants. Experiments with Excess [O2-] in Seawater. While the seawater experiments discussed above, with subnanomolar [O2-] and nanomolar [Mn(II)], are more representative of marine conditions, we conducted several experiments at higher [O2-] to determine whether Mn-catalyzed decay of O2- could also occur. Addition of Mn(II) at concentrations less than those of [O2-]0 significantly accelerated the rate of [O2-] decay in the GoA2, GoA3, and GoA4 water samples (Figures 3 and 4). All of the decay curves contributing to Figure 3 were well described by pseudo-first-order kinetics (R2 values of exponential fits > 0.998), indicating that second-order processes such as reactions R3a and R3b are not contributing greatly to decay. The reaction sequence in Table 1 does not predict Mn-catalyzed decay in the presence of excess O2- because the Mn(II) should be consumed; thus, other reactions must be occurring. Two possibilities we considered initially are (i) MnO2þ is consumed in a process that regenerates Mn(II), such that no Mn(III) is produced, or 2814

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817

Environmental Science & Technology

ARTICLE

Figure 3. Effect of added Mn(II) on O2- pseudo-first-order decay coefficients (kdecay) in three seawater samples with [O2-]0 in excess over added Mn(II) (2-9 nM for GoA2, 3-9 nM for GoA3, and approximately 20 nM for GoA4). ( 3 3 3 ) Linear regressions for each water sample; (—) expected behavior if [Mn(II)] ≈ [Mn(II)]0. () Model simulations of GoA4 with k5 = 7  107 M-1 3 s-1.

(ii) the Mn(III) that is produced is rapidly reduced back to Mn(II) by solution constituents such as dissolved organic compounds. For either of these possibilities, [Mn(II)] would remain equal to (or slightly less than) its initial value. In that case, the dependence of kdecay as a function of added [Mn(II)] should have a slope equal to (or slightly less than) keff (eq 5). However, the slopes of the linear regressions of the data points (dotted lines in Figure 3) are significantly greater than keff (solid lines) in the GoA3 and GoA4 samples (one-tailed t test, 95% confidence level, if it is assumed, as indicated by the Gepasi output, that the uncertainty in keff is negligible compared to the uncertainty in the observed slopes). Thus, neither of these possibilities offers a good explanation for the GoA3 and GoA4 results. The greater effect of Mn(II) addition on O2- decay rates when O2- is in excess, compared to when Mn is in excess, can only be explained if two conditions are met: (i) the product of reaction R2 reacts with O2- at a faster rate than Mn(II) does and (ii) reaction of that product is sufficiently slow that the reaction is negligible when Mn(II) is in excess. We propose that reduction of Mn(III) by O2- also occurs: k5

MnðIIIÞ þ O2 - sf MnðIIÞ þ O2

ðR5Þ

Reaction R5 has been shown to be thermodynamically feasible,18 though its rate constant in seawater is unknown. If reaction R5 with rate constant k5 of 7  107 M-1 3 s-1 is added to the simple model of the GoA4 data, the effect of added Mn in the presence of excess O2can be simulated ( in Figure 3). At the same time, this reaction makes minimal difference to the simulations of data sets with added Mn(II) in excess over O2-: the difference between calculated [O2-] when reaction R5 is included and excluded was less than 5 nM) in regions receiving dust or coastal inputs, such as many parts of the North Atlantic;22 however, there are at present no published accounts of superoxide decay rates from these areas. Although we were unable to measure Mn(III) formation directly (to our knowledge, no method is capable of doing so at subnanomolar concentrations), the kinetic interpretations of our O2- decay data strongly suggest that a fast redox cycle of dissolved Mn(II/III) will occur in the presence of a continuous source of O2-. For example, with keff and k5 values of 6.9  106 and 7  107 M-1 3 s-1, respectively (Table 1), Mn(II) oxidation (Reff) and reduction R5 proceed at the same rates when 11% of dissolved Mn is present as Mn(III). Thus, a continuous source of superoxide could maintain a significant percentage of the dissolved Mn as Mn(III). Using the same rate constants keff and k5 and an estimate of [O2-] of 10-10 M, based on recently published values measured in the Gulf of Alaska,8 the Costa Rica Dome,23 and the Great Barrier Reef,24 we calculate that such a superoxide-driven Mn(II/III) cycle will turn over at a rate of ∼7  10-4 s-1 or ∼2.5 h-1, much faster than the redox cycle of oxidative precipitation of 2815

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817

Environmental Science & Technology

ARTICLE

Table 2. Reactions and Rate Constants Used for the Kinetic Model of the GoA2 Data Set fitted parameter value 2H þ

Model 1 1.08  107 M-1 3 s-1

Reff: Mn(II) þ O2- sf Mn(III) þ H2O2 R4:

O2-

0.013-0.017 s-1 (low O2-); 0.022-0.024 s-1 (high O2-)a

f products

R5: Mn(III) þ

O2-

1.66  107 M-1 3 s-1

f Mn(II) þ O2 2H þ

Model 2 1.15  107 M-1 3 s-1

Reff: Mn(II) þ O2- sf Mn(III) þ H2O2 R4:

O2-

0.013-0.017 s-1 (low O2-); 0.022-0.024 s-1 (high O2-)a

f products

1.72  104 s-1

R6: Mn(III) f Mn(II)

Gepasi fits were performed only on the data obtained after Mn addition. Data obtained before Mn addition were fitted by use of eq 1, and the kdecay values thus obtained for each data set were used as k4 in the Gepasi fitting. The kdecay values were slightly higher at higher [O2-]0, but the difference was small compared to the large effect of [O2-]0 on kdecay observed and discussed by Rose et al.24 a

Mn(II)25-27 and subsequent photoreduction of manganese oxides.27,28 Thus, even if much of the Mn(III) produced by O2- may be rereduced by O2-, a continuous supply of aqueous Mn(III) could be maintained, which could then participate in other reactions. For example, aqueous Mn(III) is known to be an effective oxidant, capable of oxidizing recalcitrant organic compounds such as lignins.29 Furthermore, in the presence of ligands that stabilize Mn(III), such as siderophores, Mn(II) oxidation by O2- could represent an important source of complexed Mn(III). While it is usually assumed that all dissolved Mn in marine waters is Mn(II), Trouwborst et al.30 have measured 0.2 μm filterable Mn(III) at concentrations up to 100% of the total dissolved Mn in suboxic regions of Black Sea and Chesapeake Bay. Competition by Mn(III) for ligands associated with biological Fe uptake has been suggested to be an important ecological consequence of Mn(II) oxidation.31,32 Our results also show that the reaction of nanomolar Mn(II) with O2- is fast in simulated freshwater environments at pH 8.6. However, because keff could depend strongly on the concentrations and type of ligands present, more experiments spanning the range of pH values typical in freshwater environments, and in matrices containing ligands other than carbonate and fulvic acid, need to be performed.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional text with experimental details including bottle-cleaning protocols, preparations of solutions, and auxiliary data on the seawater samples; and five tables and four figures showing complete data sets with their model fits and calculations of Mn(II). This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]; phone: (303) 273-3152; fax: (303) 273-3629. Present Addresses †

Florida Geological Survey, Tallahassee, FL 32304. E-mail: paul. [email protected].fl.us. ‡ Admiralty Environmental, 431 N. Franklin, Suite 301, Juneau, AK 99801.

’ ACKNOWLEDGMENT We thank Ken Bruland and the captain and crew of the R.V. Thomas G. Thompson for the opportunity to participate in the GoA cruise. Whitney King (Colby College) was generous with his advice regarding O2- stock solutions and the Fe-Lume system. This work was supported by NSF Grant OCE-0551715 to B.M.V. ’ REFERENCES (1) Pick-Kaplan, M.; Rabani, J. Pulse radiolytic studies of aqueous Mn(ClO4)2 solutions. J. Phys. Chem. 1976, 80 (17), 1840–1843. (2) Cabelli, D. E.; Bielski, B. H. J. Pulse-radiolysis study of the kinetics and mechanisms of the reactions between manganese(II) complexes and HO2/O2- radicals. 2. The phosphate complex and an overview. J. Phys. Chem. 1984, 88 (25), 6291–6294. (3) Cabelli, D. E.; Bielski, B. H. J. Pulse-radiolysis study of the kinetics and mechanisms of the reactions between manganese(II) complexes and HO2/O2- radicals. 1. Sulfate, formate, and pyrophosphate complexes. J. Phys. Chem. 1984, 88 (14), 3111–3115. (4) Jacobsen, F.; Holcman, J.; Sehested, K. Manganese(II)-superoxide complex in aqueous solution. J. Phys. Chem. A 1997, 101 (7), 1324–1328. (5) Barnese, K.; Gralla, E. B.; Cabelli, D. E.; Valentine, J. S. Manganous phosphate acts as a superoxide dismutase. J. Am. Chem. Soc. 2008, 130 (14), 4604–4606. (6) Learman, D. R.; Voelker, B. M.; Vazquez-Rodriguez, A. I.; Hansel, C. M. Formation of manganese oxides by bacterially produced superoxide. Nat. Geosci. 2011, 4, 95–98. (7) Nico, P. S.; Anastasio, C.; Zasoski, R. J. Rapid photo-oxidation of Mn(II) mediated by humic substances. Geochim. Cosmochim. Acta 2002, 66 (23), 4047–4056. (8) Hansard, S. P.; Vermilyea, A. W.; Voelker, B. M. Measurements of superoxide radical concentration and decay kinetics in the Gulf of Alaska. Deep-Sea Res., Part I 2010, 57 (9), 1111–1119. (9) McDowell, M. S.; Bakac, A.; Espenson, J. H. A convenient route to superoxide ion in aqueous solution. Inorg. Chem. 1983, 22, 847–848. (10) Roitz, J. S.; Bruland, K. W. Determination of dissolved manganese(II) in coastal and estuarine waters by differential pulse cathodic stripping voltammetry. Anal. Chim. Acta 1997, 344 (3), 175–180. (11) Gimpel, J.; Zhang, H.; Davison, W.; Edwards, A. C. In situ trace metal speciation in lake surface waters using DGT, dialysis, and filtration. Environ. Sci. Technol. 2003, 37 (1), 138–146. (12) Rose, A. L.; Moffett, J. W.; Waite, T. D. Determination of superoxide in seawater using 2-methyl-6-(4-methoxyphenyl)-3,7dihydroimidazo[1,2-a]pyrazin-3(7H)-one chemiluminescence. Anal. Chem. 2008, 80 (4), 1215–1227. (13) Mendes, P. Gepasi - A software package for modeling the dynamics, steady-states and control of biochemical and other systems. Comput. Appl. Biosci. 1993, 9 (5), 563–571. 2816

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817

Environmental Science & Technology

ARTICLE

(14) Barnese, K.; Gralla, E. B.; Cabelli, D. E.; Valentine, J. S. Manganous phosphate acts as a superoxide dismutase. J. Am. Chem. Soc. 2008, 130 (14), 4604–4606. (15) Zafiriou, O. C. Chemistry of superoxide ion-radical (O2-) in seawater. I. pK*asw (HOO) and uncatalyzed dismutation kinetics studied by pulse radiolysis. Mar. Chem. 1990, 30, 31–43. (16) Minakawa, M.; Noriki, S.; Tsunogai, S. Manganese in the Bering sea and the northern North Pacific Ocean. Geochem. J. 1998, 32 (5), 315–329. (17) Hering, J. G.; Morel, F. M. M. Humic acid complexation of calcium and copper. Environ. Sci. Technol. 1988, 22, 1234–1237. (18) Luther, G. W. The role of one- and two-electron transfer reactions in forming thermodynamically unstable intermediates as barriers in multi-electron redox reactions. Aquat. Geochem. 2010, 16 (3), 395–420. (19) Landing, W. M.; Bruland, K. W. Manganese in the North Pacific. Earth Planet. Sci. Let.s 1980, 49 (1), 45–56. (20) Heller, M. I.; Croot, P. L. Superoxide decay kinetics in the Southern Ocean. Environ. Sci. Technol. 2010, 44 (1), 191–196. (21) Sedwick, P. N.; Edwards, P. R.; Mackey, D. J.; Griffiths, F. B.; Parslow, J. S. Iron and manganese in surface waters of the Australian subantarctic region. Deep-Sea Res., Part I 1997, 44 (7), 1239–1253. (22) Shiller, A. M. Manganese in surface waters of the Atlantic Ocean. Geophys. Res. Lett. 1997, 24 (12), 1495–1498. (23) Rose, A. L.; Webb, E. A.; Waite, T. D.; Moffett, J. W. Measurement and implications of nonphotochemically generated superoxide in the equatorial Pacific Ocean. Environ. Sci. Technol. 2008, 42 (7), 2387– 2393. (24) Rose, A. L.; Godrant, A.; Furnas, M.; Waite, T. D. Dynamics of nonphotochemical superoxide production and decay in the Great Barrier Reef lagoon. Limnol. Oceanogr. 2010, 55 (4), 1521–1536. (25) Sunda, W. G.; Huntsman, S. A. Effect of sunlight on redox cycles of manganese in the southwestern Sargasso Sea. Deep-Sea Res., Part A 1988, 35 (8), 1297–1317. (26) Moffett, J. W.; Ho, J. Oxidation of cobalt and manganese in seawater via a common microbially catalyzed pathway. Geochim. Cosmochim. Acta 1996, 60 (18), 3415–3424. (27) Waite, T. D.; Szymczak, R. Manganese dynamics in surface waters of the Eastern Caribbean. J. Geophys. Res.: Oceans 1993, 98 (C2), 2361–2369. (28) Sunda, W. G.; Huntsman, S. A.; Harvey, G. R. Photo-reduction of manganese oxides in seawater and its geochemical and biological implications. Nature 1983, 301 (5897), 234–236. (29) Hofrichter, M. Review: lignin conversion by manganese peroxidase (MnP). Enzyme Microb. Technol. 2002, 30 (4), 454–466. (30) Trouwborst, R. E.; Clement, B. G.; Tebo, B. M.; Glazer, B. T.; Luther, G. W. Soluble Mn(III) in suboxic zones. Science 2006, 313 (5795), 1955–1957. (31) Parker, D. L.; Sposito, G.; Tebo, B. M. Manganese(III) binding to a pyoverdine siderophore produced by a manganese(II)-oxidizing bacterium. Geochim. Cosmochim. Acta 2004, 68 (23), 4809–4820. (32) Duckworth, O. W.; Bargar, J. R.; Sposito, G. Coupled biogeochemical cycling of iron and manganese as mediated by microbial siderophores. Biometals 2009, 22 (4), 605–613.

2817

dx.doi.org/10.1021/es104014s |Environ. Sci. Technol. 2011, 45, 2811–2817