Effects of Molecular Composition of Natural Organic Matter on Ferric

Mar 17, 2014 - and T. D. Waite. ‡. †. Department of Civil ...... (13) Witter, A. E.; Hutchins, D. A.; Butler, A.; Luther, G. W.. Determination of ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Effects of Molecular Composition of Natural Organic Matter on Ferric Iron Complexation at Circumneutral pH Manabu Fujii,*,† Akira Imaoka,† Chihiro Yoshimura,† and T. D. Waite‡ †

Department of Civil Engineering, Tokyo Institute of Technology, 2-12-1-M1-4 Ookayama, Tokyo 152-8552, Japan School of Civil and Environmental Engineering, The University of New South Wales, Sydney, New South Wales 2052, Australia



S Supporting Information *

ABSTRACT: Thermodynamic and kinetic parameters for ferric iron (Fe[III]) complexation by well-characterized humic substances (HS) from various origins were determined by a competitive ligand method with 5-sulfosalicylic acid at circumneutral pH (6.0−8.0) and an ionic strength of ∼0.06 M. The measured Fe binding properties including conditional stability constants and complexation capacities ranged over more than 2 orders of magnitude, depending on the origin and the particular operationally defined fraction of HS examined. Statistical comparison of the complexation parameters to a range of chemical properties of the HS indicated a strong positive correlation between Fe(III) complexation capacity and aromatic carbon content in the HS at all pHs examined. In contrast, the complexation capacity was determined to be up to a few orders of magnitude smaller than the concentration of carboxylic and phenolic groups present. Therefore, specific functional groups including those resident in the proximity of aromatic structures within the HS are likely preferable for Fe(III) coordination under the conditions examined. Overall, our results suggest that the concentration of dissolved Fe(III) complexes in natural waters is substantially influenced by variation in HS characteristics in addition to other well-known factors such as HS concentration and nature and concentration of competing cations present.



INTRODUCTION Biogeochemistry of iron (Fe) in natural waters is of great interest to many researchers, particularly oceanographers and environmental chemists, given the accumulated evidence that oceanic primary productivity is limited in one-third of the world’s ocean1 and in some coastal waters2 primarily because of the low availability of Fe. Any change in Fe availability in such waters is anticipated to significantly affect the biogeochemical cycle of other elements, including C, N, and Si.3 In addition to its importance in influencing marine phytoplankton growth and global elemental cycles, Fe availability regulates the biosynthesis of secondary metabolites including cyanotoxins in freshwater cyanobacteria.4 Therefore, Fe is a potentially important trace metal with regard to the proper management of water supply reservoirs, particularly those subject to growth of toxic algae. Over the last three decades, the chemical speciation of Fe in surface waters and its relevance to environmental parameters have been primarily investigated for the purpose of improving our understanding of the biogeochemical functioning of Fe, including its availability to primary producers and the effect of Fe chemical transformation on the speciation of trace metals (e.g., Cu, Zn, and Ni).5,6 Because Fe complexes with naturally occurring organic matter are generally considered to be impermeable to cellular membranes,7 with the exception of specific molecules such as siderophore-type ligands, unchelated Fe (Fe′, defined as the sum of the concentration of all dissolved © 2014 American Chemical Society

inorganic Fe species) is considered to be the most readily available form of Fe.7,8 The concentration of unchelated Fe is typically controlled by a range of dynamic processes including complexation, precipitation and oxidation−reduction with recent studies suggesting that both photochemical and nonphotochemical transformations of organically complexed Fe are critical to Fe bioavailability at circumneutral pH.7,9,10 It is widely accepted that nearly all dissolved Fe in surface seawater forms complexes with Fe-binding ligands and that unchelated Fe is buffered by the excess Fe-binding ligands present.11 Thus, the concentration of dissolved ferric iron (Fe[III]) in oxygenated surface seawater, in which the oxidized form of Fe is thermodynamically preferable, generally exceeds the solubility of inorganic Fe(III) species, which is extremely low at circumneutral pH (e.g., ∼7 × 10−11 M at pH 8).12 Oceanic Fe-binding ligands have very high stability constants for Fe(III) complexation, which are typically comparable with those of siderophores;11,13 however, structural and compositional information on marine Fe-binding ligands is limited.14 In freshwater and coastal seawater systems, natural organic matter (NOM) including humic substances15 (HS, a macromolecular assemblage Received: Revised: Accepted: Published: 4414

December 10, 2013 February 19, 2014 March 17, 2014 March 17, 2014 dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

DKK, Japan) with either 0.1−10 M hydrochloric acid (HCl, Kanto Chemical) or sodium hydroxide (NaOH, Kanto Chemical) added as required. Standard humic substances, including fulvic and humic acids (FA and HA), were purchased from the International Humic Substances Society (IHSS) and the Japanese Humic Substances Society (JHSS). The fundamental chemical properties of standard HS are listed in Table 1. Stock solutions of fulvic and humic acids at approximately 10 g·L−1 from seven and eight different origins, respectively, were prepared by dissolving in 0.5−1.0 mL NaOH solutions of pH 13, followed by pH adjustment to circumneutral values of 8.0 ± 0.1, 6.5 ± 0.1, or 6.0 ± 0.1. A stock solution of 0.5 mM Fe(III) was prepared by combining a 1.0 g-Fe·L−1 ferric nitrate solution (0.2 M HNO3, Kanto Chemical) with ultrapure Milli-Q water (MQ; Millipore, 18.2 MΩ·cm resistivity). Stock solutions of 93 mM 5-sulfosalicylic acid (SSA; Kanto Chemical) and 100 mM citric acid (Cit; Kanto Chemical) were prepared in MQ followed by pH adjustment to 8.0 ± 0.05, 6.5 ± 0.05, or 6.0 ± 0.05. For pH control in the competitive ligand experiments, 10 mM 4-(2-hydroxyethyl)-1piperazine-ethanesulfonic acid (HEPES; Sigma) adjusted to pH 8.0 ± 0.05 and 20 mM 2-morpholinoethanesulfonic acid (MES; Sigma) at pH 6.5 ± 0.05 and 6.0 ± 0.05 were prepared. All experiments were performed in a temperature-controlled room at 25 ± 1 °C. Absorptivities of Organic Ligands and Fe Complexes. Although SSA itself does not exhibit significant absorbance at visible wavelengths, its complex with Fe(III) (denoted here as FeSSA) absorbs strongly in the 400−500 nm wavelength region.26,28 The molar absorptivity of FeSSA was determined by measuring the absorbance spectra for the FeSSA complex with given Fe(III) concentrations at pH 6.0−8.0. To this end, 5−50 μL of 0.5 mM Fe(III) stock was added to mixtures of SSA and the appropriate pH buffers, and were prepared in 1 cm path length polystyrene spectrophotometer cuvettes to obtain final concentrations of 2.5−25 μM for Fe(III), 7.6−8.3 mM for SSA, 8.5− 8.9 mM for HEPES, and 16−17 mM for MES. After the cuvette lid was tightly closed, the sample was gently mixed by turning the cuvette upside down and was left for 2 h under dark conditions to equilibrate. Absorbance measurement was then performed using a UV/vis spectrophotometer (UV-1800, Shimadzu) with background absorbance of the cuvette containing SSA and the buffer mixture recorded prior to the addition of the Fe(III) stock. Maximum absorbance was observed at wavelengths of 425 nm for pH 8.0 and 460 nm for pH 6.0−6.5. Molar absorptivities were −1 −1 determined to be εFeSSA 425/460 = 4,300 ± 74 M ·cm for pH 6.0−8.0, which was comparable to that reported previously (εFeSSA 425 =4800 M−1.cm−1 for seawater at pH 8.1).26 The slight difference in values derived in this and previous studies may be associated with different methodologies used including media. The average molar absorptivity of the Fe(III) complex with HS (FeHS) was determined by a method similar to that employed for SSA. In this case, SSA was replaced by HS, and the HS concentrations were adjusted to 34−500 mg·L−1, which are the same values as those used in the competitive ligand experiment, as described subsequently. The absorbance measurement of FeHS complex at wavelengths of 425 nm for pH 8 and 460 nm for pH 6.0−6.5 yielded molar absorptivities of εFeL 425 =150−860 −1 −1 at pH 6.5, and M−1·cm−1 at pH 8, εFeL 460 =75−1400 M ·cm −1 −1 εFeL 460 =650−1300 M ·cm at pH 6.0 for a range of HS (SI Table S1). The mass absorptivity for HS was similarly determined by measuring the absorbance of 100 mg·L−1 HS solutions excluding

of diverse, relatively small components likely stabilized by hydrophobic interaction and hydrogen bonds16) and other biogenic ligands such as saccharides17 play a significant role in the behavior of Fe. Recent studies indicate that terrigenous HS accounts for a majority of Fe-binding ligands in some coastal seawaters (e.g., Thurso Bay18 and Irish Sea,19 U.K.) even though it has long been recognized that a substantial amount of riverinedissolved Fe can be removed in the estuarine mixing zone.20 In the same coastal waters, Fe solubility (5.4−573 nmol.L−1) is larger than typical dissolved Fe concentrations in the open ocean (average, median, 10th percentile, and 90th percentile were 0.69, 0.35, 0.10, and 1.6 nmol.L−1, respectively, for the data set [n = 173] in Tables A1−A4 provided by de Baar and de Jong21) and maintained at levels just below the Fe-binding capacity of HS (7.5−604 nmol.L−1).18,19 Metal binding by HS is considered to be mainly mediated by O-containing acidic functional groups such as carboxyls and phenolics and less often by N- and S-bearing ligands, which are considered to form stronger complexes with particular trace metals.15,22 Carboxyl groups clustered in the short-chain aliphatic region are likely dominant in NOM23 and may play an important role in multidentate complex formation with divalent metals (e.g., Ca and Cd) in conjunction with aromatic functional groups via inner- or outer-sphere coordination.24 In addition, recent studies have reported that the aromaticity of NOM as determined by UV absorbance is relevant to the association with transition metals, particularly with regard to the binding affinities to Cd, Cu, and Zn.25 Because HS has various origins with HS from different sources potentially exhibiting different chemical properties, variation of HS characteristics in addition to the possible range of HS concentrations may influence metal complexation, and ultimately, its bioavailability in natural waters. Although extensive research has been conducted during recent decades on the molecular composition and chemical properties of HS, including its association with trace metals and other divalent cations, 15,22,24,25 the particular moieties of HS responsible for Fe(III) complexation remain poorly understood even though the Fe-binding affinity and capacity of NOM are recognized to vary substantially depending on the origin of NOM.26 The handling of Fe(III) in NOM complexation experiments at circumneutral pH may be problematic in some cases, because Fe(III) rapidly precipitates at concentrations greater than subnanomolar. However, effective techniques for the determination of the kinetics and thermodynamics of Fe(III) complexation by NOM at circumneutral pH and nanoor micromolar Fe concentrations have been developed over the last ten years and include competitive ligand methods that use adsorptive stripping voltammetry11,13 and visible spectrophotometry.26,27 In this study, spectrophotometric measurement is used to determine thermodynamic constants, such as complexation capacity and stability constant, and kinetic constants, such as complex formation rate constants, for the Fe(III) complexation of 15 types of chemically well-defined standard HS at pH 6.0−8.0. The molecular composition of HS influencing Fe(III) complexation are then investigated by examining the statistical correlation between the complexation parameters and the chemical properties of the HS used.



MATERIALS AND METHODS General. Full descriptions of the procedures including cleaning of plastic ware, chemical preparation, and storage are provided in the Supporting Information (SI.1). pH adjustment was performed using a HM25R pH meter (GST-5731C, TOA 4415

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

f f f

Aromaticity was determined from aromatic carbon[%]/(aromatic carbon[%] + carbohydrate[%] + Fujitake et al. (2009).

53 d

f f

40 c 52 b

International Humic Substances Society. Watanabe et al. (1994). aliphatic carbon[%]). eMass ratio (%). fValues not determined.

FA

Fe(III) addition, which yielded εL425 =7.1 × 10−4 to 1.0 × 10−2 L·mg−1.cm−1 at pH 8, εL460 =3.4 × 10−4 to 8.2 × 10−3 L·mg−1.cm−1 at pH 6.5, and εL460 =6.3 × 10−4 to 7.7 × 10−3 L·mg−1·cm−1 at pH 6.0 (SI Table S1). Competitive Ligand Experiments. The Fe(III) complexation capacity and stability constants for various types of standard HS were determined by the competitive ligand method. SSA stock (0.1 mL of 93 mM) and appropriate amounts of HS stock were added to 1 mL of 10 mM HEPES (pH 8) or 20 mM MES (pH 6.0−6.5) in 1 cm spectrophotometer cuvettes to obtain concentrations of 34−500 mg·L−1 for HS and 7.6−8.3 mM for SSA. Absorbance of the HS and SSA mixtures was recorded at 425 nm for the sample at pH 8 and 460 nm for that at pH 6.5−6.0 as background signals (Ab). After Fe(III) stock was added to the mixture at a final concentration of 2.5−25 μM, the sample was gently mixed and equilibrated for 2 h under dark conditions. The sample absorbance (Af) was again measured. The pH change of the sample solution after the addition of stock solution was measured to be less than ±0.05 unit. The complex formation rate constants were determined in the HEPES buffer (pH 8) by using an SSA competition method identical to that described previously.27 Briefly, the FeSSA concentration was spectrophotometrically monitored shortly after the addition of Fe(III) stock at 2.5−25 μM into the HEPES buffer amended with 7.6−8.3 mM SSA and 34−500 mg·L−1 HS. As previously reported,27 at constant SSA and HS concentrations, the concentration of the FeSSA complex formed is a function of the two bimolecular rate constants for the competing reactions of Fe complexation by SSA and HS. To ensure the validity and reliability of the parameters obtained by the complexation experiment, the following analytical aspects were considered. According to baseline stability (5 × 10−6 abs·min−1) and noise level (5 × 10−5 abs) of visible light absorbance for the given spectrophotometric analysis, the analytical detection limit of absorbance was calculated to be 1.5 × 10−4 abs at maximum by assuming 3-fold of the background noise level, corresponding to an FeSSA concentration of 3.5 × 10−8 M. This value was well below the lowest concentration of FeSSA determined in this study (5.4 × 10−7 M). Factors such as ligand concentrations that prevent significant Fe(III) precipitation and the balance of HS and SSA concentrations to achieve effective competition between HS with SSA for Fe(III) complexation were also considered in the experimental setup. Equilibrium calculation of inorganic and organic Fe(III) species under the conditions of competitive ligand experiment are also provided in SI.2. A 2-h equilibration period was selected, because the goodness of fit of the thermodynamic model to the experimental complexation data was higher than that with longer incubations of 24 h to 5 days, as discussed in SI.3. Although it is recognized that HS are capable of reducing Fe(III) to Fe(II) at an appreciated rate even in the dark,29 most likely via the redox active quinone-like groups present in HS, the contribution of Fe(II) toward Fe complexation by HS is likely to be small under the conditions examined, as described in SI.4. Thus, Fe(III) was considered to be the dominant oxidation state for Fe. In addition, standard HS contained measurable amounts of “intrinsic” Fe with concentrations ranging from 0.07 to 32 nmol.mg−1, corresponding to 0.08−130% of externally added Fe. However, the Fe that was initially present was determined to react with SSA negligibly slowly (less than 10−7 M in 2h) with the FeSSA so-formed well below the lowest FeSSA concentration determined in the competitive ligand experiment (5.4 × 10−7 M).

a

9.08 10.9 1.36 1.07 7.72 9.81

f f f

5.88 6.31 7.53 7.33 7.50 2.18 1.96 1.53 1.49 1.66 3.70 4.35 6.00 5.85 5.84

f f f

6.76 5.90 6.16 6.23 6.55 1.96 1.09 1.08 1.47 1.72 4.81 4.81 5.08 4.76 4.83

0.60 0.68 0.65 0.72 0.57 0.62 0.71 0.66 0.47 0.50 0.48 0.40 0.63 0.47 0.46 21 24 25 23 29 26 14 19 27 22 34 18 22 30 20 31 50 47 58 38 42 33 37 24 22 31 12 37 27 17 20 10 9 5 18 14 23 22 16 22 19 8.6 19 16 24 29 16 19 14 15 18 30 22 33 35 18 61 22 26 39 0.0038 0.0028 0.0047 0.0045 0.0041 0.0025 0.0020 0.0018 0.0031 0.0033 0.0033 0.0216 0.0001 0.0007

0.019 0.061 0.056 0.017 0.019 0.023 0.073 0.063 0.012 0.011 0.011 0.106 0.014 0.034 0.041 0.62 0.58 0.62 0.53 0.68 0.60 0.44 0.54 0.62 0.62 0.71 0.37 0.86 0.93 0.37 0.60 0.44 0.50 0.37 0.61 0.53 0.52 0.50 0.60 0.62 0.65 0.45 0.76 0.89 0.47 0.97 0.75 0.81 0.69 0.89 0.88 1.2 0.93 0.98 0.99 0.91 1.2 0.88 0.96 1.3 0.54 0.44 0.71 0.76 0.58 0.36 0.29 0.26 0.44 0.46 0.46 3.0 0.01 0.08

1.2 4.1 3.7 1.2 1.2 1.5 4.5 4.0 0.72 0.67 0.68 6.5 0.77 1.7 2.3 42 34 37 31 43 39 37 37 42 43 45 31 48 51 36 53 58 56 64 53 55 53 55 52 52 52 52 48 43 56 HA

2S101H 1S102H 1S103H 1S104H 1R105H 1R107H DHA IHA 1S101F 2S101F 1R105F 1R109F DFA IFA BFA

Suwannee River IIa Elliott Soila Pahokee Peata Leonarditea Nordic Lakea Waskish Peata Dando Soilb Inogashira Soilb Suwannee River Ia Suwannee River IIa Nordic Lakea Pony Lakea Dando Soilb Inogashira Soilb Biwa Lakec

4.3 3.7 3.8 3.7 4.0 4.0 5.3 4.3 4.3 4.4 4.0 5.4 3.5 3.5 6.1

total acidity phenolic groups carboxyl groups carbonyl aromatic carbohydrate N/C O/H H/C S N O C

H

Article

origin code fract ion

elemental composition (%)e

Table 1. Chemical Properties of Humic Substances Used in This Study

O/C

elemental ratio

S/C

alip hatic

carbon species from 13C NMR (%)

aromat icityd

acidic functional groups (meq/g)

Environmental Science & Technology

4416

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

Therefore, the Fe that was initially present was not considered to be an important pool with regard to complexation of Fe by HS (SI.5). It appears that the effect of HS concentration on the binding of Fe(III) by HS is relatively minor, as discussed in SI.6, suggesting that ternary complexes of the form SSA-Fe-HS are unimportant. Finally, HEPES is unlikely to affect the Fe(III) complexation by SSA under the concentrations examined in this work (SI.7) although HEPES has been reported to be a strong metal binding agent (e.g., log K = ∼7 for Cu2+).30 Determination of Concentrations for Fe Species. The intrinsic absorbance (Aint) solely attributed to the formation of FeSSA and FeHS complexes was calculated by the following procedure: (i) the background signal (Ab) was subtracted from the absorbance after equilibration (Af); (ii) HS absorbance reduction (Ared) due to the dilution effect by the Fe(III) stock addition was calculated by using HS mass absorptivities (εL425 and εL460); and (iii) Aint was calculated by adding Ared from the second step to Af − Ab from the first step; that is, Aint = Af−Ab + Ared. Given that the relationship between absorbance and concentration for the Fe(III) complex follows the Lambert− Beer law, the intrinsic absorbance defined here can be described as follows: A int = εFeSSA [FeSSA] + εFeL[FeL]

The mass balance equation for Fe can be written as follows: [Fe]T = [Fe′] + [FeSSA] + [FeL]

where [Fe]T indicates the total Fe concentration added to the system in the competitive experiment. By solving eqs 1, 3, and 6 with the approximation of [SSA′] being equal to total SSA concentration ([SSA]T), which is valid given that [SSA]T = ∼8 mM ≫ 25 μM ≥ [FeSSA], the concentration of each Fe species can be described as follows: FeL [FeSSA] = (A int − ε425/460 [Fe]T )

⎧ ⎫−1 ⎛ ⎞⎪ ⎪ 1 FeSSA FeL ⎜ ⎟ × ⎨ε425/460 − ε425/460⎜1 + cond ⎟⎬ ⎪ KFeSSA,Fe ′[SSA]T ⎠⎪ ⎝ ⎩ ⎭

[Fe′] =

cond KFeSSA,Fe ′ =

[FeSSA] [Fe′][SSA]

cond KFeL,Fe ′ =

[FeL] [Fe′][L′]

(8) (9)

Kcond FeSSA,Fe′

Thus, the substitution of Aint, molar absorptivites, and in eq 7 provides the equilibrated concentration of FeSSA at a given pH. Concentrations for other Fe species can be calculated from eqs 8 and 9. cond Determination of Kcond FeSSA,Fe′. As described in SI.9, KFeSSA,Fe′ was determined in this study by the competitive ligand method in which citrate was used as a reference ligand with spectrophotometric detection of the FeSSA complex. Conditional stability constants for ferric citrate complexes including FeIII-citrate and FeIII-citrate2 have been extensively investigated under the pH and ionic strength (58−68 mM) conditions used in this study,34 although citrate forms an achromatous complex with Fe(III) and is as such unmeasurable by visible absorbance. The Kcond FeSSA,Fe′ values determined at different pH were very similar (1.6 × 108 M−1 for pH 8.0 and 6.5, and 1.9 × 108 M−1 for pH 6.0). Comparison of determined stability constants with literature values is provided in SI.8. Determination of Complexation Parameters for FeHS. To examine the effects of fitting methodology on the determination of complexation parameters, three different fitting procedures were employed, as described in SI.10. The fitting methods included linear fits with linear-transformed equations, such as Scatchard plotting and Van den Berg linearization and nonlinear fit with the Langmuir isotherm equation.35,36 Among these, the complexation parameters from the Langmuir nonlinear regression was used for further analysis, since this fitting approach yielded the highest average goodness of fit to experimentally determined data. The relatively weak goodness of fit of the linear models is likely associated with the overestimation of free metal concentration due to the analytical errors at low concentrations of total metal.36 Moreover, the linear models have been suggested to be more susceptible to outliers.36 Given the mass balance for the Fe-binding ligands (i.e., [L]T = [L′] + [FeL], where [L]T represents total ligand concentration), the Langmuir isotherm can be derived from eq 5 as follows:

(1)

(2)

(3)

where Fe′ is the sum of all dissolved inorganic Fe(III) species at a given pH, SSA′ is the SSA species not bound to Fe(III), and −1 Kcond FeSSA,Fe′ (M ) is the conditional stability constantfor the FeSSA complex (with respect to Fe′) under the conditions used. Metal-binding sites in HS are classically characterized by multiligand models (e.g., the NICA-Donnan model)15,31,32 with ligands exhibiting a range of metal-binding affinity constants. While recognized to be a simplification, the Fe complexation by HS in this work was described by a model assuming a single ligand class only. In addition, the Fe-binding sites in HS were assumed to form a monomeric 1:1 complex with Fe(III),33 yielding the following equilibrium equation and mass law expression: Fe′ + L′ ⇄ FeL

(7)

[FeSSA] cond KFeSSA,Fe ′[SSA]T

[FeL] = [Fe]T − [FeSSA] − [Fe′]

It has been reported that SSA forms mononuclear Fe(III) complexes with stoichiometric ratios of 1:1 to 1:3.28 However, as described in SI.8, the actual stoichiometry of the Fe SSA complex under the conditions used in this study was uncertain. Thus, a conditional stability constant for FeSSA was determined by assuming a stoichiometric ratio of 1:1. The equilibrium equation and corresponding mass law expression are given by the following: Fe′ + SSA′ ⇄ FeSSA

(6)

(4)

(5)

where L′ represents free ligands not bound to Fe(III), and −1 Kcond FeL,Fe′ (M ) represents the conditional stability constant for the FeHS complex (with respect to Fe′). The complexation parameters determined in this work are average values within the specific detection window of Fe′ concentrations from 3 × 10−12 to 2 × 10−11 M for pH 6−8 (e.g., SI.2), comparable to those in the coastal seawater where Fe complexation by HS dominates (10−12−10−10 M).18

[FeL] =

cond [L]T KFeL,Fe ′[Fe′] cond 1 + KFeL,Fe ′[Fe′]

(10)

By fitting this single-binding model to the measured Fe species data, Kcond FeL,Fe′ and [L]T were determined for each HS, as shown in SI Figure S1. The Fe-binding capacity of HS (CFe) was then determined in units of mol·mg−1 by dividing [L]T by the mass 4417

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

concentration of HS (mg·L−1). The conditional stability constant −1 with respect to Fe3+ (Kcond FeL,Fe3+, M ) was also calculated by cond cond KFeL,Fe3+ = αFe3+KFeL,Fe′ at given pH (where αFe3+ = [Fe′]/[Fe3+] represents side reaction coefficient, SI Table S6). The rate constants for complex formation (kf M−1·s−1) and dissociation (kd s−1) are related by the following: cond KFeL,Fe ′ =

kf kd

Effects of pH on Complexation Parameters. Substantial increase in Kcond FeL,Fe3+ (by 2−4 orders of magnitude) was observed as pH increased from 6.0 to 8.0 (Figure 1). Indeed, Fe3+ concentration decreased by 4−5 orders of magnitude with increasing pH, while Fe-HS concentrations for the same Fe:HS ratio varied by up to 5-fold for all pHs. The positive relationship observed between pH and Kcond FeL,Fe3+is consistent with the results of previous studies for Cu2+ and Pb2+,38 which suggested that the strength of binding of these free metals by HS increased with increasing pH (4.0−8.0) in accord with the weaker competition with protons at higher pHs. In contrast to Kcond FeL,Fe3+, systematic trends were not observed for CFe and Kcond FeL,Fe′ (Figures 1 and SI S2) except for apparent decreases in Kcond FeL,Fe′ for three HS (1S101F, 1S102H, and 1S103H) with decreasing pH. Comparisons of individual HS at pHs between 6.5 and 8.0 revealed both positive and negative effects of pH on the complexation parameters, whereas these parameters were almost invariant at the same pH for some HS. A comparison of particular fractions (HA and FA) for different sources of HS also revealed little distinct trend with change in pH. Therefore, the pH effect on CFe and Kcond FeL,Fe′ is either negligible or, for some HS, Kcond FeL,Fe′ varies inversely with pH. cond The different pH effects on Kcond FeL,Fe3+ and KFeL,Fe′ can be attributed to the different definition of each stability constant. As noted above, Fe3+ concentrations varied substantially as a function of pH. In contrast, concentrations of total inorganic Fe species (i.e., Fe′) were relatively comparable for all pHs at the same Fe/HS ratios, although concentrations of ferric hydrolysis products varied with pH (e.g., major inorganic Fe species were calculated to be Fe(OH)2+ (97−98%) for pH 6−6.5 and Fe(OH)2+(33%), Fe(OH)3(aq)(22%) and Fe(OH)4−(44%) for pH 8.0). Whether any particular inorganic Fe species preferentially bind to HS remains unclear. However, it should be noted that (i) hydrolysis species may have particular affinity for HS due to their higher water exchange rates (k−w = ∼106−107 s−1) compared to Fe3+ (k−w = 1.7 × 102 s−1), and (ii) complexation parameters determined in this work are average values for the ensemble of hydrolysis products. As such, the relatively smaller effect of pH on Kcond FeL,Fe′ can be rationalized as the various Fe hydrolysis products bind to HS with comparable affinity constants. However, more detailed investigation of the mechanism of the effect of pH is beyond the scope of this study. Effects of Molecular Composition of HS on Complexation Capacity. By using simple linear regression analysis, the correlation coefficients between CFe and the parametrized values for a range of chemical properties of HS were determined (Table 2). Among the parameters tested, CFe and aromatic C content had the most significant positive correlation at all pHs examined (R = 0.69−0.89, p < 0.05; Table 2, Figure 2A). In addition, aromaticity,39 C and H mass ratio40 and specific UV absorbance (SUVA), which are classically used as indices of aromatic content of dissolved organic matter (DOM) or humification degree, had significant correlations with CFe at all pHs examined (R = 0.57−0.96, p < 0.05; Figures 2B and SI S3). Categorizing HS by origin and fraction into four different groups including soil/peat HA, aquatic HA, soil/peat FA, and aquatic FA revealed that aromatic C and CFe still had positive relationships in most of groups at pH 8.0 and FA groups at pH 6.5 (SI Figure S4). These results suggest that the Fe-binding capacity for similar groups of HS increases with an increasing degree of humification, as previously reported for heavy metal (e.g., Cu, Zn, and Pb) complexation by HA derived from sewage sludge and

(11)

Dissociation rate constants were calculated by substituting the formation rate constant, which was experimentally determined by the SSA competitive method, and the conditional stability constant into eq 11. Statistical Analysis. Statistical calculations including linear and nonlinear regression analyses were undertaken by using R statistical software version 2.13.0, as described in SI.11.



RESULTS AND DISCUSSION Variation of Complexation Parameters. The complexation parameters exhibited values ranging from 2.1 to 240 nmol·mg−1 for CFe, 3.1 × 1010 to 1.4 × 1012 M−1 for Kcond FeL,Fe′ 6 7 −1 (3.5 × 1016 to 8.3 × 1021 for Kcond FeL,Fe3+), 2.4 × 10 to 1.7 × 10 M · s−1 for kf, and 5.9 × 10−6 to 8.0 × 10−5 s−1 for kd. The values for CFe and Kcond FeL,Fe′ ranged over more than 2 orders of magnitude and Kcond FeL,Fe3+ more than 5 orders of magnitude, depending on the origin and operationally defined fraction of HS and solution pH (Figure 1, SI Table S2 and S7). When the parameters were compared for particular HS fractions at pH 6.5−8.0, the HA fraction exhibited significantly (p < 0.05) higher values than that of FA for CFe at pHs of 6.5 and 8.0 and for Kcond FeL,Fe′ at pH 6.5 cond (note that Kcond FeL,Fe3+ is proportional to KFeL,Fe′ at the given pH). The CFe averages were 2.4−3.2-fold higher for HA than for FA and Kcond FeL,Fe′ values 3.1-fold higher for HA than for FA (SI Table S2). The greater binding ability for HA is consistent with that reported by Yang and van den Berg37 in which stabilities of HA complexes with Cu, Zn, Co, Fe, and Al were found to be higher than those for FA in a seawater medium. Examination with respect to the HS source also indicated that HS originating from soil/peat had significantly higher Kcond FeL,Fe′ than those from aquatic environments including rivers and lakes at pH 6.5 (p = 0.037). Although statistically significant differences were not observed for other parameters (e.g., CFe at pH 6.5 and 8.0 and Kcond FeL,Fe′ at pH 8), HS of soil/peat origin had average CFe values higher than those for HS with aquatic origin with relatively lower p-values (0.067−0.072) obtained for pH 8 and 6.5. In general, the magnitude of the complexation capacity was in the order of soil/peat HA > aquatic HA > soil/peatFA > aquatic FA, whereas the order for the stability constant was soil/peat HA ≥ aquatic FA ≥ aquatic HA≈soil/peat FA (SI Table S3). Although discernible order was not observed for either the formation or dissociation rate constants, the results generally suggest that Fe-binding characteristics of HS are influenced to a significant extent by their origin and the particular operationally defined fraction. As summarized in SI Table S4, the stability constants for the FeHS complexes were in close agreements with those previously determined for soluble and colloidal Fe-HS 10 complexes in coastal seawaters. Values of Kcond FeL,Fe′ =1.7 × 10 to 9.3 × 1011 M−1 have been previously determined at pHs of 7.9−8.418 and values for kf = 2.1 × 105 to 9.7 × 107 M−1·s−1 and kd = 1.0 × 10−6 to 4.0 × 10−3 s−1 have been previously determined at pH 8.1.26 4418

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

cond 3+ Figure 1. (A) Complexation capacity (CFe), (B) stability constant for Fe′ (Kcond FeL,Fe′), (C) stability constant for Fe (KFeL,Fe3+)and (D) complex formation and dissociation rate constants (kf and kd, respectively) for 15 types of humic and fulvic acids at pH 6−8. At pH 6.0, the complexation parameters were measured only for 5 types of humic substances. kf and kd were determined only under the condition at pH 8. Error bars indicate ± standard deviation (n ≥ 2). In the calculation of error bars for kf and kd, average values for CFe and Kcond FeL,Fe′ were used, respectively.

reported to be 5.9−11, 1.1−2.2, and 3.7−9.8 μmol·mg−1, respectively. Karlsson and Persson (2010),33 in studies using X-ray absorption fine structure (XAFS) analysis of solutions at pH 3.0−7.2 and Fe:HS ratios (90−180 nmol-Fe·mg-HS−1) (i.e., conditions comparable to our work) suggested that mononuclear Fe(III)-HS complexes that form a five-membered chelate ring structure dominate, although the numbers of carboxyl and phenolic hydroxyl groups involved in the coordination of HS and Fe center remain unclear. Thus, assuming (i) Fe(III) is coordinated by five-membered chelate ring structures within HS, and (ii) the potential maximum coordination number for each group is five, the number of Fe-binding ligands in HS accounts for

sludge-amended soils.41 Because of the negative correlation between aromatic and aliphatic C contents (R = −0.85; SI Figure S5), a negative correlation was also observed between CFe and aliphatic C. The amount of acidic functional groups, such as carboxyl and phenolic groups and total acidity determined by aqueous acid− base titrimetry had negative correlations with CFe for all pHs (e.g., Figure 2C), except for CFe and phenolic group at pH 6.5, although such functional groups are commonly considered as major metal-binding sites.15 Indeed, our competitive ligand experiment determined CFe to be 0.0021−0.24 μmol·mg−1 for pH 8 and 0.0058−0.13 μmol·mg−1 for pH 6.5, whereas the number of carboxyl and phenolic groups and total acidity were 4419

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

k d (s−1) k f.(M−1·s−1)

K FeL,Fe′ (M−1)b

Article

only 1.0−25% of the carboxyl functional groups present (with averages of 7.9 ± 5.5%) and 8.8−109% of the phenolic hydroxyl functional groups present (with averages of 28 ± 22%). One of the reasons that Fe-binding capacities were determined to be much smaller than the concentration of acidic functional groups present is that Fe′ concentrations in the competitive experiment must be maintained below the Fe(III) solubility limit and, for this reason, the Fe′ concentrations in the titration curves for the Langmuir model fit were far from saturation for not all but many HS samples (SI Figure S1). Other plausible explanations include the possibility that Fe′ can be competitive only at sites possessing a much higher affinity for Fe′, given that the proton activities are orders of magnitude higher than Fe′ activities in all experiments: i.e., the possibility exists that only specific acidic functional groups but not all of the proton-binding sites quantified in acid−base titrations are involved in Fe(III) binding. Given the higher correlation of CFe and aromatic C content at all pHs examined, the functional groups attached to or in close proximity to the aromatic region appear likely candidates for major Fe-binding sites. Deshmukh et al.42 used NMR and FTIR spectroscopy to determine that aromatic carboxyls such as salicylate- and furoate-type structures occur as minor structures in IHSS soil and river HS samples, whereas aliphatic or alicyclic carboxyl and hydroxyl structures constitute the majority of acidic functional groups. From a more quantitative aspect, for example, aromatic carboxyl groups for SRFA (1S101F) have been reported to account for 22% of total carboxyl groups, whereas the remaining 78% have been classified as belonging to aliphatic carboxyl groups.43 If we assume that the aromatic carboxyl content is proportional to the total aromatic contents for each HS, then the aromatic carboxyl content is calculated to be comparable to the content of phenolic hydroxyl groups (the ratio ranged from 0.6 to 2.5), and we estimate roughly that the Febinding ligand accounts for up to 3.9−54% (with an average of 22 ± 14%) and 6.8−43% (with an average of 24 ± 12%) of carboxyl groups in aromatic moieties at pHs of 8 and 6.5, respectively. In addition, the proportion of Fe-binding sites relative to strongly acidic aromatic carboxylates was calculated to be ∼3-fold greater than these values, because 35% of the total aromatic carboxyl groups is likely accounted for by strong carboxyl groups with pKa values of 3.0 or less.43 Therefore, the estimated amount of functional groups with higher pKa is still comparable to the Fe-binding capacity. Effects of Molecular Composition of HS on Fe(III) Complex Stability. The relationship between Kcond FeL,Fe′ and N content in HS (%) was positive at all pHs, and relatively higher correlation coefficients were determined at pHs of 8.0 and 6.0 (R = 0.71−0.80; Figure 2D). These results may suggest that FeHS complex stability increases when both O- and N-containing ligands are involved in the multidentate coordination. The results are also consistent with the fact that chelating compounds with hydroxylamine groups, such as the bacterial siderophore desferrioxamine,13 have stronger affinities for Fe(III), and the hard and soft acid−base (HSAB) rule by which the hard acid Fe3+ prefers to coordinate with N and O atoms.44 However, it should be noted that the significant positive correlation apparent between the N content and Kcond FeL,Fe′ at pH 8 was not detected after the FA from Pony Lake was removed from the analyzed data set. In addition, significant correlations were not seen at pH 6−6.5, indicating that the relationship between Kcond FeL,Fe′ and N content was observed only in limited data sets. Similarly, the correlation between S content and Kcond FeL,Fe′ observed at pH 8 became insignificant when the Pony Lake FA results were removed.

a Asterisks represent statistically significant levels (*: p < 0.05, **: p < 0.01). Bold-faced numbers indicate correlation coefficient greater than 0.50. Numbers of HS samples subjected to statistical test were b cond cond 5 for CFe and Kcond FeL,Fe′ at pH 6.0 and 12−15 for other conditions. Note that KFeL,Fe3+ is proportional to KFeL,Fe′ at given pH.

−0.56 −0.70* −0.75 −0.17 −0.49 −0.57 0.17 −0.037 −0.28 0.15 −0.44 0.34 0.15 −0.62 −0.28 0.15 −0.44 −0.65* −0.65 −0.22 −0.47 −0.41 0.22 −0.067 0.57* 0.61* 0.96** −0.28 0.58* 0.59 0.27 0.22 0.76** 0.66** 0.86* −0.46 0.50 0.76 0.35 0.072 0.018 −0.0068 −0.29 −0.39 −0.35 0.08 0.24 −0.42 0.78** 0.69** 0.89* −0.43 0.58* 0.76 0.34 0.14 −0.51 −0.025 −0.64 −0.26 −0.35 −0.66 −0.012 0.094 −0.063 0.063 −0.24 −0.56* 0.14 0.037 −0.49 −0.68** −0.41 0.89* 0.41 −0.81 −0.47 0.73** 0.89** 0.72** −0.33 0.33 −0.01 −0.29 −0.37 0.79 0.04 −0.74 0.39 −0.66** −0.57* −0.44 −0.15 0.14 −0.046 −0.028 −0.46 −0.17 −0.72 −0.30 −0.59* −0.70 0.21 −0.20 −0.41 0.13 −0.20 −0.65**. −0.07 0.082 −0.48 −0.59* −0.75 0.89* 0.47 −0.79 −0.45 0.71**. 0.88** 0.44 −0.62* 0.37 0.03 −0.37 −0.76 0.80 0.10 −0.82 0.30 −0.64* −0.56* −0.43 −0.24 0.16 −0.040 −0.016 −0.41 −0.47 −0.39 0.44 −0.12 −0.46 −0.42 0.067 pH 8 0.61* pH 6.5 0.27 pH 6.0 0.74 pH 8 −0.02 pH 6.5 0.63* pH 6.0 0.71 pH 8 0.046 pH 8 0.16 CFe (mol·mg−1)

total acidity phenolic carboxyl aromatic carbohydrate parameters

C

H

O

N

S

H/C

O/C

O/H

N/C

S/C

aliphatic

carbon species (%) atomic ratio elemental composition (%) complexation

Table 2. Correlation Coefficients between Fe Complexation Parameters and Chemical Properties of Humic Substancesa

carbonyl aromaticity

SUVA (L·mg−1·cm−1)

acidic functional groups (meq/g)

Environmental Science & Technology

4420

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

Figure 2. Relationship between chemical properties of HS and Fe complexation parameters: (A) aromatic carbon and complexation capacity, (B) aromaticity and complexation capacity, (C) total acidity and complexation capacity, (D) nitrogen content and stability constant, (E) aromatic carbon and stability constant, and (F) nitrogen content and dissociation rate constant. Symbols represent the experimentally determined values at pH 8.0 (◊), 6.5 (×), and 6.0 (▲), respectively. Symbols and error bars indicate averaged data and ± standard deviation (n ≥ 2), respectively. The solid black, dotted black, and solid gray lines are linear regression lines for data obtained at pH 8.0, 6.5, and 6.0, respectively. Equations for linear regression lines are also indicated at each pH.

properties, implying that the nature of HS has a lesser influence on the rate of formation of the Fe(III) complexes. Potential Mechanism of Fe Complexation by HS. The direct coordination of Fe to aromatic functional groups may be possible, such as in the case of salicylate- or phthalate-type binding with metals stabilized via multidentate coordination by two adjacent aromatic carboxyl or phenolic hydroxyl groups. Approximate but reasonable quantitative agreement with specific aromatic carboxyl groups was also determined when multidentate coordination was presumed for Fe(III) complexation by HS. Baken et al.25 also observed a metal-exchange reaction between DOM and metal-loaded Chelex 100 resin at pH 6.5 and determined that metal binding affinities for freshwater DOM in natural, agricultural, and urban areas differ by 4-fold for Cu and

The effect of aromaticity on the stability constant was dependent on the pH. The stability constant had significant positive correlations with aromatic C at lower pH (R = 0.58−0.76 for pH 6.0−6.5; Figure 2E), whereas an insignificant negative correlation was observed at higher pHs. The discrepancy between lower and higher pH may be associated with complex processes that are unable to be described solely by ligand protonation, as Fe-ligand complex stability is likely governed by multiple factors including relative affinities of ferric hydrolysis species and protons to ligand atoms (e.g., HSAB rule), chelation effects, and possibly π-bonding effects. The observed correlation of kd with the N content and N/C ratio (R = −0.66 to −0.64; Figure 2F) is consistent with the data for the stability constant. In contrast, kf had no discernible relationship with chemical 4421

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

inherently present in HS samples (SI 5), effects of HS concentration on complexation parameters (SI 6), effects of HEPES concentration on FeSSA formation (SI 7), FeSSA speciation (SI 8), stability constants for Fe complexed by model ligands (SI 9), linear and nonlinear model fits (SI 10), statistical analysis (SI 11), and all associated tables and figures (SI 12 and SI 13, respectively). This material is available free of charge via the Internet at http:// pubs.acs.org.

10-fold for Cd, Zn, and Ni and that SUVA obviously accounts for such variation. This evidence is consistent with our findings that the Fe-binding capacity correlates with the amount of aromatic C in HS. However, our results do not necessarily exclude the importance of other structures for potential Fe-binding via inner- and outer-sphere coordination, which may account for the deviation from linearity of some of the dependencies examined here (Figure 2). For example, carboxyl groups clustered from aliphatic moieties, such as succinic, malic, and malonic acids, have been suggested as the best candidates for multidentate coordination sites for Cu (pH 4.5−5.5, Cu to C ratio of 13− 850 μmol.g-C),45 Ca, and Cd (≤pH 6.0) binding by HS in a simplified intramolecular model presented by Leenheer and co-workers.24 This model accounts for the abundance and geometrical fit of the relevant ligand structures to the sizes of hydrated or unhydrated metal ions. The presence of relatively longer side chains with associated reduction in steric constraints has been also suggested as a reason for the apparent stronger affinity for Fe of biogenic strong Fe-binding ligands in seawater.13 A recent study by Hertkorn et al.14 has suggested that refractory molecules with a mixture of carboxylated and fused alicyclic structures are expected to constitute strong metal-binding ligands with multiple coordination in marine environments. In addition, Leenheer et al.,24 in a study of the molecular structure of metal-binding sites for SRFA, suggested that aliphatic metalbinding sites are likely attached to or in close proximity to the aromatic structures. Although the use of different metal species, metal/HS concentration ratios, and solution pH makes it difficult to directly compare previous findings to the results of this study, both aromatic and aliphatic structures appear to be important in metal complexation in a direct or indirect manner. Environmental Implications of Findings. The Fe complexation capacity per mg-unit of HS and the stability constant of these FeHS complexes for 15 different HS indicated that the extent of Fe complexation by HS varies widely depending on the origin, operationally defined fraction, and molecular composition of HS. Overall, HS from soil/peat environments were more prone to form FeHS complexes than those of aquatic origins. Similarly, HA were more prone to form complexes with Fe than were FA. The results of correlation analysis suggest that aromatic C content is a strong molecular descriptor for complexation capacity for all HA and FA examined in this study. Although direct Fe coordination by aromatic functional groups of HS is possible, our results do not necessarily exclude the contributions of other structures including aliphatic moieties in Fe complexation. In natural waters, both DOM concentration and characteristics, including aromaticity, are significantly influenced by a range of factors including hydrological and seasonal characteristics of watershed,46,47 agricultural48 and wastewater49 effluents, and biological and physicochemical degradation50 and DOM precipitation51 during the longitudinal transport. Thus, such spatial and temporal profiles of DOM may affect the concentration of dissolved Fe complexes, Fe complexation kinetics, and availability for uptake by microorganisms.





AUTHOR INFORMATION

Corresponding Author

*Tel:+81-3-5734-2597; fax:+81-3-5734-3577; e-mail: fujii.m.ah@ m.titech.ac.jp. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was partially supported by CREST project from Japan Science and Technology Agency and Grant-in-Aid for Young Scientists (A) (25709045) and bilateral joint research project from the Japan Society for the Promotion of Science.



REFERENCES

(1) Boyd, P. W.; Jickells, T.; Law, C. S.; Blain, S.; Boyle, E. A.; Buesseler, K. O.; Coale, K. H.; Cullen, J. J.; de Baar, H. J. W.; Follows, M.; Harvey, M.; Lancelot, C.; Levasseur, M.; Owens, N. P. J.; Pollard, R.; Rivkin, R. B.; Sarmiento, J.; Schoemann, V.; Smetacek, V.; Takeda, S.; Tsuda, A.; Turner, S.; Watson, A. J. Mesoscale iron enrichment experiments 1993−2005: Synthesis and future directions. Science 2007, 315 (5812), 612−617. (2) Hutchins, D. A.; Bruland, K. W. Iron-limited diatom growth and Si:N uptake ratios in a coastal upwelling regime. Nature 1998, 393 (6685), 561−564. (3) Hales, B.; Takahashi, T. Mesoscale biogeochemical responses to iron fertilization in the upper layers of the Southern Ocean Iron Experiment areas. J. Geophys. Res.-Oceans 2012, 117. (4) Alexova, R.; Fujii, M.; Birch, D.; Cheng, J.; Waite, T. D.; Ferrari, B. C.; Neilan, B. A. Iron uptake and toxin synthesis in the bloom-forming Microcystis aeruginosa under iron limitation. Environ. Microbiol. 2011, 13 (4), 1064−1077. (5) Tipping, E.; Rey-Castro, C.; Bryan, S. E.; Hamilton-Taylor, J. Al(III) and Fe(III) binding by humic substances in freshwaters, and implications for trace metal speciation. Geochim. Cosmochim. Acta 2002, 66 (18), 3211−3224. (6) Rey-Castro, C.; Mongin, S.; Huidobro, C.; David, C.; Salvador, J.; Garces, J. L.; Galceran, J.; Mas, F.; Puy, J. Effective affinity distribution for the binding of metal ions to a generic fulvic acid in natural waters. Environ. Sci. Technol. 2009, 43 (19), 7184−7191. (7) Fujii, M.; Dang, T. C.; Bligh, M. W.; Rose, A. L.; Waite, T. D. Effect of natural organic matter on iron uptake by the freshwater cyanobacterium Microcystis aeruginosa. Environ. Sci. Technol. 2014, 48 (1), 365−74. (8) Morel, F. M. M.; Kustka, A. B.; Shaked, Y. The role of unchelated Fe in the iron nutrition of phytoplankton. Limnol. Oceanogr. 2008, 53 (1), 400−404. (9) Shaked, Y.; Lis, H. Disassembling iron availability to phytoplankton. Front. Microbiol. 2012, 3, 123. (10) Fujii, M.; Dang, T. C.; Rose, A. L.; Omura, T.; Waite, T. D. Effect of light on iron uptake by the freshwater cyanobacterium Microcystis aeruginosa. Environ. Sci. Technol. 2011, 45 (4), 1391−1398. (11) Rue, E. L.; Bruland, K. W. Complexation of iron(III) by natural organic-ligands in the Central North Pacific as Determined by a new competitive ligand equilibration adsorptive cathodic stripping voltammetric method. Mar. Chem. 1995, 50 (1−4), 117−138.

ASSOCIATED CONTENT

* Supporting Information S

Experimental procedures (SI 1), equilibrium calculation of inorganic and organic Fe(III) species (SI 2), effect of equilibration period on complexation parameters (SI 3), Fe(II) formation in FeHS solution (SI 4), concentration and chemical reactivity of Fe 4422

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

(12) Liu, X. W.; Millero, F. J. The solubility of iron hydroxide in sodium chloride solutions. Geochim. Cosmochim. Acta 1999, 63 (19−20), 3487− 3497. (13) Witter, A. E.; Hutchins, D. A.; Butler, A.; Luther, G. W. Determination of conditional stability constants and kinetic constants for strong model Fe-binding ligands in seawater. Mar. Chem. 2000, 69 (1−2), 1−17. (14) Hertkorn, N.; Benner, R.; Frommberger, M.; Schmitt-Kopplin, P.; Witt, M.; Kaiser, K.; Kettrup, A.; Hedges, J. I. Characterization of a major refractory component of marine dissolved organic matter. Geochim. Cosmochim. Acta 2006, 70 (12), 2990−3010. (15) Tipping, E. Cation Binding by Humic Substances; Cambridge University Press: Cambridge, U.K., 2002. (16) Sutton, R.; Sposito, G. Molecular structure in soil humic substances: The new view. Environ. Sci. Technol. 2005, 39 (23), 9009− 9015. (17) Hassler, C. S.; Schoemann, V.; Nichols, C. M.; Butler, E. C. V.; Boyd, P. W. Saccharides enhance iron bioavailability to Southern Ocean phytoplankton. Proc. Natl. Acad. Sci. U.S.A. 2011, 108 (3), 1076−1081. (18) Batchelli, S.; Muller, F. L. L.; Chang, K. C.; Lee, C. L. Evidence for strong but dynamic iron-humic colloidal associations in humic-rich coastal waters. Environ. Sci. Technol. 2010, 44 (22), 8485−8490. (19) Laglera, L. M.; van den Berg, C. M. G. Evidence for geochemical control of iron by humic substances in seawater. Limnol. Oceanogr. 2009, 54 (2), 610−619. (20) Boyle, E. A.; Edmond, J. M.; Sholkovitz, E. R. Mechanism of iron removal in estuaries. Geochim. Cosmochim. Acta 1977, 41 (9), 1313− 1324. (21) de Baar, H. J. W.; de Jong, J. T. M. Distributions, Sources and Sinks of Iron in Seawater; John Wiley: Chichester/New York, 2001; pp 123− 254. (22) Croue, J. P.; Benedetti, M. F.; Violleau, D.; Leenheer, J. A. Characterization and copper binding of humic and nonhumic organic matter isolated from the South Platte River: Evidence for the presence of nitrogenous binding site. Environ. Sci. Technol. 2003, 37 (2), 328−336. (23) Hay, M. B.; Myneni, S. C. B. Structural environments of carboxyl groups in natural organic molecules from terrestrial systems. Part 1: Infrared spectroscopy. Geochim. Cosmochim. Acta 2007, 71 (14), 3518− 3532. (24) Leenheer, J. A.; Brown, G. K.; MacCarthy, P.; Cabaniss, S. E. Models of metal binding structures in fulvic acid from the Suwannee River, Georgia. Environ. Sci. Technol. 1998, 32 (16), 2410−2416. (25) Baken, S.; Degryse, F.; Verheyen, L.; Merckx, R.; Smolders, E. Metal complexation properties of freshwater dissolved organic matter are explained by its aromaticity and by anthropogenic ligands. Environ. Sci. Technol. 2011, 45 (7), 2584−2590. (26) Rose, A. L.; Waite, T. D. Kinetics of iron complexation by dissolved natural organic matter in coastal waters. Mar. Chem. 2003, 84 (1−2), 85−103. (27) Fujii, M.; Rose, A. L.; Waite, T. D.; Omura, T. Effect of divalent cations on the kinetics of Fe(III) complexation by organic ligands in natural waters. Geochim. Cosmochim. Acta 2008, 72 (5), 1335−1349. (28) Ogawa, K.; Tobe, N. A spectrophotometric study of complex formation between iron(III) and sulfosalicylic acid. Bull. Chem. Soc. Jpn. 1966, 39 (2), 223−&. (29) Pullin, M. J.; Cabaniss, S. E. The effects of pH, ionic strength, and iron-fulvic acid interactions on the kinetics of nonphotochemical iron transformations. II. The kinetics of thermal reduction. Geochim. Cosmochim. Acta 2003, 67 (21), 4079−4089. (30) Mash, H. E.; Chin, Y. P.; Sigg, L.; Hari, R.; Xue, H. B. Complexation of copper by zwitterionic aminosulfonic (good) buffers. Anal. Chem. 2003, 75 (3), 671−677. (31) Kinniburgh, D. G.; Milne, C. J.; Benedetti, M. F.; Pinheiro, J. P.; Filius, J.; Koopal, L. K.; VanRiemsdijk, W. H. Metal ion binding by humic acid: Application of the NICA-Donnan model. Environ. Sci. Technol. 1996, 30 (5), 1687−1698. (32) Milne, C. J.; Kinniburgh, D. G.; Van Riemsdijk, W. H.; Tipping, E. Generic NICA-Donnan model parameters for metal-ion binding by humic substances. Environ. Sci. Technol. 2003, 37 (5), 958−971.

(33) Karlsson, T.; Persson, P. Coordination chemistry and hydrolysis of Fe(III) in a peat humic acid studied by X-ray absorption spectroscopy. Geochim. Cosmochim. Acta 2010, 74 (1), 30−40. (34) Ito, H.; Fujii, M.; Masago, Y.; Yoshimura, C.; Waite, T. D.; Omura, T. Mechanism and kinetics of ligand exchange between ferric citrate and desferrioxamine B. J. Phys. Chem. A 2011, 115 (21), 5371−5379. (35) Bruland, K. W.; Rue, E. L. Analytical Methods for the Determination of Concentrations and Speciation of Iron; John Wiley: Chichester/New York, 2001; pp 255−290. (36) Gerringa, L. J. A.; Herman, P. M. J.; Poortvliet, T. C. W. Comparison of the linear Vandenberg Ruzic transformation and a nonlinear fit of the Langmuir isotherm applied to Cu speciation data in the estuarine environment. Mar. Chem. 1995, 48 (2), 131−142. (37) Yang, R. J.; van den Berg, C. M. G. Metal complexation by humic substances in seawater. Environ. Sci. Technol. 2009, 43 (19), 7192−7197. (38) Christl, I.; Milne, C. J.; Kinniburgh, D. G.; Kretzschmar, R. Relating ion binding by fulvic and humic acids to chemical composition and molecular size. 2. Metal binding. Environ. Sci. Technol. 2001, 35 (12), 2512−2517. (39) Baldock, J. A.; Oades, J. M.; Nelson, P. N.; Skene, T. M.; Golchin, A.; Clarke, P. Assessing the extent of decomposition of natural organic materials using solid-state C-13 NMR spectroscopy. Aust. J. Soil. Res. 1997, 35 (5), 1061−1083. (40) Watanabe, A.; Itoh, K.; Arai, S.; Kuwatsuka, S. Comparison of the composition of humic and fulvic-acids prepared by the Ihss method and Nagoya method. Soil Sci. Plant. Nutr. 1994, 40 (4), 601−608. (41) Plaza, C.; Brunetti, G.; Senesi, N.; Polo, A. Molecular and quantitative analysis of metal ion binding to humic acids from sewage sludge and sludge-amended soils by fluorescence spectroscopy. Environ. Sci. Technol. 2006, 40 (3), 917−923. (42) Deshmukh, A. P.; Pacheco, C.; Hay, M. B.; Myneni, S. C. B. Structural environments of carboxyl groups in natural organic molecules from terrestrial systems. Part 2: 2D NMR spectroscopy. Geochim. Cosmochim. Acta 2007, 71 (14), 3533−3544. (43) Leenheer, J. A.; Wershaw, R. L.; Reddy, M. M. Strong-acid, carboxyl group structures in fulvic-acid from the Suwannee River, Georgia 0.1. minor structures. Environ. Sci. Technol. 1995, 29 (2), 393− 398. (44) Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; Wiley: New York, 1996; p xvi, p 1022. (45) Manceau, A.; Matynia, A. The nature of Cu bonding to natural organic matter. Geochim. Cosmochim. Acta 2010, 74 (9), 2556−2580. (46) Rosario-Ortiz, F. L.; Snyder, S. A.; Suffet, I. H. Characterization of dissolved organic matter in drinking water sources impacted by multiple tributaries. Water Res. 2007, 41 (18), 4115−4128. (47) Park, J. F.; Lee, J. H.; Kang, S. Y.; Kim, S. Y. Hydroclimatic controls on dissolved organic matter (DOM) characteristics and implications for trace metal transport in Hwangryong River Watershed, Korea, during a summer monsoon period. Hydrol. Proc. 2007, 21 (22), 3025−3034. (48) Wilson, H. F.; Xenopoulos, M. A. Effects of agricultural land use on the composition of fluvial dissolved organic matter. Nat. Geosci. 2009, 2 (1), 37−41. (49) Imai, A.; Fukushima, T.; Matsushige, K.; Kim, Y. H.; Choi, K. Characterization of dissolved organic matter in effluents from wastewater treatment plants. Water Res. 2002, 36 (4), 859−870. (50) Ward, N. D.; Keil, R. G.; Medeiros, P. M.; Brito, D. C.; Cunha, A. C.; Dittmar, T.; Yager, P. L.; Krusche, A. V.; Richey, J. E. Degradation of terrestrially derived macromolecules in the Amazon River. Nat. Geosci. 2013, 6, 530−533. (51) Sleighter, R. L.; Hatcher, P. G. Molecular characterization of dissolved organic matter (DOM) along a river to ocean transect of the lower Chesapeake Bay by ultrahigh resolution electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Mar. Chem. 2008, 110 (3−4), 140−152. (52) IHSS. International Humic Substance Society Homepage Database (http://www.humicsubstances.org/index.html). (05 June 2013 accessed). 4423

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424

Environmental Science & Technology

Article

(53) Fujitake, N.; Kodama, H.; Nagao, S.; Tsuda, K.; Yonebayashi, K. Chemical properties of aquatic fulvic acids isolated from Lake Biwa, a clear water system in Japan. Humic Subst. Res. 2009, 5/6, 45−53.

4424

dx.doi.org/10.1021/es405496b | Environ. Sci. Technol. 2014, 48, 4414−4424