Environ. Sci. Technol. 2008, 42, 2367–2373
EXAFS Study on the Reactions between Iron and Fulvic Acid in Acid Aqueous Solutions J O R I S W . J . V A N S C H A I K , * ,† INGMAR PERSSON,‡ DAN BERGGREN KLEJA,† AND JON PETTER GUSTAFSSON§ Department of Soil Sciences, Swedish University of Agricultural Sciences, Box 7014, SE-750 07 Uppsala, Sweden, Department of Chemistry, Swedish University of Agricultural Sciences, Box 7015, SE-750 07 Uppsala, Sweden, and Department of Land and Water Resources Engineering, Royal Institute of Technology, SE-100 44 Stockholm, Sweden
Received August 21, 2007. Revised manuscript received November 29, 2007. Accepted December 12, 2007.
Iron(III) competes with trace metals for binding sites on organic ligands. We used X-ray absorption fine structure (EXAFS) spectroscopy to determine the binding mode and oxidation state of iron in solutions initially containing only iron(III) and fulvic acid at pHs 2 and 4. EXAFS spectra were recorded at different times after sample preparation. Iron was octahedrally configured with inner-sphere Fe-O interactions at 1.98–2.10 Å, depending on the oxidation state of iron. Iron(III) formed complexes with fulvic acid within 15 min. Iron(III) was reduced to iron(II) with time at pH 2, whereas no significant reduction occurred at pH 4. No signs of dimeric/trimeric hydrolysis products were found in any of the solution samples (0.45 µm) showed Fe · · · Fe distances, indicating the presence of tightly packed iron(III) trimers and/or clusters of corner-sharing octahedra. It is suggested that the binding mode of iron(III) to fulvic acid at low pH may be phase-dependent: in solution mononuclear complexes predominate, whereas in the solid phase hydrolyzed polynuclear iron(III) complexes form, even at very low pH values. The observed pH dependence of iron(III) reduction was consistent with expected results based on thermodynamic calculations for model ligands.
Introduction Iron is an abundant element in natural systems, and it plays various roles in geochemistry and environmental chemistry. Iron(III) has long been known to affect the solubility of humic substances through charge neutralization and to compete strongly with other metals for binding sites on organic ligands (1–4). However, the specific details of the chemistry and mechanisms involved in the complex formation remain unclear. For example, no agreement exists with respect to the presence and nature of dissolved complexes of iron(III). It is often observed that the solubility of iron(III) in natural waters is higher than expected on the basis of iron oxide * Corresponding author e-mail:
[email protected]. † Department of Soil Sciences, Swedish University of Agricultural Sciences. ‡ Department of Chemistry, Swedish University of Agricultural Sciences. § KTH (Royal Institute of Technology). 10.1021/es072092z CCC: $40.75
Published on Web 02/21/2008
2008 American Chemical Society
solubility. Several authors claim that this high solubility of iron(III) is due to the formation of soluble iron(III)-humic acid complexes (5, 6), whereas others argue that it is due to the presence of small iron oxide particles stabilized by adsorbed humic substances (7). For iron, understanding of the mechanisms governing its interactions with humic substances is further complicated by the occurrence of redox processes in soils and waters. The iron(III)/iron(II) redox couple in relation to oxidation and reduction of iron and humic substances has been described in detail, and both iron(III) and iron(II) complexes have been reported to be important iron species in natural waters (8, 9). It has been shown that natural humic substances are capable of abiotical reduction of iron(III), a process which was found to be pH-dependent (10). In another study, however, high iron(III) concentrations were reported in peat bog pore water despite anoxic conditions (11, 12). This apparent stabilization of iron(III) was explained by the greater thermodynamic stability of the iron(III)-organic complexes compared with those of iron(II). Recent modeling efforts have provided generic parameters for the complex formation of iron(II) and iron(III) with humic and fulvic acids (13, 14). However, the database used for these model calibrations was limited, covering only a restricted pH range and containing data with poor statistics. The need for increased understanding of iron(III) binding to isolated humic substances, especially in aqueous systems, was particularly emphasized (14). In addition, the importance of including iron(III) analyses in the characterization of humic samples was highlighted. Iron binding to insolubilized humic acid was studied, and a new set of iron binding parameters both for Model VI and the NICA-Donnan model was derived (15). The modeling results indicated that carboxylic groups are the dominant functional groups involved in the binding of iron(III). It was also recently demonstrated, using electron paramagnetic resonance spectroscopy, that the binding of iron(III) to organic matter is a process of geochemical significance in the Amazon river (16). During the past decade, several studies have been conducted in which structural parameters for trace metals in the presence of humic substances were investigated with extended X-ray absorption fine structure (EXAFS) spectroscopy. Important information, such as coordination number, neighboring atoms, oxidation state and structure geometry, can be extracted from the EXAFS spectra (17). The measurements are element-specific and can be performed on any physical state of sample, and information can be obtained from dilute samples. In a recent paper in which the binding of iron(III) to organic mor layers from forest soils was studied with EXAFS spectroscopy, we showed that a major fraction of the organically complexed iron was hydrolyzed, most likely in a mixture of dimeric and trimeric complexes (18). The results were successfully used to constrain a geochemical model for metal ion-humic complexation, the Stockholm Humic Model (19), considering one dimeric hydrolyzed iron(III) humic complex. To our knowledge, no EXAFS studies have been performed to date on iron in aqueous solutions that contain dissolved humic substances. The present study examines the interaction of iron(III) with fulvic acids at pH values of 2 and 4, which are comparable with the pH range used in the study of the organic soil samples (18). Our objective was to study the binding of iron to a natural fulvic acid on a molecular level. To study reaction kinetics, we recorded EXAFS spectra on several occasions, ranging from 15 min to 34 months after preparation. Thermodynamic calculations were made on the redox VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2367
TABLE 1. Summary of Sample Parameters and Elemental Composition (IHSS, Elliott Soil I, 1S102F (45)) composition %(w/w) ash C H O
0.86 50.57 3.77 43.7
charge density parameters mequiv g-1 C
%(w/w) N S P
2.72 0.56 0.03
a
carboxylic phenolicb
13.24 2.27
a Charge density (mequiv g-1 C) at pH 8.0. b Twice the change in charge density (mequiv g-1 C) between pH 8.0 and pH 10.0.
properties of FeIIIL/FeIIL (L ) acetate, phenolate, phthalate, salicylate, and catecholate) couples in order to understand the observed redox properties of the iron-fulvic acid aqueous system.
Materials and Methods Sample Preparation. EXAFS measurements were performed on samples that initially contained 0.9 mmol · dm-3 iron(III) and 1.0 g · dm-3 fulvic acid. Iron(III) perchlorate solution (1.8 mmol · dm-3) was prepared by dissolving iron(III) perchlorate hexahydrate, [Fe(OH2)6](ClO4)3 (G. F. Smith), in hydrochloric acid (0.01 mmol · dm-3). An aqueous solution of 2.0 g · dm-3 fulvic acid (IHSS Soil Fulvic Acid Standard IIs1S102F; isolated following Swift (20), Table 1) was prepared by dissolving 20.0 mg in 10 mL of hydrochloric acid (0.01 mmol · dm-3). Two 5.00 mL aliquots of the two solutions were thoroughly mixed in 20 mL polyethylene vials, giving two samples with a final concentration of 0.9 mmol · dm-3 iron(III) perchlorate and 1.0 g · dm-3 fulvic acid. This corresponds to a molar ratio of iron(III) to fulvic acid–acidic (carboxylic plus phenolic) groups of approximately 0.12 mol · equiv-1. The pH value in the samples obtained was approximately 1.9. In one sample, pH was immediately adjusted to approximately 4.1 with a 1.0 mol · dm-3 sodium hydroxide solution. The solutions were pale yellowish-brown at both pH values. A total of three sample pairs (A-C), each containing one pH 2 and one pH 4 sample, were prepared. Sample pairs A-C were prepared 30 days, 7 days, and 15 min, respectively, prior to the first EXAFS measurements. The first EXAFS measurements on series A were made after 30 days, on series B after 7 days, and on series C after 15 min, 24 h, and 48 h. Additional EXAFS measurements were made on series A after 12 months and on series C after 28 (pH 4) and 34 (pH 2) months. The initial composition of the samples is presented in Table 3. The solutions were stored in darkness for up to 34 months; the first month at room temperature (+20 °C) and then at +5 °C to prevent microbial activity. Storage was under oxic conditions, since no attempts were made to exclude oxygen during preparation and storage of the solutions. The samples were wrapped in aluminum foil during transport to minimize temperature fluctuations and avoid exposure to light. The total iron concentration chosen, 0.9 mmol · dm-3, is the lowest possible concentration to obtain acceptable EXAFS data. Using comparable high fulvic acid concentrations, a field-realistic iron:fulvic acid ratio was obtained. Prior to EXAFS measurements, subsamples of 1–2 mL were filtered (Acrodisc minispike syringe filter; poly(tetrafluoroethylene) (PTFE) membrane; pore size, 0.45 µm) and consequently injected into a specially designed cell, which consisted of a 2 mm Teflon spacer and 6 µm polypropylene X-ray film windows held together with a brass frame (see ref 17, Figure 44b, p 112). Before collection of the final set of EXAFS spectra, the total remaining sample volume was filtered (Acrodisc syringe filter; PTFE membrane; pore size, 0.45 µm), and spectra were recorded for both the aqueous sample and the solid phase, which remained on the filter. 2368
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
EXAFS. A first set (15 min to 1 month old samples) of iron K-edge X-ray absorption spectra was recorded at the bending magnet beam line 2–3 at Stanford Synchrotron Radiation Laboratory (SSRL). The beam line was equipped with a Si[220] double crystal monochromator, the storage ring was operated at 3.0 GeV, and a maximum current of 100 mA was used (21). Two more sets (>1 month old samples) of spectra were recorded at the wiggler beam line I811 at MAX-Laboratory, Lund University, Sweden. This beam line was equipped with a Si[111] double crystal monochromator, the storage ring was operated at 1.5 GeV, and a maximum current of 200 mA was used (22). Higher-order harmonics were reduced by detuning the second monochromator crystal to reflect 40% of the maximum intensity at the end of the scans. Iron foil was used as an internal energy calibrant. The experiments were performed in fluorescence mode using a Lytle detector with a manganese filter at ambient temperature. EXAFS Data Analysis. The primary treatment (energy calibration and averaging) of the EXAFS data was performed with the EXAFSPAK program package (23). The energy scale of each X-ray absorption spectrum was calibrated by assigning the first inflection point of the K edge of metallic iron foil as 7111.3 eV (24). After primary data treatment, the EXAFSPAK (23) and GNXAS (25, 26) program packages were used for further data treatment. The GNXAS code is based on calculation of the EXAFS signal and subsequent refinement of the structural parameters (25, 26). The GNXAS method accounts for multiple scattering (MS) paths by including the configurational average of all the MS signals to allow fitting of correlated distances and bond distance variances (Debye–Waller factors). A correct description of the distribution of the ion– ligand distances in a coordination shell should in principle account for asymmetry (27, 28). Therefore the Fe-O two body signals associated with the first coordination shells were modeled with Γ-like distribution functions, which depend on three parameters: the coordination number N, the average distance R, and the mean-square variation σ. R is the first moment of the function 4π∫g(r)r2 dr; all distances have been modeled assuming Gaussian distribution. The standard deviations given for the refined parameters were obtained from k2-weighted least-squares refinements of the EXAFS function χ(k) and do not include systematic errors of the measurements. These statistical error values provide a measure of the precision of the results and allow reasonable comparisons of e.g. the significance of relative shifts in the distances. However, the variations in the refined parameters, including the shift in the Eo value (for which k ) 0), using different models and data ranges, indicate that the absolute accuracy of the distances given for the separate complexes is within (0.01-0.02 Å for well-defined interactions. The “standard deviation” values given in the text have therefore been increased accordingly to include estimated additional effects of systematic errors. Chemical Analysis. A portion of each sample was filtered (0.45 µm) of the pH 2 sample contained iron in oxidation state +III only. In addition, contrary to any of the aqueous samples, the precipitate showed a Fe · · · Fe distance at 3.30(2) Å, which fits perfectly within the range of 3.28–3.32 Å such as that found for an iron trimer, O(FeOR)3 (see ref 18, and references therein) or clusters of cornersharing FeO6 octahedra. The FT spectra also indicate longer Fe · · · Fe distances at approximately 4.4 and 5.5 Å. This suggests a very tight packing of iron(III) ions in the solid phase. 2370
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
Structure of the Iron(III)-Fulvic Acid Complexes. The aqueous samples in this study were dominated by a first shell Fe-O distance, the exact distance depending on the degree of reduction of iron(III), which was favored by a lower pH. The samples in which iron(III) predominated showed Fe-O-C and Fe · · · C distances typical of iron(III)-carboxylate and -phenolate complexes (see ref 18, and references therein). The absence of a Fe · · · Fe distance at 3.3 Å in all filtered samples suggests that iron(III) formed a mononuclear complex with the fulvic acid. In contrast, the EXAFS spectra of the precipitate isolated from the pH 2 sample showed Fe · · · Fe distances that indicate the presence of hydrolyzed iron trimers and/or FeO6 clusters in a tightly packed configuration. The presence of polynuclear iron forms in the precipitate is in line with our previous findings for organic soils (18). In that study we investigated the chemical state of iron in organic mor layers at ambient iron levels and with added iron(III). The results showed Fe · · · Fe distances indicating complexes containing a mixture of dimers and trimers with one and three µ-oxo bridges between the iron(III) ions, respectively. The iron(III) addition experiments were carried out in strongly acidic conditions (2.2–2.9); i.e., dimeric/ trimeric iron(III) complexes can be formed at low pH values in the presence of solid-phase humic substances. Our data suggest that trimeric iron(III) complexes and/or clusters of corner-sharing octahedra can be formed with precipitated (>0.45 µm) humic material at pH values as low as 1.9. Furthermore, the formation of trimeric iron(III) complexes with the particulate phase appears to stabilize iron in a trivalent state. The iron found in solution at pH 2 was almost exclusively iron(II) (Figure 1, Tables 2 and 3). The fact that no polynuclear iron complexes were found in the solution phase (
