Differential Adsorption of Molybdate and Tetrathiomolybdate on Pyrite

2r), on synthetic pyrite (FeS2) as a function of solution composition. Both MoO4. 2r and MoS4. 2r partitioned strongly on FeS2 under a range of condit...
0 downloads 0 Views 291KB Size
Environ. Sci. Technol. 2003, 37, 285-291

Differential Adsorption of Molybdate and Tetrathiomolybdate on Pyrite (FeS2) BENJAMIN C. BOSTICK AND SCOTT FENDORF* Department of Geological and Environmental Sciences, Stanford University, Stanford, California 94305-2115 GEORGE R. HELZ Department of Chemistry & Biochemistry, University of Maryland, Chemistry Building 091, Room 3101, College Park, Maryland 20742-2021

Molybdenum is a nutrient important for a variety of biological functions, most notably nitrogen fixation. Molybdenum availability is limited through sorption reactions, particularly in environments rich in sulfide minerals. This study examines the sorption of two major molybdenum species, molybdate (MoO42-) and tetrathiomolybdate (MoS42-), on synthetic pyrite (FeS2) as a function of solution composition. Both MoO42- and MoS42partitioned strongly on FeS2 under a range of conditions and ionic strengths. Molybdate and tetrathiomolybdate adsorption obeyed a Langmuir isotherm with a calculated site density between 2 and 3 sites/nm2 under acidic and circumneutral conditions, which decreased to less than 1 site/ nm2 at pH 9. Although both MoO42- and MoS42- adsorbed most strongly under moderately acidic conditions, MoO42readily desorbed while MoS42- remained adsorbed even at high pH. The reversibility of MoO42- adsorption suggests the formation of labile surface complexes while MoS42-likely forms strong inner-sphere complexes. X-ray absorption spectroscopy was used to determine the structure of the surface complexes. Molybdate formed bidentate, mononuclear complexes on FeS2. The Mo-S and Mo-Fe distances for tetrathiomolybdate on pyrite are consistent with the formation of Mo-Fe-S cubane-type clusters. The high affinity of MoS42- for FeS2, as well as its resistance to desorption, supports the hypothesis that thiomolybdate species are the reactive Mo constituents in reduced sediments and may control Mo enrichment in anoxic marine environments.

Under such conditions, molybdenum is stable as either hexavalent molybdate, MoO42-, or octahedral polymolybdate complexes that each react only weakly with aluminosilicates and metal (hydr)oxides (1-4). Molybdate reacts more strongly with manganese oxides (5, 6) and some clay minerals at low pH (1); however, MoO42- is mobile and available for biological uptake in most systems. Molybdenum chemistry in reduced (anoxic) environments is more complex. Molybdenum is scavenged in such environments (7-12), leading to nonconservative behavior in sulfidic sediments (13), estuaries (14), and other reduced environments. In some cases, the decrease in soluble Mo concentrations may contribute to Mo deficiencies (15, 16), although nitrogen fixation is most commonly limited by iron supply in marine environments (17). The enrichment observed in anoxic sediments is directly correlated to the presence of sulfidessuboxic (sulfide-free, intermediate redox status) environments exhibit little or no Mo enrichment (9). Consequently, Mo enrichment has been used as a proxy for the presence of sulfidic, anoxic conditions during sediment deposition and genesis (e.g., refs 9, 10, 17, and 18). The mechanism of Mo retention in sulfidic environments is not well established; however, molybdenum enrichment has often been attributed to the reduction of MoO42- to form molybdenum(IV) sulfide (e.g., ref 8).

MoO42- + 2e- + 2HS- + 6H+ S MoS2 + 4H2O, ∆G0 ) -314.3 kJ/mol (1) Although plausible for some systems, MoS2 rarely precipitates due to kinetic limitations (19). Elements such as U, W, and Re are concentrated in sulfidic soils and sediments through the formation of insoluble reduced oxide phases (9, 12, 20). In contrast, Mo does not form insoluble oxides and only rarely forms insoluble molybdate solids. Molybdenum enrichment may also involve simple surface complexation or it may involve coprecipitation with iron (di)sulfides including pyrite (5, 21, 22). Thiomolybdate intermediate species, MoOxS4-x2-, may also be important precursors to either precipitation or Mo adsorption in sulfidic environments (e.g., refs 11 and 19). Unfortunately, at the present time insufficient data are available to define the process(es) responsible for Mo retention in anoxic sediments and soils. Sulfide minerals may have a controlling impact on Mo sequestration in anoxic environments. Accordingly, this research examines the role of adsorption processes on pyrite in the enrichment of Mo within anoxic environments. Both molybdate and tetrathiomolybdate adsorption were examined, and surface structures were resolved using X-ray absorption spectroscopy (XAS).

Introduction

Materials and Methods

Molybdenum has many vital biological functions including nitrogen fixation and nitrogen assimilation. The Mo-Fe-S cubane in nitrogenase is responsible for the reduction of N2(g) to NH4+, a critical step in the nitrogen cycle. Molybdenum also forms the active site of nitrate reductase, thereby allowing the incorporation of nitrate N. These essential ecosystem functions are regulated by the availability of Mo; consequently, processes that control Mo availability are of great interest. Molybdenum deficiency is uncommon in nearneutral soils and sediments under typical, oxic conditions.

Reagents. All reactions were performed in a glovebox (95% N2, 5% H2) that maintained O2 concentrations below 1 ppm. Synthetic pyrite, FeS2, was purchased from Strem Chemical (>99% purity). The mineral surface was prepared by washing with water, 0.1 M HCl, and finally 0.01 M sulfide solutions to remove oxidized surface species. The washed FeS2 has a surface area of 41.7 m2/g as determined by a three-point nitrogen BET isotherm. X-ray photoelectron spectroscopy of the resulting mineral surfaces showed no evidence of sulfate, thiosulfate, or other oxidized sulfur species, although limited surface oxidation of Fe was apparent regardless of cleaning procedures. The washed material was used immediately to minimize further oxidation.

* Corresponding author. Tel: 650-723-5238. Fax: 650-725-2199. E-mail: [email protected]. 10.1021/es0257467 CCC: $25.00 Published on Web 12/06/2002

 2003 American Chemical Society

VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

285

All other chemicals were of analytical grade. Molybdenum was added to suspensions using 10 mM stock solutions of Na2MoO4‚2H2O or (NH4)2MoS4; 10 mM stock solutions of sulfide were prepared using Na2S‚9H2O. Reagents were stored under nitrogen and used as purchased. Adsorption Experiments. Adsorption of MoO42- and MoS42- on FeS2 was investigated under a wide variety of conditions. A series of FeS2 suspensions were reacted with MoO42- or MoS42- (0-100 µM Mo) as a function of suspension density (0.5-2 g/L), pH (3-11), ionic strength (0.005-0.5 M, adjusted with NaCl), and addition of sulfide (0-2 mM total sulfide). Adsorption isotherms were performed at pH 4, 7, and 9 by buffering 1 g/L mineral suspensions with 0.002 M acetate, (3-N-morpholino)propanesulfonic acid (MOPS), or borate, respectively, to maintain sample pH. The effect of pH on sorption was also investigated by adjusting the pH of mineral suspensions containing Mo(VI). Molybdate, 50 µM, was added to a 1 g/L FeS2 suspension having an initial pH between 2 and 3. The pH was measured after equilibration for 30 min, sampled, and then increased with a small volume of 0.5 M NaOH. This procedure was repeated until the pH reached ∼11. Similar experiments were performed using suspensions in which the initial pH was high (∼11) and titrated to pH ∼3 with 0.5 M HCl. Suspended solids were removed from subsamples by filtration (0.2-µm pore size), and residual solutions were analyzed for total Fe, Mo, and S using a Thermo Jarrell Ash IRIS inductively coupled plasma-optical emission spectrometer (ICP-OES). The extent of sorption was determined by the difference between initial and final Mo concentrations. Samples were prepared for XAS by reacting 100 µM molybdate or tetrathiomolybdate with 1 g/L FeS2 at pH 9 under total sulfide concentrations of 0.1, 2, and 5 mM. Sulfide concentrations were chosen to poise the suspensions at a constant redox potential and also served to stabilize either thiomolybdate or molybdate. At pH 9, MoO42- solutions are stable at total sulfide concentrations below ∼2 mM; MoS42solutions are stable at concentrations of g2 mM (19). Following reaction for 3 h, solids were isolated by filtration and sealed in Kapton film, stored under nitrogen, and analyzed within 24 h. X-ray Absorption Spectroscopy. X-ray absorption spectroscopy was performed at the Stanford Synchrotron Radiation Laboratory on beam lines 4-3 and 11-2. The storage ring operated at 3.0 GeV and at currents between 50 and 100 mA. Spectra were taken with a Si(220) double-crystal monochromator with an unfocused beam. Incident and transmitted intensities were measured with 15-cm N2-filled ionization chambers. Sample fluorescence was measured with a 13element Ge detector containing a 6 µx Ge filter. The incident beam intensity was detuned ∼50% to reject higher-order harmonic frequencies and to prevent detector saturation. Spectra were internally calibrated by placement of a molybdenum metal foil between the second and third ionization chamber. X-ray absorption spectra were collected from -200 to +1000 eV about the Mo K-edge (20 000 eV). At least five spectra were collected for each sample and averaged for analysis. XAS spectra were analyzed using WinXAS (23) in combination with Feff 8.0 software (24-26). The X-ray absorption near-edge structure (XANES) spectra provide information about the oxidation state and coordination environment of the adsorbed Mo. XANES spectra were not smoothed to preserve line shapes; however, first-derivative XANES spectra were smoothed 5% to decrease noise levels. XANES spectra were compared to several model compounds (including MoS2, MoO42- (as Na2MoO4), and MoS42- (as (NH4)2MoS4)) to determine the structural and oxidation-state changes resulting from sorption and to validate the phase and amplitude functions derived using Feff. 286

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

FIGURE 1. Adsorption isotherms for MoO42- (a) and MoS42- (b) on 1 g/L FeS2 as a function of pH. These experiments were performed without added sulfide. Error bar represents the maximum analytical uncertainty for each adsorption isotherm. Extended X-ray absorption fine structure (EXAFS) spectroscopy was also used to determine the coordination environment of adsorbed molybdenum species. The k3weighted χ(k) spectrum was isolated using a seven-point spline function and an E0 of 20 000 eV. The EXAFS spectrum was then Fourier transformed without smoothing to produce a radial structure function (RSF) using a k-range of approximately 3-14 Å-1. Preliminary fits on Fourier-filtered shells were used as a guide for final fits, which always used unsmoothed k-weighted χ(k) spectra. This procedure yields information about the local structure of Mo at the pyrite surface, including the coordination number (CN), distance (R), and the Debye-Waller factor (σ2) for each shell. EXAFS fitting provides interatomic distances accurate within 0.02 Å; larger errors in CN and σ2 (up to 30%) result due to their correlation. Differences in phase and amplitude functions and interatomic distances permit the identification of different elements.

Results and Discussion Solution-Phase Experiments. Both one- and two-site Langmuir isotherms were used to describe MoO42- and MoS42sorption on FeS2 (Figure 1). Freundlich isotherms were less effective at fitting the adsorption data. Fitting the data with two Langmuir sites did not improve the accuracy of fitting at low adsorption densities except for MoS42- at pH 7, which contained two distinct adsorption regions (Figure 2). Anion adsorption reactions commonly obey Langmuir isotherms, reaching an adsorption maximum when surface sites are saturated. Precipitation reactions, which do not exhibit adsorption maximums, are distinct from the observed data, suggesting that molybdenum does not undergo precipitation on FeS2. Isotherm data can be interpreted to yield information about the affinity and site density of Mo species on FeS2 (Table 1). The monolayer capacity for MoO42- was 130 µmol g-1 at pH 4 and 7, decreasing to 60 µmol g-1 at pH 9. The monolayer capacity for MoS42- ranged between 150 µmol

sphere complexes; therefore, these data imply that a portion of adsorbed MoO42- and MoS42- is present as outer-sphere complexessor at least a labile complex. Although the negative FeS2 surface charge at pH >3 (35) makes outer-sphere complexes unlikely, a small fraction of the species could be bound to oxidized regions of the surface that have positive surface charge at pH 7. Outer-sphere molybdate complexes commonly form on iron (hydr)oxides (34); thus, ionic strength would affect adsorption (at least to a limited extent). Interestingly, the effect of ionic strength was most pronounced when excess sulfide was present in solution (Figure 4). The reason for this combined effect of ionic strength and sulfide addition is unclear, although it could involve the adsorption of MoOxS4-x2- species formed by the partial or complete transformation of MoO42- and MoS42- to intermediate MoOxS4-x2- species. Such conversions are relatively rapid at pH 7 (11) and consequently may affect the net adsorption observed in these experiments.

FIGURE 2. One-site (a) and two-site (b) linearized Langmuir isotherm plots of MoO42- and MoS42- adsorption on 1 g/L FeS2 at pH 7. Error bars denote the uncertainty associated with Mo analysis. g-1 at pH 4 to 12.1 µmol g-1 at pH 9. The monolayer capacities correspond to site densities for both MoO42- and MoS42- of ∼2 sites/nm2 at pH 4 and 7 and much lower at pH 9. A site density of 2 sites/nm2 corresponds to 0.5 reactive sites/unit cell, assuming that the sites are uniform and on the (100) plane (the most stable surface and usual termination face for pyrite). This suggests that not all surface functional groups are reactive. Defects sites are well recognized for their important roles in adsorption (27-29), and they are also a locus for pyrite oxidation (30-32). The presence of defects may result in the high-energy site suggested by the 2-site Langmuir isotherm for MoS42- at pH 7; however, more data are needed to accurately determine which FeS2 surface reactive groups react with molybdenum. Experiments were also performed to examine the effect of pH on adsorption (Figure 3). Molybdate exhibited an adsorption maximum between pH 5 and 6, with adsorption decreasing sharply as the pH departed from this range. A similar pH of maximum adsorption was noted for molybdate on amorphous iron sulfide (33) and iron (hydr)oxide (4, 34) and correlates well with the pKa of both H2MoO4 dissociations (4.24 and 4.00, respectively). Although adsorption was strong between pH 4 and 7, MoO42- was rapidly desorbed by changes in pH (Figure 3). Reversible adsorption behavior implies that the MoO42- surface complexes are labile, consistent with a weak inner-sphere or outer-sphere complex. In contrast, MoS42- adsorption was greatest under acidic conditions and steadily decreased as pH became more basic (Figure 3). Tetrathiomolybdate adsorption was considerably less reversible than MoO42-; only a fraction of the MoS42- was released following adsorption even as pH increased to nearly 11. Hysteretic adsorption implies inner-sphere thiomolybdate complexes on FeS2, although it is not possible to discriminate between adsorption and precipitation based on irreversibility alone. The effect of ionic strength on sorption was also used to examine the binding strength of the surface complexes. Both MoO42- and MoS42- adsorption were impacted somewhat by increasing the ionic strength (Figure 4); MoO42- adsorption was affected by salt additions at all concentrations, while MoS42- adsorption was unaffected except at low ionic strengths. Ionic strength changes generally affect outer-

Sulfide additions affect the adsorption of both MoO42and MoS42- considerably (Figure 4). In all cases, the presence of excess sulfide inhibited Mo sorption. Inhibition was most pronounced for the smallest additions of sulfide; as sulfide concentrations were increased, inhibition also increased but their effect became incrementally smaller. Inhibition of Mo adsorption implies that sulfide competes for surface sites or is evolved during the binding of Mo. Molybdenum sulfide precipitation would also be enhanced by sulfide addition (eq 1). The opposite trend was observed; thus, precipitation of molybdenum sulfides is not likely. Precipitation mechanisms are also unlikely given the adherence of adsorption to the Langmuir isotherm. Nevertheless, spectroscopic confirmation is necessary to validate this conclusion. Solid-Phase Analysis. Macroscopic studies provide information about the influence of solution composition and the affinity of FeS2 surfaces for molybdenum species. Spectroscopic information is needed to define the adsorption mechanism. Here we used X-ray absorption spectroscopy to define the local structure of molybdenum adsorbed on FeS2. XANES spectroscopy is diagnostic of the oxidation state and coordination, permitting the differentiation of MoO42- and MoS42- (Figure 5). The spectrum of MoO42- contains a large preedge feature and a main edge position (defined by its inflection point) at 20 016 eV; MoS42- does not have a pronounced preedge feature and has an edge position 10 eV lower than molybdate. The difference is a consequence of the greater covalency in MoS42-and permits the simple differentiation of oxygen and sulfur coordination for Mo(VI). As such, XANES spectroscopy affords the opportunity to discern the conversion of tetrahedral Mo-O coordination in MoO42- to either tetrahedral or octahedral sulfur coordination. The spectrum of adsorbed molybdate samples contained a preedge feature and an edge position of 20 016 eV (Figure 5), consistent with MoO42-. Thus, it appears that the structure of MoO42- is largely unchanged following adsorption on FeS2. In contrast, MoS42- undergoes major spectral, and thus structural, changes during adsorption (Figure 5). While spectra of adsorbed MoS42- lack the presence of a prominent preedge feature, the derivative spectra contain two maximums at 20 007 and 20 019 eV, neither of which corresponds to MoO42- or MoS42-. Consequently, XANES spectroscopy suggests that the coordination environment of MoO42- is retained during adsorption, while the coordination environment of MoS42- undergoes conversion to a different form. Further information regarding the structural environment of Mo on FeS2 was elucidated with EXAFS spectroscopy. Molybdate adsorbed on FeS2 contains both Mo-O and MoFe shells (Figure 6). The Mo-O shell is fit with a single Mo-O VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

287

TABLE 1. Langmuir Adsorption Parameters for MoO42- and MoS42- on FeS2 Γmaxa (µmol/g)

Γmaxb (µmol/m2)

site densityc (sites/nm2)

Kd (L/µmol)

pH

MoO42-

MoS42-

MoO42-

MoS42-

MoO42-

MoS42-

MoO42-

MoS42-

4 7 9

130.0 136.8 60.9

147.7 122.5 12.1

3.12 3.28 1.46

3.54 2.94 0.291

1.88 1.97 0.88

2.13 1.76 0.17

0.12 0.21 0.014

0.28 0.061 0.042

a Maximum surface coverage (monolayer capacity). coverage. d The Langmuir adsorption affinity term.

b

Calculated using a surface area of 41.7 m2/g. c Calculated from the maximum surface

FIGURE 3. Adsorption envelopes for 50 µM MoO42- (a) and MoS42(b) on 1 g/L FeS2. These experiments were performed without added sulfide. The analytical uncertainty for measurement of Mo is 5% of the measured (dissolved) As concetration or less. distance, d(Mo-O), of ∼1.76 Å (Table 2), consistent with that of molybdate. Disordered coordination environments are commonly observed for strongly adsorbed (inner-sphere) molybdate species, which contain short ModO bonds near 1.7 Å and longer bridging MosO bonds at ∼1.9 Å (36, 37). Octahedral polymolybdate species, such as heptamolybdate, also contain a disordered Mo-O shell with distances ranging from 1.7 to 2.1 Å (38). The symmetry of the Mo-O shell agrees with the macroscopic observations that the adsorbed Mo complexes are relatively weak; strongly adsorbed complexes would lead to different bond lengths for bridging and terminal Mo-O bonds. The structure of the adsorbed molybdate can be probed in more detail using the Mo-Fe shell. The Mo-Fe shell of 12 µmol of MoO42-/g of FeS2 is fit with a single Fe atom at ∼2.85 Å (Table 2), consistent with the formation of a bidentate, mononuclear complex (Figure 7). The Mo-Fe shell was not detected for the 3 µmol of MoO42-/g of FeS2 sample; however, a more disordered shell and limited k-range prohibited detection. A similar bidentate complex is formed for molybdate adsorption to a variety of other substrates, including iron and aluminum (hydr)oxides (34). The spectra of MoS42- adsorbed on FeS2 are more complex than that for MoO42- (Figure 6). Three features are resolved in the Fourier-transformed spectra, an Mo-O shell, a Mo-S shell, and a Mo-Fe shell. The Mo-O shell is fit at a distance of 1.75 Å, similar to other molybdate species, and indicates a limited conversion of MoS42- to MoO42- during reaction. The change toward oxo-coordination is limited during the 288

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

FIGURE 4. Effect of ionic strength (a) and total sulfide (b) on MoO42and MoS42- adsorption on 1 g/L FeS2 at pH 7. For both experiments, total Mo was 100 µM. The ionic strength was 0.005 M for the added sulfide experiments (1 mM H2S). The typical error (n ) 3) is ∼5%, or twice the size of the symbols. time scale of the experiment (the Mo-O coordination number is only 1) and does not impact our ability to determine the structure of adsorbed thiomolybdate species. The second shell is fit with a Mo-S pair at 2.40 Å, elongated relative to the MoS42- standard (with Mo-S distances of 2.19 Å). In contrast, the Mo-Fe interatomic distance is contracted to ∼2.70 Å. It is therefore apparent that MoS42- undergoes extensive structural rearrangement as a consequence of adsorption on FeS2. The presence of a Mo-Fe shell at 2.7 Å also indicates that Mo forms an inner-sphere complex on FeS2, in agreement with nonreversible adsorption (Figure 3) and only a small dependence on ionic strength. Given the long (2.40 Å) Mo-S bonds, the minimum Mo-Fe distance for a bidentate, mononuclear complex would be ∼3.1 Å, even given a tilt angle (the Mo-S-Fe bond angle) of nearly 90°. Monodentate complexes or bidentate, binuclear complexes have longer Mo-Fe distances. The observed Mo-Fe distances are ∼2.7 Å, too short to be either a monodentate or bidentate complex. Thus, a face-sharing (tridentate) adsorption complex must be formed through MoS42- adsorption on FeS2. Face-sharing anion complexes are rare because they result in short interatomic distances between ions of similar and

Fe bonds at ∼2.7 Å (Table 2). Therefore, we identified MoS42adsorbed on FeS2 as a Mo-Fe-S cubane. Adsorption Processs. The combination of spectral and adsorption data allows us to create a chemical expression describing the surface reaction of Mo. A bidentate, mononuclear complex was identified for MoO42- adsorption on FeS2. One possible reaction forming such a complex is described below (eq 2). Such a reaction is consistent with -

tFe - OH + MoO42- S tFe - O2MoO2 + OH-

(2)

Mo adsorption on a range of other media, including iron and aluminum oxides (3, 34). Additionally, it explains the reactivity of protonated surface groups and the observed pH dependence. Moreover, it is consistent with reactions proposed for anion adsorption on other sulfide minerals (42, 43):

tMe - OH + HCO3 - S tMe - OHCO20 + OH-

(3)

A reaction can also be postulated for the adsorption of MoS42on FeS2, forming a Mo-Fe-S cubane as described in eq 4.

3tFe - OH + MoS42- S tFe3 - S3MoS - + 3OH -

FIGURE 5. Raw and first-derivative XANES spectra of 100 µM MoS42(a) and MoO42- (b) adsorbed on FeS2 at pH 9 under varying sulfide activities. The solid and dashed vertical lines in the derivative spectra show the edge position of tetrathiomolybdate and molybdate, respectively.

(4)

While a variety of cubane structures are possible, this cubane is chosen since the Fe atoms can remain associated with the surface, resulting in a relatively simple inner-sphere complex. Sandwich-type cubanes or other structures may have similar EXAFS spectra but require appreciable rearrangement. Furthermore, reaction 4 is consistent with the observed pH dependence of MoS42- adsorption. Equations 2-4 assume that the reactive surface groups are protonated hydroxyl groups, similar to those described for anion adsorption on other sulfides (42, 43). This assumption is only reasonable for sulfide minerals in lowsulfide-activity solutions owing to the conversion of hydroxyls to sulfhydryl functional groups (28, 44).

tFe - OH + H2S S tFe - SH + H2O

FIGURE 6. The k-weighted χ(k) spectra (a) and corresponding Fourier transforms (b) of MoO42- and MoS42- adsorbed on FeS2 at pH 9. The fits (dotted lines) of each spectrum use the fit parameters in Table 2.

high charge. However, Mo forms a wide variety of cubanes with Fe and S (e.g., Figure 7), in which Mo-S and Fe-S tetrahedra are face-shared (39, 40). Molybdenum cubanes occur in a multitude of chemical environments, including the active site of nitrogenase (41), and have a coordination environment similar to that observed in these sampless elongated Mo-S bonds (at about 2.35-2.4 Å) and short Mo-

(5)

The competition between aqueous sulfide and molybdenum species for the hydroxyl surface groups limits adsorption of the latter. The dramatic decrease in Mo adsorption observed in sulfide-rich solutions suggests that this model of sulfide competitive adsorption is valid. Although these reactions are plausible and are consistent with the observed data, further studies are needed to determine the stoichiometry and complete reaction of MoO42- and MoS42- with FeS2. Nevertheless, the data presented here serve as useful descriptors of Mo adsorption. Environmental Interpretation. Molybdate and tetrathiomolybdate adsorption on FeS2 conforms to a Langmuir isotherm. Molybdate formed a labile bidentate, mononuclear complex on FeS2 with a maximum surface concentration at pH 5. Tetrathiomolybdate adsorbed through a complex mechanism that was irreversible and resulted in a Mo-Fe-S cubane. Cubanes are often encountered in nature, although their presence on FeS2 surfaces is somewhat surprising. The accompanying structural rearrangement required for cubane formation may be responsible for the irreversible adsorption. Molybdenum enrichment in sulfidic environments has been explained by a combination of MoS2 precipitation, Mo coprecipitation with iron sulfides or other phases, and adsorption on mineral surfaces. Although Mo is probably retained through each of these mechanisms in some environments, the results presented here suggest that pyrite adsorption reactions may play an important role in Mo sequestration. Furthermore, our results support the hypothesis (11, 19, 45) that the conversion of MoO42- to MoS42drives Mo retention in anoxic sediments. Sulfidization of VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

289

TABLE 2. Molybdenum Local Structure for Selected Compounds and Adsorbed Species on FeS2a Mo-O

Mo-S σ2 (Å2)

Mo-Fe

CN

R (Å)

σ2 (Å2)

CN

R (Å)

σ2 (Å2)

sample

CN

R (Å)

Na2MoO4‚2H2O (NH4)2MoS4 MoS2 Mo-Fe-S cubaneb 12 µmol/g MoO42-, 0.1 mM total sulfide 3 µmol/g MoO42-, 5 mM total sulfide 15 µmol/g MoS42-, 0.1 mM total sulfide 12 µmol/g MoS42-, 2 mM total sulfide 5 µmol/g MoS42-, 5 mM total sulfide

4d -c 4d

1.78 1.76

0.003 0.008

4d 6d 4 -

2.20 2.43 2.35 -

0.005 0.004 0.005 -

6d,e 3 1

3.18e 2.73 2.81

0.004e 0.006 0.008

4d

1.78

0.007

-

-

-

-

-

-

1d

1.77

0.003

3e

2.39

0.005

3e

2.72

0.010

1d

1.76

0.003

3e

2.41

0.006

3e

2.68

0.011

0.1

1.77

0.003

4.0

2.40

0.006

3e

2.70

0.016

a The coordination number (CN) is typically accurate to within (1, interatomic distance (R) within (0.03 Å; the Debye-Waller factor (σ2) represents the variance in R. b Determined by Cramer et al. (1978). Other cubanes have similar structures. c Not detected. d Fixed during fitting. e Mo-Mo shell.

although it is unknown whether adsorbed Mo undergoes intermediate transformations over longer, geologic time scales. Nevertheless, our results show a plausible mechanism for the accumulation of MoS42- in preference to MoO42within anoxic sediments and soils.

Acknowledgments

molybdate (eq 6) occurs at a threshold HS- activity, ∼1 mM when buffered with seawater (19, 46).

The authors thank the anonymous reviewers, as well as the associate editor, James F. Pankow, for their editorial comments and suggestions, which improved the manuscript considerably. We also acknowledge financial support for this research from the National Science Foundation (NSFCRAEMS Grant CHE-0089215) and the Department of Energy. X-ray absorption spectroscopy was performed at the Stanford Synchrotron Radiation Laboratory, a facility operated by the Department of Energy Office of Basic Energy Sciences.

MoO42- + xHS - S MoO4 - xSx2- + xOH-, 1 < x < 4 (6)

Literature Cited

Molybdate and iron (hydr)oxides are stable in oxic environments, react reversibly, and thus are not retained permanently in most sediments. Under conditions in which molybdate and FeS2 persist, molybdate will react with sediments weakly. Molybdate and pyrite may coexist in environments in which the total sulfide concentration is measurable but below the pH-dependent threshold, ∼1 mM in seawater (19, 46), and in transient environments in which molybdate is not converted to MoS42- due to slow reaction (11). As a consequence of the weak interaction of molybdate and pyrite, Mo will quickly desorb as conditions change, thereby preventing appreciable Mo enrichment in suboxic environments.Asignificantbodyofevidencesupportsthishypothesiss Mo enrichment in sediments such as the Saanich Inlet is less than observed for other elements in the same system or Mo enrichment in more reduced systems (e.g., refs 9, 10, and 47). Under highly sulfidic zones, sulfide concentrations commonly increase above the threshold activity to convert MoO42into MoS42- (19, 48) which, on the basis of this research, react much more strongly with FeS2. Once MoS42- reacts with pyrite, it forms stable complexes that are irreversibly retained as Mo-Fe-S cubanes. Similar Mo cubanes have been identified in sulfur-rich black shales (19) but were not identified in sulfur-deficient shales; Mo was not enriched and was present as molybdate species. Our results provide a rapid mechanism of forming Mo-Fe-S cubanes observed in the sedimentary record where Mo enrichment occurs,

(1) Goldberg, S.; Forster, H. S.; Godfrey C. L. Soil Sci. Soc. Am. J. 1996, 60, 425. (2) Rietre, R. P. J. J.; Hiemstra, T.; van Riemsdijk, W. H. Geochim. Cosmochim. Acta 1999, 63, 3009. (3) Wu, C. H.; Lin, C. F.; Lo, S. L.; Yasunaga, T. J. Colloid Interface Sci. 1998, 208, 430. (4) Bourikas, K.; Hiemstra, T.; Van Riemsdijk, W. H. J. Phys. Chem. B 2001, 105, 2393-2403. (5) Bertine, K. K.; Turekian, K. K. Geochim. Cosmochim. Acta 1973, 37, 1415. (6) Calvert, S. E.; Pedersen, T. F. Mar. Geol. 1993, 113, 67. (7) Emerson, S. R.; Huested, S. S. Mar. Chem. 1991, 34, 177-196. (8) Amrhein, C.; Mosher, P. A.; Brown, A. D. Soil Sci. 1993, 155, 249. (9) Crusius, J.; Calvert, S.; Pedersen, T.; Sage, D. Earth Planet. Sci. Lett. 1996, 145, 65. (10) Morford, J. L.; Emerson, S. Geochim. Cosmochim. Acta 1999, 63, 1735. (11) Erickson, B. E.; Helz, G. R. Geochim. Cosmochim. Acta 2000, 64, 1149. (12) Johannesson, K. H.; Lyons, W. B.; Graham, E. Y.; Welch, K. A. Aquat. Geochem. 2000, 6, 19. (13) van den Berg, C. M. G. Estuaries 1993, 16, 512. (14) Vitousek, P. M.; Howarth, R. W. Biogeochemistry 1991, 13, 87. (15) Howarth, R. W.; Chan, F.; Marino, R. Biogeochem. 1999, 46, 203. (16) Paerl, H. W.; Prufertbebout, L. E.; Guo, C. Z. Appl. Environ. Microbiol. 1994, 60, 1044. (17) Shaw, T. J.; Gieskes, J. M.; Jahnke, R. A. Geochim. Cosmochim. Acta 1990, 54, 1233. (18) Adelson, J. M.; Helz, G. R.; Miller, C. V. Geochim. Cosmochim. Acta 2001, 65, 237. (19) Helz, G. R.; Miller, C. V.; Charnock, J. M.; Mosselmans, J. F. W.; Pattrick, R. A. D.; Garner, C. D.; Vaughn, D. J. Geochim. Cosmochim. Acta 1996, 60, 3631.

FIGURE 7. Structural models of MoO42- (a) and MoS42- (b) adsorption on FeS2. Molybdate forms a bidentate, mononuclear complex with an Fe polyhedron, and MoS42- forms a Mo-Fe-S cubane structure.

290

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

(20) Choppin, G. R.; Wong, P. J. Aquat. Geochem. 1998, 4, 77. (21) Bertine, K. K. Mezhdunar. Geokhim. Kongr., Dokl. 1st 1973, 4 (Pt. 1), 348. (22) Thomson, J.; Jarvis, I.; Green, D. R. H.; Green, D. A.; Clayton, T. Geochim. Cosmochim. Acta 1998, 62, 643. (23) Ressler, T. J. Phys. IV 1997, 7, C269. (24) Rehr, J. J.; Muestre de Leon, J.; Zabinski S. I.; Albers, R. C. Phys. Rev. B 1991, 44, 4146. (25) Zabinsky, S. I.; Rehr, J. J.; Ankudinov, A.; Albers, R. C.; Eller, M. J. Phys. Rev. B 1995, 52, 2995. (26) Ankudinov, A. L.; Ravel, B.; Rehr, J. J.; Conradson, S. D. Phys. Rev. B 1998, 58, 7565-7576. (27) Guevremont, J. M.; Strongin, D. R.; Schoonen, M. A. A. Surf. Sci. 1997, 391, 109. (28) Guevremont, J. M.; Strongin, D. R.; Schoonen, M. A. A. Am. Mineral. 1998, 83, 1246. (29) Bostick, B. C.; Fendorf, S.; Fendorf, M. Geochim. Cosmochim. Acta 2000, 64, 247. (30) Sasaki, K. Geochim. Cosmochim. Acta 1994, 58, 4649. (31) Martello, D. V.; Vecchio, K. S.; Diehl, J. R.; Graham, R. A.; Tamilia, J. P.; Pollack, S. S. Geochim. Cosmochim. Acta 1994, 58, 4657. (32) Rosso, K. M.; Becker, U.; Hochella, M. F., Jr. Am. Mineral. 1999, 84, 1535. (33) Bertine, K. K. Mar. Chem. 1972, 1, 43. (34) Bibak, A.; Borggard, O. K. Soil Sci. 1994, 158, 323. (35) Bebie, J.; Schoonen, M. A. A.; Fuhrmann, M.; Strongin, D. R. Geochim. Cosmochim. Acta 1998, 62, 633. (36) Bare, S. R.; Mitchell, G. E.; Maj, J. J.; Vrieland, G. E.; Gland, J. L. J. Phys. Chem. 1993, 97, 6048.

(37) Takenaka, S.; Tanaka, T.; Funabiki, T.; Yoshida S. J. Phys. Chem. B 1998, 102, 2960. (38) Hu, H. C.; Wachs, I. E.; Bare, S. R. J. Phys. Chem. 1995, 99, 10897-10910. (39) Christou, G.; Garner, C. D.; Mabb, F. E. J. Chem. Soc., Chem. Commun. 1978, 1978, 740. (40) Garner, C. D. In Comprehensive coordination chemistry; Wilkinson, G., Gillard, R. D., McCleverty, J. A., Eds.; Pergamon Press: New York, 1987; Chapter 36.6. (41) Chan, M. K.; Kim J.; Rees, D. C. Science 1993, 260, 792. (42) Ga¨rd, R.; Sun, Z.-X.; Forsling, W. J. Colloid Interface Sci. 1995, 169, 393. (43) Balsley, S. D.; Brady, P. V.; Krumhansl, J. L.; Anderson, H. L. Environ. Sci. Technol. 1996, 30, 3025. (44) Stumm, W. Adv. Chem. Ser. 1995, 244, 1. (45) Clarke, M. B.; Helz, G. R. Environ. Sci. Technol. 2000, 34, 1477. (46) Brule, J. E.; Hayden, Y. T.; Callahan, K. P.; Edwards, J. O. Gazz. Chim. Ital. 1988, 118, 93. (47) Francois, R. Mar. Geol. 1988, 83, 285. (48) Luther, G. W., III; Church, T. M.; Scudlark, J. R.; Cosman, M. Science 1986, 232, 746-749.

Received for review April 24, 2002. Revised manuscript received September 30, 2002. Accepted October 15, 2002. ES0257467

VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

291