Thermodynamic Stabilization of Hydrous Ferric Oxide by Adsorption of

May 10, 2011 - developed.1 Natural HFO is usually referred to as ferrihydrite, a .... ¼ 20.92H2O ю MgSO4 ю 2KCl ... 0 γ-FeOOH. 3 0.162H2O (cr) ю ...
0 downloads 0 Views 828KB Size
ARTICLE pubs.acs.org/est

Thermodynamic Stabilization of Hydrous Ferric Oxide by Adsorption of Phosphate and Arsenate Juraj Majzlan* Institute of Geosciences, Burgweg 11, Friedrich-Schiller University, D-07749 Jena, Germany

bS Supporting Information ABSTRACT: Hydrous ferric oxide (HFO) is an X-ray amorphous compound with a high affinity for anions under strongly or mildly acidic conditions. Because of the usually small particle size of HFO, the adsorption capacity is high and adsorption may significantly impact the thermodynamic properties of such materials. Here we show that adsorption of phosphate and arsenate stabilizes HFO by experimental determination of enthalpies of formation (by acid-solution calorimetry) and estimates of standard entropies for six phosphate- or arsenateenriched HFO samples. At pH values lower than ∼5, the phosphate-doped HFO is not only less soluble than ferrihydrite (anion-free HFO) but also crystalline FeOOH polymorphs feroxyhyte and lepidocrocite. The arsenate-doped HFO is also stabilized with respect to the ferrihydrite. Phosphate availability in soils can be controlled by the phosphate-enriched HFO which is many orders of magnitude less soluble than apatite or crystalline Fe(III) phosphates, for example strengite (FePO4 3 2H2O). Thermodynamic dissolution models for scorodite (FeAsO4 3 2H2O) and As-enriched HFO show that under mildly acidic or circumneutral conditions, scorodite dissolves, As-HFO precipitates, and a substantial amount of As(V) is released into the aqueous solution (at pH 7, log m(As) ∼ 2.5). The data presented in this paper can be used to model the equilibrium concentration of Fe(III), P(V), or As(V) in soil solutions or in natural or anthropogenic sediments polluted by arsenic.

1. INTRODUCTION Hydrous ferric oxides (HFO) are a family of X-ray amorphous substances whose physical (e.g., surface area) and chemical (e.g., water content) properties vary widely. The lack of a well-defined periodic structure makes the characterization and comparison of HFO samples difficult, although structural models are being developed.1 Natural HFO is usually referred to as ferrihydrite, a ferric oxyhydroxide mineraloid with small particle size (in the range of 220 nm), variable and uncertain chemical composition, physicochemical properties, and structure. The study of ferrihydrite, no matter how arduous and challenging, is often undertaken because of the omnipresence and fundamental importance of this mineraloid in many natural processes at or near the surface of the Earth. The properties, be they physical, chemical, or structural, of ferrihydrite can be further modified by adsorption of phosphate or silicate2 or by embedding ferrihydrite particles in organic polymers.3 In this contribution, we reserve the term ferrihydrite for HFO free of foreign anions. X-ray amorphous iron oxides with adsorbed anions will be referred to as anion-enriched HFO. Crystalline phases may dissolve and transform into shortrange ordered Fe(III) precipitates (HFO) under certain conditions, as reported for strengite reacted with excess Fe(III) at slightly acidic to neutral pH4 and for scorodite above pH 2.5 Hydrous ferric oxide also precipitates very often directly from aqueous solutions and under mildly or strongly acidic conditions, ferrihydrite is a ferocious adsorber of a wide range of anions. r 2011 American Chemical Society

The adsorption of the anions onto the surface of hydrous ferric oxide alters the interfacial free energy of these particles6,7 and therefore also their solubility and the overall thermodynamic properties of the anion-enriched HFO. The interfacial energy (or surface energy) of several simple oxides has been measured (e.g., for iron oxides8 and titanium oxides9) and the energetic effect of the surface of nanoparticles can amount to as much as 1015 kJ/mol.7 Because small nanoparticles have a very large fraction of surface atoms, another line of thought may be adopted, namely that the “surface complexation occurs at the bulk of the nanoparticles, as in a solid solution”.10 The notion of bulk reaction or “solid solution” is supported by the recent work of Voegelin et al.2 who used extended X-ray absorption fine-structure (EXAFS) spectroscopy to claim that the structure of the phosphate-doped HFO is different from that of ferrihydrite. Thibault et al.11 invoked explicitly a “solid solution” between ferrihydrite and an end-member phosphate-rich HFO with a limiting P:Fe ratio of 1:2. In our recent work on natural HFO doped with arsenate, we have observed essentially the same limiting ratio for As:Fe in arsenate-rich HFO.12 Received: December 1, 2010 Accepted: May 2, 2011 Revised: April 19, 2011 Published: May 10, 2011 4726

dx.doi.org/10.1021/es1040249 | Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology

ARTICLE

Table 1. Composition of the Samples Studied and Their Dissolution Enthalpiesa sample

chemical composition

molar mass (g 3 mol1)

AsADS1

0.2284Fe2O3 3 0.02742As2O5 3 0.7441H2O

56.1873

0.120

19.83b ( 0.23c (8)d

AsADS2

0.2035Fe2O3 3 0.03756As2O5 3 0.7590H2O 0.2065Fe2O3 3 0.04224As2O5 3 0.7513H2O

54.7968

0.185

16.90 ( 0.27 (8)

AsADS3

56.2139

0.205

16.28 ( 0.17 (4)

AsADS4

0.2090Fe2O3 3 0.04536As2O5 3 0.7456H2O

57.2389

0.217

15.82 ( 0.40 (5)

As/Fe (molar ratio)

ΔdissolutionH (kJ 3 mol1)

P/Fe PADS2 PADS3

0.5100Fe2O3 3 0.08367P2O5 3 0.4063H2O 0.4033Fe2O3 3 0.08193P2O5 3 0.5148H2O

100.644

0.164

37.23 ( 0.55 (5)

85.301

0.203

28.93 ( 0.38 (3)

a General composition of the samples is fFe2O3 3 xX2O5 3 hH2O (X = P,As). Samples were analyzed for Fe, As or P by wet chemical methods. The H2O content was calculated by difference to 100%. b average. c two standard deviations of the mean. d number of measurements.

In this contribution, we have used the assumption that the surface adsorption for nanoparticles equals to a bulk reaction and investigated the thermodynamics of a series of phosphate- and arsenate-doped HFO samples by acid-solution calorimetry. We derived their formation enthalpies, estimated standard entropies, calculated the Gibbs free energies of formation, and evaluated the thermodynamic effect of phosphate and arsenate adsorption onto HFO. Using simple thermodynamic models, we present the implications of the results for the availability of Fe, As, and P in pristine and polluted soils and sediments.

2. METHODS AND MATERIALS Ferrihydrite was synthesized by neutralization of a 1 M Fe(NO3)3 solution by 5 M NaOH solution. The end-point of the titration was the point of the fastest settling of the suspension when the aqueous solution remained visually clear. The pH of such end-point was approximately 7. The synthesized ferrihydrite was then washed repeatedly (at least six times) with copious amounts of deionized water and allowed to dry at room temperature. The resulting product was then finely ground in a agate mortar and pestle. The dry powder was redispersed in deionized water and predetermined amounts of phosphate (as a K2HPO4 solution) or arsenate (as As2O5 solution) were added. The suspensions were agitated for 2 h, then centrifuged. The aqueous solution was decanted and the solid fraction was dried and gently ground. The samples were analyzed for their chemical composition after dissolution in distilled 37% HCl. The analyses were carried out after appropriate dilution by an atomic absorption spectrometer (AAS) Analytik Jena Vario 6 (Fe, As) or inductively coupled plasma optical emission spectrometer (ICP-OES) Spectroflame (Spectro) (Fe, As, P). Both KH2AsO4 and KH2PO4 were purchased from commercial suppliers of chemicals, dissolved in water, and allowed to crystallize slowly from supersaturated solutions at room temperature. The colorless and transparent crystals of both phases (up to 1 cm) were broken, crushed, and used for further characterization and calorimetry. Lepidocrocite (γ-FeOOH) was synthesized as previously described.13 KCl was purchased from a manufacturer and treated at 800 K before the calorimetric measurements. The X-ray diffraction patterns of the samples and reference materials were collected with a Bruker D8 Advance diffractometer. The instrument was equipped with a Cu KR X-ray tube, secondary graphite monochromator and a scintillation detector. For the calorimetric experiments, we used a commercial IMC-4400 isothermal microcalorimeter (Calorimetry Sciences

Figure 1. Enthalpies of dissolution (in J/g) of the studied samples (squares, arsenate-HFO; diamonds, phosphate-HFO) and ferrihydrite (circle) and schwertmannite (stars). Data for the latter two compounds taken from ref 19.

Corporation) which we modified for the purposes of acidsolution calorimetry as described in ref 14. The thermodynamic calculations were done with the program PHREEQC.15 The database was augmented with the data for iron oxides minerals (see ref 8 for a review), the values for scorodite (FeAsO4 3 2H2O)16 and aqueous arsenate and arsenite complexes (ref 16 in their Table 3, values in parentheses taken from ref 17). The data for strengite (FePO4 3 2H2O) were taken from ref 18.

3. RESULTS Chemical composition of the anion-enriched HFO samples was recalculated to a formula fFe2O3 3 xX2O5 3 hH2O (X = P,As), where f þ x þ h = 1. The resulting formulas and the corresponding molar masses used in the calculations are listed in Table 1. All HFO samples are X-ray amorphous, with only two broad maxima instead of sharp X-ray diffraction peaks in the angular range used for the XRD measurements (580°2θ). The XRD patterns and the positions of the broad maxima are given in Supporting Information (SI) Figure S1. 3.1. Energetics of the HFO with Adsorbed Oxyanions. Adsorption of the oxyanions leads to a significant stabilization of the HFO. Because these poorly crystalline substances have no stoichiometric or nominal formulas which would allow an easy comparison of the measured values, we use the dissolution enthalpies normalized to the mass of the sample (data in J/g) as a rough measure of their energetics. The dissolution enthalpy of ferrihydrite is much more exothermic than those for the 4727

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology

ARTICLE

Table 2. Thermodynamic Cycle for the Sample PADS2a ΔH (kJ 3 mol1)

reaction and reaction number

0.54

H2O (l) = H2O (aq)

1

0.51Fe2O3 3 0.08367P2O5 3 0.4063H2O (s) þ 2.55798Hþ (aq) = 1.02Fe3þ (aq) þ 0.16734PO43 (aq) þ 1.68529H2O (aq)

2

37.23 ( 0.55

KCl (cr) = Kþ (aq) þ Cl (aq)

3

17.85 ( 0.11

HCl 3 9.96H2O (l) = Hþ (aq) þ Cl (aq) þ 9.96H2O (aq)

4

0

γ-FeOOH 3 0.162H2O (cr) þ 3Hþ (aq) = Fe3þ (aq) þ 2.162H2O (aq)

5

46.62 ( 0.13

KH2PO4 (cr) = Kþ (aq) þ 2Hþ (aq) þ PO43 (aq)

6

24.61 ( 0.46

γ-FeOOH 3 0.162H2O (cr) = γ-FeOOH (cr) þ 0.162H2O (l)

7

1.41 ( 0.17

2.02142H2O (l) þ (0.51Fe2O3 3 0.08367P2O5 3 0.4063H2O) (s) þ 0.16734KCl (cr) = 0.16734 HCl 3 9.96H2O (l) þ 1.02γ-FeOOH (cr) þ 0.16734KH2PO4 (cr)

8

ΔH8 = 2.18666ΔH1 þ ΔH2 þ 0.16734(ΔH3  ΔH4  ΔH6) þ 1.02(ΔH7  ΔH5)

9.45 ( 0.60

H2 (g) þ 1/2O2 (g) = H2O (l)

9

285.8 ( 0.1

Fe (cr) þ O2 (g) þ 1/2H2 (g) = γ-FeOOH (cr)

10

549.4 ( 1.4

K (cr) þ 1/2Cl2 (g) = KCl (cr)

11

436.5 ( 0.1

10.46H2 (g) þ 4.98O2 (g) þ 1/2Cl2 (g) = HCl 3 9.96H2O (l) K (cr) þ H2 (g) þ P (cr) þ 2O2 (g) = KH2PO4 (cr)

12

3007.9 ( 1.0

13

1573.6 ( 1.0

1.02Fe (cr) þ 0.16734P (cr) þ 1.177325O2 (g) þ 0.4063H2 (g) = 0.51Fe2O3 3 0.08367P2O5 3 0.4063H2O (s) ΔH14 = ΔfHo (PADS2) = ΔH8  2.02142ΔH9  0.16734(ΔH11  ΔH12  ΔH13) þ 1.02ΔH10

14 685.7 ( 1.6

The calculation of ΔH8 is an unnecessary intermediate step which is shown here only to simplify the final equation for ΔfHo (PADS2). Abbreviations: g = gas, l = liquid, s = solid, cr = crystalline. Thermodynamic cycle for the sample AsADS1 and a generic cycle for all samples in this study are given in SI Table S2 and S3, respectively. The enthalpy values from the right-hand side column of this table are given in SI Table S1 with the corresponding references or notes. a

phosphate- or arsenate-enriched HFO (Figure 1). In other words, more energy is released upon dissolution of ferrihydrite than arsenate- or phosphate-HFO, meaning that ferrihydrite is located higher on an energy scale and is therefore the least stable substance. Even the adsorption of relatively small amount of phosphate or arsenate leads to a substantial stabilization. For comparison, the data for schwertmannite19 are also shown in Figure 1. Schwertmannite is a poorly crystalline iron sulfateoxyhydroxide whose structure may be related to the crystal structure of akaganeite.20 The enthalpies of formation of our samples were derived from the measured enthalpies of dissolution in 5 N HCl by the appropriate thermochemical cycles. The enthalpies of dissolution of the HFO samples are in Table 1. The enthalpies of dilution, dissolution, desorption, and formation used in the cycles are listed in Table 2. An example of a thermochemical cycle for the sample PADS2 is given in Table 2. All other cycles with their reactions are listed in the electronic supplement to this paper. The formation enthalpies for all samples studied are listed in Table 3. The reliability of the values which are used as input for the thermochemical cycles can be checked by other, appropriately constructed thermochemical cycles. We have tested the accuracy of the enthalpies of dissolution and formation for H2O, KCl, and HCl 3 9.96H2O by measuring the enthalpy of the reaction 2HCl 3 9:96H2 O þ K 2 SO4 þ MgO ¼ 20:92H2 O þ MgSO4 þ 2KCl by stepwise measurements of the enthalpies of dissolution for the individual phases. The resulting ΔrH = 84.29 ( 1.39 kJ 3 mol1 compares very well to the ΔrH = 85.66 ( 3.00 kJ 3 mol1 calculated from the published enthalpies of formation for the phases in this reaction. From all the phases used as reference compounds, KH2PO4 and KH2AsO4 and their enthalpies of formation are the least certain ones. We determined the formation

Table 3. Thermodynamic Data for the Studied Samplesa sample AsADS1 AsADS2 AsADS3 AsADS4 PADS2 PADS3

ΔfHo S° (kJ 3 mol1) (J 3 mol1 3 K1) 426.2 ( 0.7 420.1 ( 1.1 425.6 ( 1.0 429.6 ( 1.2 685.7 ( 1.6 626.8 ( 1.7

65.3 65.0 65.5 65.8 84.6 78.4

Δ f Go (kJ 3 mol1)

log K

ΔrHo (kJ 3 mol1)

364.4 ( 1.2 358.2 ( 1.4 363.2 ( 1.4 366.9 ( 1.5 612.8 ( 2.0 556.3 ( 2.0

0.331 0.842 1.170 1.411 4.294 4.220

30.44 26.32 25.81 25.45 61.84 46.63

auxiliary data 3þ

Fe PO43 AsO43 H2O

reference 49.9 1284.4 890.21 285.83

16.7 1025.491 646.36 237.14

18 21 17 21

a

Enthalpies of formation were derived from the appropriate thermochemical cycles. Standard entropies were estimated as described in the text. Gibbs free energies were calculated by combining the formation enthalpies and the standard entropies. The log K and the ΔrHo values refer to the reaction f Fe2 O3 3 xX2 O5 3 hH2 O þ ð6f  6xÞH þ ¼ 2f Fe3 þ þ 2xXO4 3  þ ð3f þ h  3xÞH2 O where X = P, As. The auxiliary data used in calculating the log K and ΔrHo for this reaction are given in the bottom portion of the table.

enthalpies of these two phases by measuring the heats of dissolution at infinite dilution and combining these values with the formation enthalpies of the appropriate aqueous ions at infinite dilution. The results and their comparison with the literature values are elaborated in the Supporting Information. 3.2. Entropy Estimates. Using measured entropies for a suite of ferric iron and aluminum oxides, Majzlan et al.19 have estimated the standard entropy of ferrihydrite. We note that this entropy estimate, when combined with the experimental 4728

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology

ARTICLE

enthalpy of formation, leads to a reasonable Gibbs free energy of formation and solubility product, when compared to earlier22 or later23 works. The entropy estimates for the arsenate- or phosphate-HFO are more difficult to make. If these samples were crystalline, a first-order estimate could be made as a sum of the components, usually oxides (Kopp’s rule). The procedure is complicated because of the poor crystallinity of the studied phases which must be taken into account. We make an arbitrary assumption that the percentual increase of entropy for each component, when transforming that component from a well to a poorly crystalline state, is the same. Therefore, using our previous estimate for ferrihydrite as Fe(OH)3, we can calculate that percentual increase, taking goethite (R-FeOOH) and hexagonal ice (Ih) as the components. The standard entropy at T = 298.15 K (So) of goethite is 59.69 J 3 mol1 3 K1, that of hexagonal ice is 41.94 J 3 mol1 3 K1 (cf. ref 13). Consider reaction FeðOHÞ3 ðferrihydriteÞ ¼ R  FeOOHðcrystallineÞ þ H2 OðIh, crystallineÞ with the corresponding equation which describes the increase of the entropy upon transformation from well to poorly crystalline state, the variable x ð1 þ xÞSo ðR  FeOOHÞ þ ð1 þ xÞSo ðIhÞ ¼ So ðferrihydriteÞ or x ¼ ½So ðfhÞ  So ðR  FeOOHÞ  So ðIhÞ= ½So ðR  FeOOHÞ þ So ðIhÞ The So for Fe(OH)3 ferrihydrite is bracketed by values of 122 and 135 J 3 mol1 3 K1.19 The value of x for these two estimates is 0.20 and 0.32, respectively, meaning that the entropy of hypothetical disordered components is 2032% larger than the entropy of the real, well crystalline components. The entropies of the studied samples were then estimated as a sum of the entropies of the appropriate components (hexagonal ice, goethite, As2O5, P2O5) increased by 26%. The estimates are listed in Table 3.

4. DISCUSSION The measured formation enthalpies and estimated entropies can be combined to calculate the Gibbs free energies of formation for the studied samples (Table 3). The variable composition of the studied samples precludes a simple comparison of their stability, either among themselves or with respect to ferrihydrite or the crystalline iron oxides, for example goethite (R-FeOOH). To facilitate such comparison, we have calculated the solubility of all these compounds over a wide range of pH values (Figure 2). As expected, the lowest solubility is that of goethite, the most stable phase in the system Fe2O3H2O together with hematite (R-Fe2O3). Solubility of lepidocrocite (γ-FeOOH) and feroxyhyte (δ0 -FeOOH), minerals metastable with respect to goethite, is higher, but still lower than the solubility of ferrihydrite, an X-ray amorphous compound. Doping of ferrihydrite with phosphate stabilizes this compound so that it is much less soluble than ferrihydrite (Figure 2a). At pH values lower than ∼5, the phosphate-doped HFO is even less soluble than feroxyhyte and lepidocrocite. The arsenatedoped HFO is also stabilized with respect to the ferrihydrite; one of the samples (AsADS1) is less soluble than ferrihydrite in a wide range of pH values (less than 8), the other ones by pH less

Figure 2. Solubility curves for crystalline (goethite, lepidocrocite, feroxyhyte) and amorphous (ferrihydrite) iron oxides, together with (a) curves for phosphate-enriched HFO, strengite, and schwertmannite, and (b) curves for arsenate-enriched HFO and scorodite. Sources of the data are listed in text. All calculations done by PHREEQC.

than 5 (Figure 2b). The larger stabilization of ferrihydrite by the adsorption of phosphate than by the adsorption of arsenate agrees well with the stronger adsorption of phosphate than arsenate observed in bulk studies2426 and theoretical prediction of the stability of HFO nanoparticles with adsorbed phosphate or arsenate.10 For comparison, the solubility curves for the mineral schwertmannite, a poorly ordered ferric sulfate, are also plotted in Figure 2a. This mineral probably deviates structurally from ferrihydrite; the early assumption that it may be related to akaganeite20 is being adhered to in the later literature but a satisfactory proof is missing. The adsorption of sulfate is weaker than that of phosphate or arsenate. Despite that, the arsenatedoped HFO samples are comparable in their solubility with schwertmannite. The additional stabilization of schwertmannite may come from the partially ordered structure which is manifested by more well-defined peaks (humps) in the X-ray diffraction pattern. One should note that all solubility curves presented in Figures 24 were calculated for the given solids and water with no additional ligands. Under natural conditions, ligands such as small organic molecules (oxalate, citrate) or siderophores may be present (e.g., refs 27 and 28) and increase the concentration of Fe(III) in the aqueous solutions. This fact should be kept in mind when attempting a stringent and critical comparison of thermodynamic predictions with field observations. In many soils, phosphate availability is limited in the horizons where Fe and Al oxides are abundant (e.g., ref 29) although the dissolved organic phosphorus may represent a significant fraction of the total P pool (e.g., ref 30). Among the inorganic pool of phosphate, the phosphate-enriched HFO is the sink of the 4729

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology

Figure 3. Equilibrium concentrations of PO4,total in aqueous solutions in contact with phosphate-enriched HFO, strengite, and hydroxylapatite. Thermodynamic data for hydroxylapatite were taken from the phreeqc.dat database. All calculations done by PHREEQC.

Figure 4. Equilibrium concentrations of Fe(III)total and As(V)total for four different models. The curves are labeled with the number of the model and the element whose concentration the curve represents. Model 1: congruent dissolution of scorodite. Model 2: congruent dissolution of scorodite and precipitation of ferrihydrite. Model 3: congruent dissolution of scorodite and precipitation of arsenate-enriched HFO (sample AsADS1). Model 4: congruent dissolution of arsenate-enriched HFO (sample AsADS1) and precipitation of scorodite. Details in text. All calculations done by PHREEQC.

phosphate released by the weathering of primary minerals (especially apatite from magmatic and metamorphic rocks) or crystalline iron phosphates (e.g., strengite) (Figure 3). Furthermore, the phosphate-enriched HFO determines the amount of Fe accessible to the soil biota. At moderate loading of the surface of HFO with phosphate, at P/Fe (molar ratio) of 0.1, the equilibrium aqueous Fe(III) concentration for phosphate-HFO is about 2 orders of magnitude lower than the equilibrium concentration for ferrihydrite (Figure 2a). Thus, modeling the biota-available Fe(III) with ferrihydrite can lead to substantial errors. The dissolution of scorodite as a function of pH was extensively reviewed in a recent work16 and new studies on the topic appeared since.31,32 They all showed that above pH ∼2.5, scorodite dissolves incongruently and the Fe/As ratio in the solution is much larger than 1, the Fe:As ratio in scorodite. Although most of the studies did not explicitly identify HFO as the precipitate, they did not report a crystalline phase as the Fe sink. Therefore, one must assume that the precipitate is HFO. The same reaction pathway of scorodite dissolution and HFO precipitation can be modeled by consideration of arsenate-enriched HFO samples used in this study. The solubility curve for

ARTICLE

scorodite is depicted in Figure 4, together with the equilibrium solubilities in a system where scorodite is dissolving and ferrihydrite is allowed to precipitate, and solubilities in a system where scorodite dissolves and arsenate-enriched HFO is allowed to precipitate. Under strongly acidic conditions, scorodite dissolves congruently, meaning that the molalities of Fe(III) and As(V) are equal in the solution, and HFO does not precipitate. At a certain pH value, ferrihydrite begins to precipitate (model 2 in Figure 4), withdrawing Fe(III) from the solution, and hence making the solution undersaturated with respect to scorodite and forcing more scorodite to dissolve. The two processes, dissolution of scorodite and simultaneous precipitation of ferrihydrite cause drop in the aqueous Fe(III) concentration and increase in the As(V) concentration. Qualitatively similar trend can be observed if As-enriched HFO is allowed to precipitate instead of ferrihydrite (model 3 in Figure 4). Counterintuitively, the aqueous As(V) concentration is higher than in the previous model, although the precipitate contains As. This paradox can be explained by the greater stability (lower solubility) of the AsHFO than ferrihydrite. As-HFO is thus able to withdraw more Fe(III) from the aqueous solution, thus forcing more scorodite to dissolve. At the same time, because the As/Fe ratio in our samples of the As-enriched HFO is relatively low (0.10.2, cf. Table 1), precipitation of As-enriched HFO cannot compensate for the increased scorodite dissolution and the aqueous As(V) concentration rises. These models show that storing scoroditecontaining waste forms in mildly acidic or circumneutral aqueous environments can be very dangerous because much arsenic can be released under these conditions, as previously reported in refs 33 and 34. As already noted by Welham et al.,35 scoroditecontaining waste forms must be kept under acidic conditions or in solutions which high Fe(III) concentrations to prevent the buildup and release of aqueous As(V). A different scenario can be modeled if As-enriched HFO is allowed to dissolve and scorodite is allowed to precipitate (model 4 in Figure 4). Under mildly to moderately acidic conditions (pH > 3), the aqueous concentrations of Fe(III) and As(V) are controlled by As-HFO. Only in the strongly acidic region (pH < 3), scorodite begins to precipitate and draws a substantial amount of As(V) from the solution. These models, however, neglect the fact that the As-enriched HFO itself is metastable with respect to goethite and will slowly transform to this phase and release As(V) into the aqueous solution. The presented data clearly show that the solubility of anionenriched HFO differs significantly from the solubility of ferrihydrite. Adsorption of anions such as phosphate or arsenate modifies the thermodynamic properties of HFO and similar effects, probably smaller in magnitude, could be expected for the adsorption of silicate, perhaps also carbonate and organic matter onto ferrihydrite in natural systems. Hence, when modeling geochemical processes, the thermodynamic data for the anionfree ferrihydrite should be used as the limiting case for availability and equilibrium solubility of Fe(III). In addition, HFO can be a major sink for oxyanions and can control their availability, either for inorganic reactions or for biota, in natural systems.

’ ASSOCIATED CONTENT

bS

Supporting Information. Selection of the values of ΔfHo

for KH2PO4 and KH2AsO4 (p. S2). Enthalpy values used in the thermochemical cycles in this study (p. S3). Thermochemical cycle for the sample AsADS1 and a generic sample AsADS 4730

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology (pp. S4 and S5). X-ray diffraction patterns of the HFO samples (p. S6). This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT I thank M. Gomez and two anonymous reviewers for constructive criticism and R. Kretzschmar for the editorial handling of the manuscript. I am grateful to S. Hirth-Walther (Freiburg) and D. Merten (Jena) for the chemical analyses of the HFO samples. This work was financially supported by the Deutsche Forschungsgemeinschaft, Grant No. MA 3927/2-1. ’ REFERENCES (1) Michel, F. M.; Ehm, L.; Antao, S. M.; Lee, P. L.; Chupas, P. J.; Liu, G.; Strongin, D. R.; Schoonen, M. A. A.; Phillips, B. L.; Parise, J. B. The structure of ferrihydrite, a nanocrystalline material. Science 2007, 316, 1726–1729. (2) Voegelin, A.; Kaegi, R.; Frommer, J.; Vantelon, D.; Hug, S. J. Effect of P, Si, and Ca on Fe(III)-precipitates formed in aerated Fe(II) and As(III) containing water studied by X-ray absorption spectroscopy. Geochim. Cosmochim. Ac. 2010, 74, 164–186. (3) Mikutta, C.; Mikutta, R.; Bonneville, S.; Wagner, F.; Voegelin, A.; Christl, I.; Kretzschmar, R. Synthetic coprecipitates of exopolysaccharides and ferrihydrite. Part I: Characterization. Geochim. Cosmochim. Acta 2008, 72, 1111–1127. (4) McLaughlin, J. R.; Syers, J. K. Stability of ferric phosphates. J. Soil Sci. 1978, 29, 499–504. (5) Robins, R. G. Solubility and stability of scorodite, FeAsO4 3 2 H2O: Discussion. Am. Mineral. 1987, 72, 842–844. (6) Parks, G. A. Surface energy and adsorption at mineral/water interfaces: An introduction. Rev. Mineral. 1990, 23, 133–176. (7) Kulik, D. A. Thermodynamic concepts in modeling sorption at the mineral-water interface. Rev. Mineral. Geochem. 2009, 70, 125–180. (8) Navrotsky, A.; Mazeina, L.; Majzlan, J. Size-driven structural and thermodynamic complexity in iron oxides. Science 2008, 319, 1635–1638. (9) Ranade, M. R.; Navrotsky, A.; Zhang, H. Z.; Banfield, J. F.; Elder, S. H.; Zaban, A.; Borse, P. H.; Kulkarni, S. K.; Doran, G. S.; Whitfield, H. J. Energetics of nanocrystalline TiO2. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6476–6481. (10) Fukushi, K.; Sato, T. Using a surface complexation model to predict the nature and stability of nanoparticles. Environ. Sci. Technol. 2005, 39, 1250–1256. (11) Thibault, P.-J.; Rancourt, D. G.; Evans, R. J.; Dutrizac, J. E. Mineralogical confirmation of a near-P:Fe = 1:2 limiting stoichiometric ratio in colloidal P-bearing ferrihydrite-like hydrous ferric oxide. Geochim. Cosmochim. Acta 2009, 73, 364–376. (12) Majzlan, J.; Lalinska, B.; Chovan, M.; Bl€ass, U.; Brecht, B.; G€ottlicher, J.; Steininger, R.; Hug, K.; Ziegler, S.; Gescher, J. A mineralogical, geochemical, and microbiogical assessment of the antimony- and arsenic-rich neutral mine drainage tailings near Pezinok, Slovakia. Am. Mineral. 2011, 96, 1–13. (13) Majzlan, J.; Lang, B. E.; Stevens, R.; Navrotsky, A.; Woodfield, B. F.; Boerio-Goates, J. Thermodynamics of iron oxides: Part I. Standard entropy and heat capacity of goethite (R-FeOOH), lepidocrocite (γFeOOH), and maghemite (γ-Fe2O3). Am. Mineral. 2003, 88, 846–854. (14) Ackermann, S.; Armbruster, T.; Lazic, B.; Doyle, S.; Grevel, K.-D.; Majzlan, J. Thermodynamic and crystallographic properties of kornelite (Fe2(SO4)3 3 ∼7.75H2O) and paracoquimbite (Fe2(SO4)3 3 9H2O). Am. Mineral. 2009, 94, 1620–1628.

ARTICLE

(15) Parkhurst, D. L.; Appelo, C. A. J. User’s Guide to PHREEQC (Version 2)—a Computer Program for Speciation, Batch-Reaction, OneDimensional Transport, And Inverse Geochemical Calculations; U.S. Geological Survey Water-Resources Investigation Report, 1999; 994259, 312 pp. (16) Langmuir, D.; Mahoney, J.; Rowson, J. Solubility products of amorphous ferric arsenate and crystalline scorodite (FeAsO4 3 2H2O) and their application to arsenic behavior in buried mine tailings. Geochim. Cosmochim. Acta 2006, 70, 2942–2956. (17) Nordstrom, D. K.; Archer, D. G. Arsenic Thermodynamic Data and Environmental Geochemistry. In Arsenic in Ground Water: Geochemistry and Occurrence; Welch, A. H., Stollenwerk, K. G., Eds.; Kluwer Academic Publishers: Norwell, MA2002, pp 126. (18) Robie, R. A.; Hemingway, B. S. Thermodynamic properties of minerals and related substances at 298.15 K and 1 bar (105 Pascals) and at higher temperatures. U.S. Geol. Surv. Bull. 1995, 2131, 461. (19) Majzlan, J.; Navrotsky, A.; Schwertmann, U. Thermodynamics of iron oxides. Part III. Enthalpies of formation and stability of ferrihydrite (∼Fe(OH)3), schwertmannite (∼FeO(OH)3/4(SO4)1/8), and ε-Fe2O3. Geochim. Cosmochim. Acta 2004, 68, 1049–1059. (20) Bigham, J. M.; Schwertmann, U.; Carlson, L.; Murad, E. A poorly crystallized oxyhydroxysulfate of iron formed by bacterial oxidation of Fe(II) in acid mine waters. Geochim. Cosmochim. Acta 1990, 54, 2743–2758. (21) Grenthe, I.; Fuger, J.; Konings, R. J. M.; Lemire, R. J.; Muller, A. B.; Nguyen-Trung, C.; Wanner, H. Chemical Thermodynamics of Uranium; Nuclear Energy Agency, OECD, Elsevier, 1992. (22) Nordstrom, D. K.; Plummer, L. N.; Langmuir, D.; Busenberg, E.; May, H. M.; Jones, B. F.; Parkhurst, D. L. Revised chemical equilibrium data for major water-mineral reactions and their limitations. In Chemical Modeling in Aqueous Systems II, Symposium Series 416; Melchior, D. D., Bassett, R. L., Eds.; American Chemistry Society: Washington. DC, 1990; pp 398413. (23) Stefansson, A. Iron(III) hydrolysis and solubility at 25 °C. Environ. Sci. Technol. 2007, 41, 6117–6123. (24) Liu, F.; de Cristofaro, A.; Violante, A. Effect of pH, phosphate and oxalate on the adsorption/desorption of arsenate on/from goethite. Soil Sci. 2001, 166, 197–208. (25) Pigna, M.; Krishnamurti, G. S. R.; Violante, A. Kinetics of arsenate sorption-desorption from metal oxides: Effect of residence time. Soil Sci. Soc. Am. J. 2006, 70, 2017–2027. (26) Rubinos, D.; Iglesias, L.; Devesa-Rey, R.; Díaz-Fierros, F.; Barral, M. T. Arsenic release from river sediments in a gold-mining area (Anllons River basin, Spain): effect of time, pH and phosphorous concentration. Eur. J. Mineral. 2010, 22, 665–678. (27) Kraemer, S. M. Iron oxide dissolution and solubility in the presence of siderophores. Aquat. Sci. 2004, 66, 3–18. (28) Dehner, C. A.; Awaya, J. D.; Maurice, P. A.; DuBois, J. L. Roles of siderophores, oxalate, and ascorbate in mobilization of iron from hematite by the aerobic bacterium Pseudomonas mendocina. Appl. Environ. Microbiol. 2010, 76, 2041–2048. (29) V€a€an€anen, R.; Hristov, J.; Tanskanen, N.; Hartikainen, H.; Nieminen, M.; Ilvesniemi, H. Phosphorus sorption properties in podzolic forest soils and soil solution phosphorus concentration in undisturbed and disturbed soil profiles. Boreal Environ. Res. 2008, 13, 553–567. (30) Ron Vaz, M. D.; Edwards, A. C.; Shand, C. A.; Cresser, M. S. Phosphorus fractions in soil solution: Influence of soil acidity and fertilizer additions. Plant Soil 1993, 148, 175–183. (31) Bluteau, M.-C.; Demopoulos, G. P. The incongruent dissolution of scorodite  Solubility, kinetics and mechanism. Hydrometallurgy 2007, 87, 163–177. (32) Paktunc, D.; Bruggeman, K. Solubility of nanocrystalline scorodite and amorphous ferric arsenate: Implications for stabilization of arsenic in mine wastes. Appl. Geochem. 2010, 25, 674–683. (33) Moldovan, B. J.; Jiang, D. T.; Hendry, M. J. Mineralogical characterization of arsenic in uranium mine tailings precipitated from iron-rich hydrometallurgical solutions. Environ. Sci. Technol. 2003, 37, 873–879. 4731

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732

Environmental Science & Technology

ARTICLE

(34) Harvey, M. C.; Schreiber, M. E.; Rimstidt, J. D.; Griffith, M. M. Scorodite dissolution kinetics: Implications for arsenic release. Environ. Sci. Technol. 2006, 40, 6709–6714. (35) Welham, N. J.; Malatt, K. A.; Vukcevic, S. The stability of iron phases presently used for disposal from metallurgical systems—A review. Miner. Eng. 2000, 13, 911–931.

4732

dx.doi.org/10.1021/es1040249 |Environ. Sci. Technol. 2011, 45, 4726–4732