American Mineralogist, Volume 93, pages 1641–1646, 2008

0003-004X/08/0010–1641$05.00/DOI: 10.2138/am.2008.2858 1641

Incorporation of molybdate anion into β-FeOOH

Dale Carroll anD William r. riChmonD*

Nanochemistry Research Institute and Parker Cooperative Research Centre for Integrated Hydrometallurgy Solutions, Curtin University of Technology, GPO Box U1987, Perth, Western Australia, 6845, Australia

abstraCt

An analog of β-FeOOH (akaganéite) containing structurally incorporated molybdate has been synthesized and the conditions for its formation have been studied in the presence of chloride, sulfate, and nitrate anions. The incorporation of molybdate into the β-FeOOH structure has been shown to greatly increase the stability of this phase with respect to its transformation to hematite at elevated temperature. Molybdate incorporation also has been shown to change the particle morphology, with increasing amounts of molybdate causing the rod-shaped nanoparticles to become progressively shorter until spherical nanocrystals are observed. The combined observations of increased stability, morphology changes, and measured levels of incorporated molybdenum suggest that although inclu-sion of molybdate anions into the β-FeOOH tunnel sites is possible, it is unlikely to be the only mode of molybdate incorporation.

Keywords: β-FeOOH, molybdate, anion incorporation, phase transformation

introDuCtion

The iron oxyhydroxide β-FeOOH occurs in nature as the mineral akaganéite. It has been long recognized as an iron oxide component of acid soils (MacKay 1962) and geothermal brines (Holm et al. 1983), and also has been identified as a cor-rosion product on iron meteorites (Buchwald and Clarke 1989). The presence of chloride ions was previously considered to be essential for formation of the β-FeOOH phase, and synthetic β-FeOOH is typically prepared by hydrolysis of acidic Fe3+ chloride solutions at elevated temperatures (Cornell and Schw-ertmann 2003).

Crystallographic studies of β-FeOOH have shown that it is isostructural with hollandite (Post and Buchwald 1991; Post et al. 2003). A key feature of its structure is that it contains double chains of edge-sharing octahedra that are linked by corners to form a framework with large tunnels having a cross section of about 0.5 nm2. Chloride ions typically occupy sites within these tunnels, although the tunnels are not fully occupied, and the extent of incorporation of Cl– appears to be related to the concentration of the starting FeCl3 solution from which the β-FeOOH is produced. The average occupancy for a range of β-FeOOH samples reported from both natural and synthetic sources is about 1.23 chloride ions per unit cell (Cornell and Schwertmann 2003), which corresponds to an average occupancy of 2/3 of the tunnel sites.

It has been shown that hydroxide, fluoride, and bromide ions can be exchanged with chloride in β-FeOOH (Cai et al. 2001), but the larger iodide and perchlorate ions cannot be accommodated in the tunnel sites (Paterson and Rahman 1984). It has been pro-posed that other anions may be incorporated into the β-FeOOH

structure, provided that the oxyhdroxide phase is formed in the presence of a sufficient concentration of the anion. An example is the iron oxyhydroxide known as schwertmannite, which forms as a hydrolysis product from Fe3+ sulfate solutions (Bigham et al. 1994). The structure of schwertmannite remains the subject of some debate (Loan et al. 2004), but it has been suggested that schwertmannite is structurally similar to β-FeOOH, and that sulfate anions in the tunnels share oxygen atoms with FeO6 octahedra of the tunnel wall, thereby causing some distortion of the structure (Bigham et al. 1994, 1996). There is also an ex-ample in the literature of a naturally occurring β-FeOOH mineral containing tungstate (Walenta 1982), and a report describing a poorly crystalline phase containing nitrate anion (Schwertmann et al. 1996). The formation of a molybdate analog of β-FeOOH, in which the molybdate anion appears to occupy the tunnel sites within the β-FeOOH lattice also has been reported previously in the literature (Linehan et al. 1997; Richmond et al. 2004). This phase will form readily in the absence of chloride ion and is considerably more stable than the chloride-containing form of β-FeOOH with regard to its transformation to hematite.

In this study, we seek to investigate the relative abilities of different anions to act as templates in the formation the β-FeOOH structure and, in particular, to more clearly define the conditions under which the molybdate-β-FeOOH phase will form. This phase appears to be considerably more stable than the chloride or sulfate forms of β-FeOOH and its increased stability, coupled with its high surface area, make it an interesting material for potential catalytic applications (Matson et al. 1994).

experimental methoDs

Effect of molybdate on kinetics of β-FeOOH formationA series of condensed ferrihydrite suspensions was prepared to study the effects

of molybdate anion on the rate of transformation of the ferrihydrite. A set of control * E-mail: w.richmond@curtin.edu.au

CARROLL AND RICHMOND: INCORPORATION OF MOLyBDATE ANION INTO β-FeOOH1642

samples containing ferric chloride alone, and another set containing added sodium molybdate, were prepared using the following general procedure.

In a 100 mL Nalgene screw-top plastic jar, an Fe3+ solution (25 mL, 2 mol/L, Fe3+) was added to sodium hydroxide (20 mL, 6 mol/L, NaOH). In the control samples this sodium hydroxide solution was pre-mixed with 8.5 mL of deionized water, and in the molybdate-containing samples, with 8.5 mL of 2.0 mol/L Na2MoO4

solution [Mo:(Fe + Mo) = 0.25]. A gel-like suspension formed immediately and this was mixed thoroughly by vigorous shaking of the jar. The pH was then adjusted to 2 (±0.1) with a few drops of sodium hydroxide and/or hydrochloric acid. The suspensions were then aged at 98 °C for various time intervals up to 7 days in a laboratory oven. On removal from the oven the suspensions were cooled and the precipitates thoroughly washed with water (twice) and ethanol (twice) in repeated settling/re-suspension cycles using a centrifuge, then dried in air at 55 °C for two days.

Effect of suspension pH and anions on molybdate incorporation

The general procedure described above was also used to prepare samples in experiments aimed at examining the effects of varying the amount of added molybdate anion on the products of hydrolysis of concentrated ferrihydrite gels under various conditions. The volumes of reagents used, and the compositions of the suspensions are summarized in Table 1. In one set of experiments, using ferric chloride as the starting Fe3+ salt, the effects of pH were investigated by adjusting the suspension pH to either pH = 2, 6, or 9 before aging at 98 °C for 24 h. A further set of experiments was carried out at pH = 2 to investigate the effect of the other anions in the suspension on molybdate incorporation. In these experiments, ferric sulfate and ferric nitrate solutions were used to prepare samples with the range of molybdate ion concentrations outlined in Table 1, and the resulting products were characterized for comparison with those formed from ferric chloride solutions.

Characterization of productsIron and molybdenum content of the iron oxyhydroxide products was de-

termined by ICP-AES analysis after digestion of the solids in concentrated HCl. X-ray diffraction was carried out at 298 K using CuKα radiation (40 kV, 30 mA) on a Seimens D500 diffractometer. The samples were step scanned from 3 to 70 °2θ, at 0.02° increments, using a counting time of 5 s per increment. Transmission

electron micrographs (TEM) were recorded on a JEOL 2011 TEM operating at 200 kV. To prepare specimens for TEM, samples were dispersed in de-ionized water with the aid of ultra-sonication, and then a drop was placed onto a conventional holey carbon-coated copper grid and dried in air.

results

Effect of molybdate on kinetics of β-FeOOH formationThe series of time-resolved powder XRD patterns shown in

Figure 1 demonstrates that incorporation of molybdate into the β-FeOOH structure results in a material that has a much higher stability with respect to the transformation to hematite, than the β-FeOOH phase that contains chloride ion alone. In Figure 1a, it can be seen that when the preparation is carried out in ferric chloride media, and in the absence of any added molybdate anion, the characteristic reflections for β-FeOOH are observed after only 2 min and are well established after a period of 24 h. Over longer reaction times however, this mixture transforms to hematite, as seen in the pattern measured after 7 days. In Figure 1b, we see the patterns obtained from a ferric chloride solution when the system contains molybdate anion at a Mo:Fe mole ratio of 0.25. In this case the formation of the β-FeOOH structure is slower; the initial 2-line ferrihydrite phase transforms to β-FeOOH over about 24 h, which then remains stable for a period of 7 days and no hematite is observed.

Effect of pH on molybdate incorporationTable 2 provides a summary of the product phases obtained

from hydrolysis of condensed ferrihydrite gels containing vary-ing amounts of added molybdate under conditions chosen to examine the effects of pH, and the effects of different anions in the starting Fe2+ solution. The influence of pH was studied with samples prepared from ferric chloride solutions and Table 2 reports the products obtained at pH = 9, 6, and 3. At pH = 9, the molybdate-free suspension is transformed to hematite within the 24 h timeframe of this experiment, but the addition of only a small amount of molybdate (2 mol%) is sufficient to prevent this transformation, and ferrihydrite is observed as the product phase at this and all higher molybdate concentrations. For samples

Table 1. Volumes of reagents used to prepare concentrated iron oxyhydroxide suspensions in the synthesis of molybdate β-FeOOH samples

Part 1V H2O V 2 mol/L Fe conc Mo conc Mo(mL) Na2MoO4 (mL) (mol/L) (mol/L) Mo + Fe in reaction mixture8.5 0 0.943 0.000 0.008 0.5 0.019 0.027.5 1.0 0.038 0.047.0 1.5 0.057 0.066.0 2.5 0.094 0.095.0 3.5 0.132 0.124.0 4.5 0.170 0.153.0 5.5 0.208 0.182.0 6.5 0.245 0.201.0 7.5 0.283 0.230 8.5 0.321 0.25

Part 2V H2O V 2 mol/L Fe conc Mo conc Mo(mL) Na2MoO4 (mL) (mol/L) (mol/L) Mo + Fe in reaction mixture21 4 0.714 0.114 0.1417 8 0.229 0.2414 11 0.314 0.3012 13 0.371 0.348 17 0.486 0.40Notes: Two different Fe3+ concentrations were used, as represented in the two parts to this table. Part 1 = 25 mL 2 mol/L Fe3+ solution, 20 mL 6 mol/L NaOH was made up to 53.5 mL with H2O and 2 mol/L Na2MoO4. Part 2 = 25 mL 2 mol/L Fe3+ solution, 20 mL 6 mol/L NaOH was made up to 70 mL with H2O and 2 mol/L Na2MoO4. Part 2, in order to achieve Mo:(Mo + Fe) ratios > 0.25 without exceeding the solubility limit for Na2MoO4, a lower final Fe3+ concentration had to be used.

Table 2. A summary of product phases obtained for samples pre-pared from different Fe3+ salts and under various condi-tions of suspension pH, showing the effects of increasing molybdate ion concentration in each case

Mo Mo + Fe in reaction FeCl3 Fe(NO3)3 Fe2(SO4)3

mixture pH = 9 pH = 6 pH = 2 pH = 2 pH = 2 0.00 Hematite Fh β-FeOOH Fh Na-jarosite 0.02 Fh 0.04 Na-jarosite + 0.06 β-FeOOH 0.09 β-FeOOH 0.12 β-FeOOH 0.14 0.15 0.18 0.20 Na-jarosite + 0.23 β-FeOOH 0.24 0.25 0.30 Fh Na-jarosite 0.34 0.40

CARROLL AND RICHMOND: INCORPORATION OF MOLyBDATE ANION INTO β-FeOOH 1643

formed at pH = 6, the powder XRD patterns of the products showed no indication of hematite formation, only the broad re-flections characteristic of 2-line ferrihydrite. At pH = 2, the XRD patterns of all the products matched that of β-FeOOH.

Curves shown in Figure 2 illustrate the incorporation of molybdate into iron oxyhydroxide products formed from ferric chloride solutions at the 3 different pHs. The molybdate levels measured here represent molybdate retained in the sample after the standard washing procedure described in the experimental section. Levels of adsorbed and/or incorporated molybdate are significantly higher in the products formed at pH = 6 than those formed at pH = 9, as we might expect on the basis of surface charge considerations (the isoelectric point for ferrihydrite is in the range pH = 6–8) (Cornell and Schwertmann 2003). These results show that the β-FeOOH phase will form at pH = 2 in the presence of molybdate, but that the molybdate ion does not tem-plate the formation of the β-FeOOH structure at higher pH. This suggests that the formation of the molybdate-β-FeOOH phase requires the presence of the same precursor polyions needed to form the chloride-based β-FeOOH phase.

Formation of the molybdate-β-FeOOH phase in the presence of other anions

To investigate the ability of molybdate to compete with other anions in the formation of β-FeOOH phases, forced hydrolysis of ferric ion solutions was carried out with various amounts of sodium molybdate, added to either ferric chloride, sulfate, or nitrate solutions, and the molybdate content of the resulting iron oxyhydroxides measured. The range of product phases observed in these experiments is presented in Table 2, and the combined data from elemental analysis of the products is presented in Figure 3, showing that molybdate incorporation in all systems follows the same general relationship.

When ferric nitrate is used as the starting solution for the

forced hydrolysis reaction, the presence of a small amount of molybdate anion (up to 9 mol%) inhibits the transformation of ferrihydrite to hematite, but no detectable β-FeOOH phase forms. This indicates that the nitrate anion on its own does not template the β-FeOOH structure. At molybdenum levels above 9 mol%, the molybdate-β-FeOOH phase is formed, but it can be seen that there is a discrete range of molybdate ion concentrations under which β-FeOOH is observed. At levels of added molybdate

Figure 1. (a) Partial powder XRD patterns of products isolated from a condensed ferrihydrite gel formed by hydrolysis of FeCl3 (no molybdate present). The gel was aged over a period of 7 days at 98 °C and pH = 2. The single sharp peak in two of the patterns (marked with an arrow) is due to residual sodium chloride in the samples. (b) Partial powder XRD patterns of hydrolysis products isolated from a condensed ferrihydrite gel formed by hydrolysis of Fe(NO3)3 containing molybdate (Mo:Fe mole ratio = 0.25) and aged under the same conditions as in a.

Figure 2. Relationship between molybdenum added to the reaction mixture and molybdenum incorporated into the product for iron oxyhydroxides formed by hydrolysis of ferric chloride at pH = 2, 6, and 9. Temperature = 100 °C, reaction time = 24 h.

a b

CARROLL AND RICHMOND: INCORPORATION OF MOLyBDATE ANION INTO β-FeOOH1644

greater than 25%, β-FeOOH no longer forms, and ferrihydrite is the only reaction product.

For reaction mixtures in which ferric sulfate is used as the source of iron, the threshold level of molybdate required to form the β-FeOOH phase is lower than in the nitrate case (4 mol% Mo), but in samples formed at low molybdate levels, the products were found to contain a mixture of sodium jarosite and β-FeOOH. There is, once again, a range of molybdate concentrations over which β-FeOOH is the only phase present in the product, in this case from 12 to 18 mol% Mo. Mixtures of sodium jarosite and β-FeOOH are obtained when the molybdate content is increased further, and when high levels of sodium molybdate (>25% mol% Mo) are added to the reaction mixture, the β-FeOOH structure is no longer observed and a sodium-jarosite phase becomes the dominant product of the reaction.

When ferric chloride solution is used as the source of iron, it is possible to approach the maximum level of molybdate incorpora-tion in the β-FeOOH phase, as seen in Figure 3 for molybdenum levels above 20 mol%. When the added molybdate is above 20 mol%, the relationship between added and incorporated molyb-denum is no longer linear, and the incorporated molybdenum tends to flatten out at a value of about 25 mol%. Several attempts to prepare samples with higher molybdate content have indicated that 25 mol% is the highest level of molybdate incorporation that can be achieved for the β-FeOOH phase.

effeCt of molybDate on morphology

β-FeOOH formed from ferric chloride solutions at pH = 2 was found to be composed of nanocrystalline rod-shaped particles of width ~10 nm and a broad distribution of particle lengths in the range of 20–300 nm. The addition of molybdate anions to the reaction induces a change in particle morphology that is displayed in TEM images and particle size distributions shown in Figure 4. It is clear that the presence of molybdate anions, even

in amounts as low as ~3 mol% leads to a decrease in the average particle length, and that with increasing molybdate content, the particles become shorter until they are essentially equiaxial at 25 mol% molybdenum. In this sample, the particles are roughly spherical with an average diameter of ~6 nm and in the TEM image presented here they resemble ferrihydrite nanocrystals. An interesting feature of our molybdate-incorporated β-FeOOH is that we have not observed the spiculed-ball morphology that is typical of schwertmannite, and which has been reported for both molybdate and tungstate-based β-FeOOH by other researchers (Linehan et al. 1997).

Morphology changes caused by adsorbed ions usually arise when the ions bind preferentially to a particular crystal face and inhibit crystal growth on that face. Rod-like β-FeOOH particles of the type shown in Figure 3b are known to be elongated along the direction parallel to the tunnel axis (Atkinson et al. 1977), so the effect of molybdate incorporation is inhibition of growth on (010) faces of the β-FeOOH crystals. Given that the pres-ence of molybdate promotes the formation of a tunnel structure (β-FeOOH), it is surprising that it also inhibits the development of the crystals along the tunnel direction. This may indicate that molybdate incorporation is not limited to tunnel sites alone, as we would expect that this type of incorporation might well favor the development of crystals along the tunnel direction.

DisCussion

In a previous paper, we have shown that co-precipitation of iron with molybdate anion produces an iron oxyhydroxide phase with structural incorporation of molybdenum (Richmond et al. 2004). The presence of molybdate enables the formation of the β-FeOOH phase even in the absence of chloride anions and the resulting molybdate-containing β-FeOOH phase appears to be considerably more stable than the analogous chloride form with respect to its phase transformation to hematite. At pH = 2 the presence of chloride anions facilitates rapid formation of the β-FeOOH structure within a concentrated suspension of fer-rihydrite, but hematite is the more stable phase in this system, and thus at 100 °C, we would expect β-FeOOH suspensions to transform to hematite after about 2–4 days (Sugimoto and Mura-matsu 1996). The XRD patterns shown in Figure 1, demonstrate that in the presence of molybdate, no hematite is detected in the suspensions even after 7 days at 98 °C, indicating that the molybdate-based β-FeOOH phase is considerably more stable than the chloride form. The patterns presented in Figure 1b also show that the initial formation of β-FeOOH is slower in the pres-ence of molybdate than in the presence of chloride.

β-FeOOH is known to crystallize via a dissolution-reprecip-itation mechanism in ferrihydrite suspensions where the pH lies within the range 1–3 (Atkinson et al. 1977; Cornell and Schwert-mann 2003). This is believed to be due to the stability of a soluble precursor comprising pairs of edge-sharing FeO6 octahedra, which is the main solution species present at this pH (Atkinson et al. 1977). It has been shown that the nucleation mechanism for β-FeOOH involves the formation of dimers composed of pairs of edge-sharing FeO6 octahedra, followed by formation of trimers with additional corner sharing (Bottero et al. 1994). The coalescence of these trimers into Fe24 polycations follows, and this final step appears to be templated by chloride anions

Figure 3. Combined data for samples prepared from ferric chloride (triangle), ferric sulfate (square), and ferric nitrate solutions (circle), showing the relationship between molybdenum added to the reaction mixture and molybdenum incorporated into the iron oxyhydroxide product (β-FeOOH).

CARROLL AND RICHMOND: INCORPORATION OF MOLyBDATE ANION INTO β-FeOOH 1645

to produce a polycation that has the tunnel structure typical of β-FeOOH. Although the molybdate ion must clearly template the β-FeOOH structure as does chloride, the slower formation of β-FeOOH in the presence of molybdate ion points to some significant differences in the interaction of the two anions with FexOy precursor polycations.

First, it would seem that the high degree of surface adsorption of MoO4

2– and HMoO4– anions to the initial ferrihydrite phase

must, to some extent slow the re-dissolution of ferrihydrite and thus inhibit the formation of the key embryonic precursor oligom-ers for β-FeOOH crystallization. Second, the greater stability of the molybdate-templated β-FeOOH phase provides evidence for the possibility that molybdate incorporation might involve the formation of Fe-O-Mo bonds between the molybdate anions and the FeO6 octahedra making up the sides of the β-FeOOH tunnels. A further consideration is that molybdenum may incorporate into the double straight-chains of FeO6 octahedra forming the β-FeOOH “backbone,” as indicated by the work of Linehan et al. (1997). If molybdate ions are incorporated into the β-FeOOH structure in this way, then the increased stability of the molybdate β-FeOOH phase seems logical, given that dissolution of this phase to facilitate transformation to hematite would then require the breakage of Fe-O-Mo bonds. It should be noted that simple

experiments, described in an earlier paper, showed that molybdate ion in Mo-akagaéite did not ex-change readily with other anions (Richmond et al. 2004)—in contrast to the properties reported for ha-lide-containing β-FeOOH (Cai et al. 2001). It is possible, however, that the differences in size and geometry of molybdate and chloride anions may mean that they are bound differently into the tun-nels, in which case they might behave differently to exchange and in their influence on the stability of β-FeOOH. In particular, one might expect molyb-date to form more and stronger H-bonds with the OH anions in β-FeOOH than chloride.

Although we have studied the formation of β-FeOOH in systems with added molybdate levels up to 40 mol%, the high-est levels of molybdate we observed in the prod-ucts was 25 mol%. In the preparations carried out

using ferric nitrate and ferric sulfate solutions, no β-FeOOH was observed at all once the molybdate concentration was in-creased above 25 mol%. However, when the ferric chloride was used, the samples obtained from systems with added molybdate >25 mol% still contained a molybdate level close to this value. These observations allow us to consider some possible structural models for the molybdate β-FeOOH phase. If the molybdate anion enters the β-FeOOH tunnel sites alone, then molybdate incorporation of 25 mol% is difficult to rationalize, given that full occupancy of tunnel sites by molybdate will correspond to an approximate stoichiometry of (FeO)8(OH)6(HMoO4)2, which has the Fe:Mo ratio 4:1. Alternatively, if the molybdenum in-corporates as octahedral MoO6 units by substitution at Fe sites within the β-FeOOH “backbone,” then an Fe:Mo ratio of 3:1 can be explained by a stoichiometry of (FeO)6(MoO)2(OH)8. In this model of the structure, the tunnel sites must be occupied by OH– anions, but if synthesized in the presence of chloride, then chloride ions could fill tunnel sites without any change to the overall Fe:Mo ratio. Previous Raman spectroscopic (Richmond et al. 2004), and XAFS experiments (Linehan et al. 1997), have indicated that molybdenum in β-FeOOH is present in both tetrahedral and octahedral environments, which would suggest that at least some of the molybdate must occupy tunnel sites,

Figure 4. Effect of molybdate addition on the morphology of β-FeOOH particles. (a) Size distributions, measured from TEM images for samples prepared in ferric chloride media with Mo:Fe molar ratios of 2.8, 6.3, and 11.4%. (b) TEM images of β-FeOOH prepared from ferric chloride solution in the absence of molybdate (top) and with Fe:Mo ratio of 3:1 (bottom).

CARROLL AND RICHMOND: INCORPORATION OF MOLyBDATE ANION INTO β-FeOOH1646

and this would mean that the phase must have an intermediate composition between the two formulas we have proposed. In the present study, we have not been able to distinguish among molybdate incorporated at tunnel sites, molybdate that might be incorporated in the lattice, and chemisorbed surface molybdate. Thus, a precise determination of the compound’s stoichiometry is not possible from the data available here, and it is likely that a full crystallographic structure refinement will be required to determine the levels of iron substitution for molybdate/β-FeOOH phases formed under different conditions, and the structural mode of molybdate incorporation.

The results we have described here raise some interesting questions in relation to the physical requirements for formation of the β-FeOOH structure. It is clear that, regardless of the anions present in the reaction mixture, there is a narrow acidic pH range in which the β-FeOOH phase can form, but within this range there is also the requirement that the anion be able to, in some way, “template” the β-FeOOH structure. Thus, β-FeOOH cannot be formed in the presence of nitrate ions or sulfate alone, while the addition of as little as 4 mol% molybdate enables the formation of a stable β-FeOOH structure from ferric sulfate solution, and in the nitrate system, at least 9 mol% molybdate is required.

Although the presence of molybdate in tunnel sites is still a possibility, the results presented here seem to indicate that this form of molybdate inclusion cannot fully account for the extent of molybdenum incorporation into the structure, nor can it account for the properties of the resulting molybdate β-FeOOH phase. Ongoing studies aimed at determining the relative amounts of tetrahedral and octahedral Mo present in these materials are expected to provide more detail on the nature of molybdate incorporation.

aCknoWleDgmentsThis research has been supported by the Australian Government’s Cooperative

Research Centre (CRC) Program, through the Parker CRC for Integrated Hydro-metallurgy Solutions.

referenCes CiteDAtkinson, R.J., Posner, A.M., and Quirk, J.P. (1977) Crystal nucleation and

growth in hydrolysing iron(III) chloride solutions. Clays and Clay Minerals, 25, 49–56.

Bigham, J.M., Carlson, L., and Murad, E. (1994) Schwertmannite, a new iron oxyhydroxy sulfate from Pyhäsalmi, Finland, and other localities. Mineralogi-

cal Magazine, 58, 641–648.Bigham, J.M., Schwertmann, U., Traina, S.J., Winland, R.L., and Wolf, M. (1996)

Schwertmannite and the chemical modeling of iron in acid sulfate waters. Geochimica et Cosmochimica Acta, 60, 2111–2121.

Bottero, J.-y., Manceau, A., Villieras, F., and Tchoubar, D. (1994) Structure and mechanisms of formation of FeOOH(Cl) Polymers. Langmuir, 10, 316–319.

Buchwald, V.F. and Clarke, R.S.J. (1989) Corrosion of Fe-Ni alloys by Cl-containing akaganéite (β-FeOOH): The Antarctic meteorite case. American Mineralogist, 74, 656–667.

Cai, J., Liu, J., Gao, Z., Navrotsky, A., and Suib, S.L. (2001) Synthesis and anion exchange of tunnel structure akaganeite. Chemistry of Materials, 13, 4595–4602.

Cornell, R.M. and Schwertmann, U. (2003) The Iron Oxides. VCH, Weinheim.Holm, N., Dowler, M.J., Wadsten, T., and Arrhenius, G. (1983) β-FeOOH-Cln

(akaganéite) and Fe1–xO(wüstite) in hot brine from the Atlantis II Deep (Red Sea) and the uptake of amino acids by synthetic β-FeOOH-Cln. Geochimica et Cosmochimica Acta, 47, 1465–1470.

Linehan, J.C., Darab, J.G., Matson, D.W., Chen, X., and Amonette, J.E. (1997) Synthesis and characterisation of akaganéite-like ferric oxyhydroxides. Ma-terials Research Society Symposium Proceedings, 432, p. 157–161. Materials Research Society, Warrendale, Pennsylvania.

Loan, M., Hart, R.D., Cowley, J.M., and Parkinson, G.M. (2004) Evidence on the structure of synthetic schwertmannite. American Mineralogist, 89, 1735–1742.

MacKay, A.L. (1962) β-Ferric oxyhydroxide-akaganéite. Mineralogical Magazine, 33, 270–280.

Matson, D.W., Linehan, J.C., Darab, J.G., and Buehler, M.F. (1994) Nanophase iron-based liquefaction catalysts: Synthesis, characterization, and model compound reactivity. Energy and Fuels, 8, 10–18.

Paterson, R. and Rahman, H. (1984) The ion exchange properties of crystalline inorganic oxide-hydroxides. Part II. Exclusion of perchlorate from β-FeOOH by an ion sieve mechanism. Journal of Colloid and Interface Science, 97, 423–427.

Post, J.E. and Buchwald, V.F. (1991) Crystal structure refinement of akaganeite. American Mineralogist, 76, 272–277.

Post, J.E., Heaney, P.J., Von Dreele, R.B., and Hanson, J.C. (2003) Neutron and temperature-resolved synchrotron X-ray powder diffraction study of akagané-ite. American Mineralogist, 88, 782–288.

Richmond, W.R., Hockridge, J.G., Loan, M., and Parkinson, G.M. (2004) A new iron oxyhydroxide phase: The molybdate-substituted analogue of akaganéite. Chemistry of Materials, 16, 3203–3205.

Schwertmann, U., Friedl, J., and Pfab, G. (1996) A New Iron(III) Oxyhydroxyni-trate. Journal of Solid State Chemistry, 126, 336.

Sugimoto, T. and Muramatsu, A. (1996) Formation Mechanism of monodispersed α-Fe2O3 particles in dilute FeCl3 solutions. Journal of Colloid and Interface Science, 184, 626–638.

Walenta, K. (1982) Akaganéit und Ferrihydrit aus der Grube Clara (Schwartzwald). Aufschluss, 33, 367–373.

Manuscript received noveMber 12, 2007Manuscript accepted May 14, 2008Manuscript handled by sergey Krivovichev