Characterization of Ferrihydrite-SoilOrganic Matter Coprecipitates byX-ray Diffraction and Mo1ssbauerSpectroscopyK A R I N E U S T E R H U E S , * , †

F R I E D R I C H E . W A G N E R , ‡

W E R N E R H A U S L E R , §

M A R I A N N E H A N Z L I K , | H E I K E K N I C K E R , §

K A I U . T O T S C H E , †

I N G R I D K O G E L - K N A B N E R , § A N DU D O S C H W E R T M A N N §

Institut fur Geowissenschaften, Friedrich-Schiller-Universitat,07749 Jena, Germany, Physik Department, TechnischeUniversitat Munchen, 85747 Garching, Germany, Lehrstuhlfur Bodenkunde, Technische Universitat Munchen, 85350Freising, Germany, and Department Chemie, TechnischeUniversitat Munchen, 85748 Garching, Germany

Received March 28, 2008. Revised manuscript receivedAugust 15, 2008. Accepted August 18, 2008.

In soils and sediments ferrihydrite often precipitates fromsolutions containing dissolved organic matter, which affectsits crystallinity. To simulate this process we prepared a seriesof 2-line ferrihydrite-organic matter coprecipitates usingwater extractable organic matter (OM) from a forest topsoil.The products were characterized by X-ray diffraction, Mossbauerspectroscopy, N2-gas adsorption and transmission electronmicroscopy. With increasing C/Fe ratios of the initial solutionthe d-spacings of the two major XRD peaks increased, while peakshoulders at 0.22 and 0.16 nm weakened. The asymmetry ofthe 0.26 nm peak decreased and disappeared at a C/Fe ratio of0.78. The quadrupole splitting of the Mossbauer spectra at300 K increased from 0.78 to 0.90 mm s-1, the mean magnetichyperfine field at 4.2 K dropped from 49.5 to 46.0 T, and thesuperparamagnetic collapse of the magnetic hyperfine splittingwas shifted toward lower temperatures. These data reflect astrong interference of OM with crystal growth leading to smallerferrihydrite crystals, increased lattice spacings, and moredistorted Fe(O,OH)6 octahedra. Even small amounts of OMsignificantlychangeparticlesizeandstructuralorderof ferrihydrite.Crystallinity and reactivity of natural ferrihydrites will thereforeoften differ from their synthetic counterparts, formed in theabsence of OM.

IntroductionFerrihydrite (Fh) is a poorly crystalline Fe oxyhydroxideubiquitously occurring in soils, sediments, mine wastes, andother aqueous environments (1-3). The combination of alarge specific surface area (up to 700 m2 g-1) and a highamount of pH-dependent functional hydroxyl groups at the

surface makes ferrihydrite an efficient sorbent for ions andorganic molecules (4). In consequence, ferrihydriteseven ifpresent at a low concentrationsmay determine the mobilityand availability of many nutrients and pollutants. In soils,ferrihydrite adsorbs high amounts of soil organic matter. Socalled “sorptive protection” by soil minerals like ferrihydriteis assumed to stabilize OM from biological degradation andto represent a major mechanism of carbon sequestration (5, 6).Conversely, association of ferrihydrite with dissolved OMmay retard or inhibit the transformation of ferrihydrite intomore crystalline Fe oxides.

The structure of ferrihydrite is still under discussion (1, 7, 8).Identification and characterization is hampered by its smallparticle size and poor crystallinity. As in all Fe oxides andhydroxides, the characteristic structural unit is the Fe (O,OH)6 octahedron, which forms (short) edge-linked di- ortrioctahedral chains. These are cross-linked at octahedralcorners to neighboring chains. The published structuralmodels differ in the proposed stacking order of the anioniclayers, the distribution of Fe on the octahedral sites, thepresence of Fe in tetrahedral coordination, or the existenceof face sharing octahedra. Several models describe ferrihydriteas consisting of microdomains of different crystal structuresand varying crystal order (9, 7, 10), whereas others emphasizethat its surface structure is different from that of thebulk (11, 12). A specific surface structure is a typical propertyof nanoparticles, which may become manifested in con-tracted or expanded surface layers, a different surfacecomposition, or an enhanced disorder at the particle surface(13). This may strongly influence structural analyses becauseof the large surface area/volume ratio of small particles. Forexample, Zhao et al. (11) assigned as much as 35% of the Featoms in ferrihydrite to the surface. Likewise, chemisorptionof Si and borate by ferrihydrite was found to significantlyalter its XRD pattern and Mossbauer spectrum (14, 15).

Especially in soils, ferrihydrite often forms in the presenceof dissolved OM. This involves processes such as adsorption,coprecipitation, flocculation/coagulation, and poisoning ofcrystal growth (16, 17). Schwertmann et al. (18) found thatcoprecipitation of ferrihydrite with water extractable soilorganic matter lowered its crystallinity and decreased themagnetic ordering temperature as well as the magnetichyperfine field. In a similar approach, Mikutta et al. (19)precipitated ferrihydrite in the presence of synthetic acidpolysaccharides to investigate the influence of microbialexudates on the mineral properties. They observed a strongeraggregation of ferrihydrite particles due to the polymers, aslightly decreasing magnetic hyperfine field, but no significantchange of the crystal structure (XRD) or the local coordinationof Fe (EXAFS).

This study aims at a better understanding of the formationof Fe(III) oxides under natural conditions, specifically innaturally occurring solutions with dissolved humic sub-stances. We therefore synthesized a series of ferrihydritesusing solutions with increasing amounts of OM extractedfrom soil. Assuming that the OM is normally soluble, butcarried down by the precipitating ferrihydrite due to sorptionor occlusion, we address the products obtained at variousC/Fe ratios as “coprecipitates”. These coprecipitates werecharacterized by X-ray diffraction, Mossbauer spectroscopy,and transmission electron microscopy. Our main objectiveswere to quantify how coprecipitation affects the loadingcapacity of ferrihydrite for OM and how the occupation ofgrowth sites changes the ferrihydrite structure. Such inves-tigations are a first step to relate the properties of natural

* Corresponding author phone:+49-3641-948642; fax:+49-3641-948622; e-mail: karin.eusterhues@uni-jena.de.

† Friedrich-Schiller-Universitat.‡ Physik Department, Technische Universitat Munchen.§ Lehrstuhl fur Bodenkunde, Technische Universitat Munchen.| Department Chemie, Technische Universitat Munchen.

Environ. Sci. Technol. 2008, 42, 7891–7897

10.1021/es800881w CCC: $40.75 2008 American Chemical Society VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 7891

Published on Web 10/08/2008

ferrihydrites to the conditions of their formation and to assessdifferences between natural and synthetic ferrihydrites.

Experimental SectionRecovery of Water Extractable Soil Organic Matter. Water-extractable soil organic matter (OM) was extracted from thelitter layer of a podzol close to Freising, Germany. Sampleswere air-dried and passed through a 2 mm sieve to removecoarse plant remains. Aliquots of 150 g of soil and 700 mLof deionized H2O were shaken end-over-end for 16 h at roomtemperature and then centrifuged for 1 h at 4000 rpm. Thesupernatant was pressure-filtered through polyvinylidenefluoride (Durapore; 0.45 µm pore width) membranes, con-centrated in low temperature rotary evaporators and freeze-dried. This material is supposed to represent the polysac-charide-rich composition of a typical soil solution. A solidstate 13C NMR spectrum (18) suggests a total aromatic Ccontent of less that 6%. The signal for carbonyl-amide Caccounts for 7% of the total C and is centered at 175 ppm(typical for esters). A small peak at 17 ppm was assigned toterminal methyl C groups. With 65-70% of the total C theO-alkyl C groups represent the most important functionalgroup of the extracted OM. The freeze-dried OM contains371 mg kg-1 C and 14 mg kg-1 N.

Synthesis of Ferrihydrite. Two-line ferrihydrite wasproduced by titrating a 0.01 M Fe(NO3)3 solution with 0.1 MKOH to a pH of 7 (20). Organic matter-ferrihydrite copre-cipitates were prepared by dissolving the Fe(NO3)3 in asolution of extracted OM and deionized H2O and raising thepH afterward. Products with different OM loadings wereproduced by varying the C/Fe ratio of the initial solutionbetween 0.1 and 2.3 (Table 1). The precipitates were separatedby centrifugation, washed twice with deionized H2O, andfreeze-dried.

Analytical Methods. Total C content was measured witha CN analyzer (Vario EL, Elementar GmbH) of solids anddried supernatants. Specific surface area (SSA) was measuredby N2 gas adsorption with an Autosorb1 instrument (Quan-tachrome) and calculated according to the BET- equationfrom 11 data points in the relative pressure range of 0.05-0.3.Prior to the measurements, samples were outgassed for atleast 16 h at 343 K in a flow of He to remove adsorbed waterfrom the sample surfaces. For transmission electron mi-croscopy (TEM), ∼2 mg of Fh were ultrasonically dispersedin ∼20 mL of deionized H2O. A drop of this suspension wasairdried on a TEM grid and examined with a JEOL JEM 2011,working at an accelerating voltage of 120 kV. Powder X-raydiffraction (XRD) patterns were obtained with a Phillipsvertical goniometer (PW 1050) using Co KR radiation at 40kV and 30 mA, step-scanning from 5 to 90 °2Θ withincrements of 0.02 °2Θ and a counting time of 5 s per step.

Data were profile-fitted with XFIT (21) using a linearbackground and pseudo-Voigt functions. The asymmetricferrihydrite peak at ∼0.25 nm (9) was fitted by two pseudo-Voigt peaks. The noise of each diffractogram was quantifiedby the difference of the observed diffractogram and a filteredcurve. For filtering we combined a running median with aSavitzky-Golay filter. If the root-mean-square error (rmse)of the fitted minus the observed curve approached the levelof noise, no further peak was included. The inclusion of allpeaks clearly increased the goodness of fit, as compared toa linear combination fit of any combination of fewer peaks:inclusion of the last accepted peak decreased the rmse by atleast a factor of 10 in all cases. Mossbauer spectra wererecorded in transmission geometry using a sinusoidal velocitywaveform and a source of 57Co in Rh. For measurementsbelow ambient temperature, a liquid He bath cryostat wasused. Temperatures above 4.2 K were reached during theslow warming of the cryostat after the liquid He had run out.These spectra average over regions of typically 2-3 K. Themagnetically split spectra were fitted with broadened patternscorresponding to static Gaussian distributions of hyperfinefields. Spectra exhibiting electric quadrupole interactionswere fitted with patterns corresponding to Gaussian distri-butions of quadrupole splittings. In the transition region,where magnetically split patterns and pure quadrupolepatterns coexist, superpositions of both distributions wereused. Further details of the fitting procedure have beendescribed in ref19.

Results and DiscussionParticle Size, Surface Area, and C Concentration. Trans-mission electron micrographs show that the OM-free refer-ence ferrihydrite forms aggregates composed of particles∼3-5 nm in diameter (Figure 1a). This is consistent with thespecific surface area (SSA) measurement: assuming a densityof ∼4 g cm-3 (3), a spherical geometry and no surfaceroughness, a SSA of 386 m2 g-1 (Table 1) corresponds to aparticle diameter of ∼4 nm. In the presence of OM,ferrihydrites form considerably smaller particles (Figure 1band c). Their SSA, however, does not reflect their small particlesize (250 m2 g-1 for Fh at a C/Fe ) 0.09 and 5-23 m2 g-1 forall other samples; Table 1). These small SSAs can be explainedby (i) a reduced accessibility for N2 due to the formation ofdenser aggregates by the associated OM or (ii) by a maskingof the mineral surface by OM. According to Chiou et al. (22)natural OM exhibits only very low N2-BET specific surfaceareas of ca. 1 m2 g-1. An OM coverage of approximately one-third of the oxide surface for the Fh at C/Fe ) 0.09, and analmost complete coverage for the other coprecipitates cantherefore account for their low N2-BET specific surface areas.

TABLE 1. Synthesis Conditions: C/Fe ratio, C Concentration, Fe Concentration, and pH of the Initial Solution; And Properties ofthe Coprecipitates: Calculated C/Fe ratio, C Concentration, Specific Surface Area (SSA), Average Quadrupole Splitting at 300 K,Average Hyperfine Field at 4.2 K, and the Superparamagnetic Blocking Temperaturea

C/Fe initialsolution

C initialsolution(mg l-1)

Fe initialsolution(mg l-1)

pH initialsolution

C contentcoprecipitate

(mg g-1)C/Fe coprecipitate

(calculated)SSA

(m2 g-1)

Quadrupolesplit at 300 K

(mm s1-)

Hyperfinefield at

4.2 K(T)

Blockingtemperature

(K)

0.00 0 436 2.3 5 0.01 386 0.78 49.5 84.00.09 188 2038 1.9 39 0.08 250 nd 47.9 56.00.28 120 430 2.3 79 0.17 13 0.83 48.3 53.00.57 317 554 2.2 122 0.30 16 0.86 47.7 47.00.78 353 450 2.1 135 0.35 23 0.87 46.9 nd1.17 307 261 2.3 159 0.46 17 0.87 46.5 nd1.48 296 200 2.3 142 0.40 5 0.89 46.5 nd2.28 568 249 2.2 170 0.54 b 0.90 46.0 36.0

a nd is not detected. b Smaller than detection limit with 0.1 g sample material, i.e. < 10 m2 g-1.

7892 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

Carbon concentrations of coprecipitates increase withincreasing C/Fe ratio of the initial solution from 39 to 170mg g-1 C in the products (Table 1, Figure 2a). The highestachieved loading of 170 mg C per g ferrihydrite correspondsto ca. 460 mg OM per g ferrihydrite. This loading is roughlytwice as large as observed in sorption experiments by Tipping(23), but smaller than the maximum loading of 318 mg C perg ferrihydrite reached in sorption experiments with waterextractable soil OM by Kaiser et al. (24). One has to bear inmind, however, that different OM sorbtives were used inthese experiments and that loadings expressed on a massbasis strongly depend on the SSA of the sorbent.

If we assume that ferrihydrites in the coprecipitates havethe same surface area as the OM-free control sample (386m2 g-1), the highest achieved loading corresponds to 0.77mg C per m2 ferrihydrite (this is an overestimation becauseof the fact that TEM and XRD results point to smaller particlesand therefore higher mineral surface area for the copre-cipitates). Compared to maximum loadings from sorptionexperiments with water extractable soil OM on ferrihydrite(1.1 mg C m-2 ferrihydrite (24);) and estimated sorptionmaxima on natural pedogenic Fe oxides from acid forestsoils (1.2 mg C m-2 Fe oxide (25);) the maximum loadingobtained in our coprecipitation experiment is rather low.We conclude that coprecipitation produces OM loadings ofthe same magnitude as pure sorption processes and do notencapsulate large amounts of OM in aggregates.

A mass balance calculation comparing the amount of Cand Fe in the initial solution and in the products shows thatincreasing proportions of both elements are not precipitated,but remain in solution as the C/Fe ratio of the initial solutionincreases (Figure 2b and c). Amounts of nonprecipitated Feand C are correlated (Figure 2c) and correspond to an atomicC/Fe ratio of 13.3 in the supernatant. The linear correlationsuggests that the formation of a specific soluble Fe-OM-complex is responsible for the mobilization of the nonprec-ipitated Fe. However, this result may also be interpreted asan experimental artifact, i.e., an increasing formation of verysmall Fe oxides with increasing C/Fe, associated with a higherloss of solid material during centrifugation.

Structural Characteristics. The XRD of the OM freereference sample (C/Fe ) 0) shows the usual peaks of 2-lineferrihydrite at ∼0.26 nm and ∼0.15 nm (Figure 3). Their peakshoulders can be explained by smaller peaks at 0.22, 0.20,0.17, and 0.16 nm, which correspond to d-spacings forferrihydrites with higher than two-line crystallinity. Theasymmetry of the 0.26 nm-peak (9) was quantified by anadditional peak at 0.29 nm. With an increasing C/Fe ratio ofthe initial solution the XRD pattern changes systematically:The small peaks at 0.22, 0.20, 0.17, and 0.16 nm, as well asthe one at 0.29 nm weaken until they finally disappear, whilethe two main peaks broaden and shift toward higherd-spacings (Figure 3 and Figure 4a-c). The decreasingnumber of peaks and the observed peak broadening indicatea decreasing size of coherent scattering domains or anincrease of stacking disorder in the anionic layers. Based onEXAFS, Waychunas et al. (26) suggested that the precipitationprocess of ferrihydrite starts with the formation of smallchains of Fe octahedra, continues by linking these chains viatheir edges to plate-like dioctahedral and trioctahedral chains,and only then proceeds by cross-linking the chains at theiroctahedral corners. This scenario is in line with our observa-tions: At high initial C/Fe ratios, frequent reactions betweenFe (O, OH)6 octahedra and organic molecules may preventcross-linking and result in particles with only two XRD peaks,ascribed to (110) and (300). This indicates that only per-pendicular to these two hk0 planes enough lattice planes arestacked to produce an XRD peak, suggesting a plate-like shapeof the coherent scattering domains. Less interference withorganic molecules allows for the development of the ad-ditional hkl planes (112) and (115), i.e, the low intensity peaksat 0.22 and 0.16 nm, which might be connected to the cross-linking process.

The observed peak shift points to increased spacings forthe 110 and 300 planes. Such peak shifts are also found forferrihydrites doped with borate (15) and Si (14). In bothstudies, the shifts were considered to reflect changes in thelattice spacings, and the non-Fe metals were assumed tochemisorb on the surface of the ferrihydrite nanoparticles.Increasing d-spacings are also in accordance with Michel et

FIGURE 1. TEM images show aggregates of pure ferrihydrite (a), and of ferrihydrite-OM coprecipitates, which formed at different C/Fe ratios of the initial solution (b, c). Ratios in parentheses refer to the C/Fec of the coprecipitates.

FIGURE 2. Carbon concentration of coprecipitates vs C/Fe of the initial solution (a), percentage of C, which is not precipitated, butremaining in the supernatant vs C/Fe of the initial solution (b), and correlation between the nonprecipitated amounts of C and Fe (c).

VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 7893

al. (8), who observed systematically changing unit celldimensions with decreasing crystal size. The asymmetry ofthe 0.26 peak is known for ferrihydrites regardless their degreeof crystallinity. It was interpreted as being caused by stackingdisorder in the close packed anionic planes by Janney et al.(7) or to represent ultra dispersed hematite in ferrihydrite byDrits et al. (9). However, the most intense hematite peakappears at ∼0.27 nm, whereas we had to include a peak at0.29 nm to quantify the peak asymmetry.

The diffractogram of the OM was fitted by a broad peakat ∼0.49 nm, which also appears in coprecipitates with higherC contents (i.e., at C/Fe of 0.78, 1.18, 1.48, and 2.28). Threesharp peaks at 0.24, 0.20, and 0.14 nm in the diffractogramsof OM and ferrihydrite at C/Fe ) 2.28 can be assigned to Aland are caused by the sample holder.

Magnetic Characteristics by Mossbauer Spectroscopy.All spectra recorded at room temperature show broadenedelectric quadrupole doublets with isomer shifts (IS) of0.35-0.36 mm s-1 againstR-Fe. This is consistent with valuesreported in other studies (0.31-0.38 mm s-1; compiled byref12) and is in the range for trivalent Fe of 0.24-0.54 mms-1 (27). The isomer shifts fall into the overlap range oftetrahedrally (0.24-0.44 mm s-1) and octahedrally (0.34-0.54mm s-1) coordinated Fe and do not help to decide whether4-fold coordinated Fe is present (27, 28). The broadening ofthe doublets can be attributed to a distribution of electricfield gradients. Good fits were obtained with an asymmetricalGaussian distribution of quadrupole splittings (QS), i.e., witha distribution having a larger variance on the high QS sidethan on the low QS side. The mean QS increases withincreasing C/Fe ratios of the initial solution from 0.77 mm

FIGURE 3. X-ray diffractograms of pure ferrihydrite, somecoprecipitates with different initial C/Fe ratios (ratios inparentheses refer to the C/Fec of the coprecipitates) and theextracted organic matter. The heavy black line denotes themeasured data points. Also shown are fit envelope,background, fitted peaks, and the residuum.

FIGURE 4. XRD data in relation to the C/Fe ratio of the initialsolution: d-spacings (a, b) and full width at half-maximum(fwhm) of the 110 and 300 X-ray peaks (c).

7894 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

s-1 for the organic free ferrihydrite to 0.90 mm s-1 for theferrihydrite with a C/Fe ) 2.28 (Figure 5a). Increasing QSvalues indicate an increasing distortion of the octahedralgeometry around the Fe atom (29).

All 4.2 K Mossbauer spectra show a broadened sextet,indicating complete magnetic ordering at this temperature.None of the spectra contain a quadrupole doublet componentrepresenting iron that does not exhibit a magnetic splitting,as has been reported by Schwertmann et al. (18). The meanmagnetic hyperfine field (Bhf) decreases with increasing C/Feof the initial solution from 49.5 T of the pure ferrihydrite to46.0 T for the Fh at C/Fe ) 2.28 (Figure 5b). Such a decreasecan be explained by decreasing particle size and crystallinity(30) or by decreasing interparticle interactions (31). Botheffects may apply here: decreasing particle size with increas-ing initial C/Fe ratios is evident from the TEM images (Figure1), decreasing crystallinity is shown by XRD, and decreasinginterparticle interactions are plausible because of the in-creasing amounts of diamagnetic organic materials betweenindividual ferrihydrite particles. Interestingly, by usingpolysaccharide-rich natural OM we found much largerchanges in QS and Bhf than that described in ref19, whereferrihydrite was synthesized in the presence of various purepolysaccharides.

For some samples, the temperature dependence of theMossbauer spectra was recorded between 4.2 and 116 K(Figure 6). In this temperature range, the magnetic hyperfinesplitting collapses, which is shown exemplary for the co-precipitate obtained at a C/Fe)0.57 (Figure 7). The vanishingof the magnetic splitting at temperatures well below the Neeltemperature TN (TN ∼ 500 K for 6-line Fh (32)) is explainedby superparamagnetic relaxation. When the relaxationbecomes sufficiently fast, the magnetic effects in the Moss-bauer spectra disappear. The superparamagnetic relaxationrate ω increases with increasing temperature and decreasingparticle size according to

ω)ωo exp[-KV ⁄ kT] (1)

where ωo is a constant depending on the material andperhaps its particle shape, as is the anisotropy constant K,V is the particle volume, k is Boltzmann’s constant, and Tis the temperature. In a ferrihydrite powder, particle sizeand particle shape would have distributions, so that adistribution of relaxation rates has to be expected at anygiven temperature. The narrower the distribution of proper-ties of the particles, the more abrupt is the transition fromthe magnetically split sextet to the quadrupole doublet (29).Therefore, the width of the transition is an indicator of thehomogeneity of particle sizes, shapes, and perhaps otherproperties. Schwertmann et al. (18) reported that for 6-lineferrihydrite, the same spectral shapes (and hence relaxationrates) occur at a temperature twice as high as with 2-lineferrihydrite. Neglecting particle shape, this means that themean particle volumes are about two times larger in 6-linethan in 2-line ferrihydrite. The particle diameter would differby much less, namely by roughly the cubic root of 2, i.e., bya factor of 1.26. Consequently, rather small differences in themean particle size result in rather drastic differences in thetemperature dependence of the Mossbauer spectra.

The magnetically split fraction (Amag) of the Mossbauerspectra and the mean hyperfine field (Bhf) in this fraction areshown in Figure 6 for all samples, for which the temperaturedependence was measured. Whereas according to Schw-ertmann et al. (18) 6-line ferrihydrite still has a magneticfraction of 100% at 100 K, the organic-free 2-line ferrihydritefrom our coprecipitation series starts to lose its magneticorder above 72 K (Figure 6b), where the Bhf has decreasedto ∼20 T. For ferrihydrite precipitated in the presence ofOM, both the proportion of the sextet as well as the Bhf

decrease even more rapidly. A surprisingly large changeoccurs between the organic free ferrihydrite and the onewith the lowest OM content (Fh at C/Fe ) 0.09), indicatingthat even small amounts of OM may have a significant effect

FIGURE 5. Dependence of the Mossbauer results for the meanelectric quadrupole splitting (QS) at ambient temperature (a)and of the mean magnetic hyperfine field (Bhf) at 4.2 K (b) onthe C/Fe ratio of the initial solution.

FIGURE 6. Mean hyperfine field (Bhf) and magnetic fraction(Amag) as obtained from Mossbauer spectra taken at differenttemperatures.

VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 7895

on the crystal size and/or structure. If the observed decreasein the superparamagnetic blocking temperature (Table 1;defined as the temperature at which the magnetic fractionhas dropped to 0.5) is only attributed to particle volume (eq1) the particle size for the coprecipitate obtained at C/Fe )2.28 is smaller by a factor of 0.75 than the pure 2-lineferrihydrite. The superparamagnetic behavior, and hence theblocking temperature may, however, also be influenced byinterparticle interactions, which will be reduced when OMseparates the ferrihydrite particles.

Mossbauer and XRD results show that even smallconcentrations of water soluble soil OM have a clear impacton structure, size, and magnetic properties of the pre-cipitated ferrihydrite. With increasing C/Fe of the initialsolution, that is with an increasing probability of reactionsbetween OM and ferrihydrite or the early polymers ofFe(O,OH)6 octahedra, structural features of ferrihydritechange systematically: organically poisoned ferrihydriteshows larger lattice spacings, fewer crystal planes, smallerparticle diameters, and a magnetic hyperfine splitting thatdisappears at lower temperatures. Natural ferrihydritesmust, therefore, be expected to differ in many propertiesfrom their synthetic analogues.

AcknowledgmentsMany thanks to Petra Muller and Maria Greiner for assistancein the laboratory, and to Roman Skala for helpful XRD-discussions. The thorough reviews by three anonymousreferees and the editorial work by D. Dzombak are gratefullyacknowledged. This study was supported by the Friedrich-Schiller-Universitat Jena and the Technische UniversitatMunchen.

Supporting Information AvailableExtension of Figure 7 shows five Mossbauer spectra taken atdifferent temperatures of one coprecipitate (C/Fe ) 0.57 orC/Fec) 0.30). This material is available free of charge via theInternet at http://pubs.acs.org.

Literature Cited(1) Jambor, J. L.; Dutrizac, J. E. Occurrence and constitution of

natural and synthetic ferrihydrite, a widespread iron oxyhy-droxide. Chem. Rev. 1998, 98, 2549–2585.

(2) Bigham, J. M.; Fitzpatrick, R. W.; Schulze D. G, Iron Oxides. In SoilMineralogy with Environmental Applications; Dixon, J. B, Schulze,D. G, Eds.; SSSA Book Ser. No. 7: Madison, WI, 2002.

(3) Cornell, R. M.; Schwertmann, U. The Iron Oxides: Structure,Properties, Reactions, Occurrences and Uses, 2nd ed.; Wiley-VCH Verlagsgesellschaft: Weinheim, 2003.

(4) Dzombak, D. A.; Morel, F. M. M. Surface Complexation Modeling.Hydrous Ferric Oxide; Wiley Interscience: New York, 1990.

(5) Jones, D. L.; Edwards, A. C. Influence of sorption on the biologicalutilization of two simple carbon substrates. Soil Biol. Biochem.1998, 30, 1895–1902.

(6) Kalbitz, K.; Schwesig, D.; Rethemeyer, J.; Matzner, E. Stabilizationof dissolved organic matter by sorption to the mineral soil. SoilBiol. Biochem. 2005, 37, 1319–1331.

(7) Janney, D. E.; Cowley, J. M.; Buseck, P. R. Structure of synthetic2-line ferrihydrite by electron nanodiffraction. Am. Mineral.2000, 85, 1180–1187.

(8) 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.

(9) Drits, V. A.; Sakharov, B. A.; Salyn, A. L.; Manceau, A. Structuralmodel for ferrihydrite. Clay Miner. 1993, 28, 185–207.

(10) Janney, D. E.; Cowley, J. M.; Buseck, P. R. Transmission electronmicroscopy of synthetic 2-and 6-line ferrihydrite. Clays ClayMiner. 2000, 48, 111–119.

(11) Zhao, J. M.; Huggins, F. E.; Feng, Z.; Huffman, G. P. Ferrihydrite- Surface-structure and its effects on phase-transformation. ClaysClay Miner. 1994, 42, 737–746.

(12) Rea, B. A.; Davis, J. A.; Waychunas, G. A. Studies of the reactivityof the ferrihydrite surface by iron isotopic exchange andMossbauer-spectroscopy. Clays Clay Miner. 1994, 42, 23–34.

(13) Waychunas, G. A.; Kim, C. S.; Banfield, J. F. Nanoparticulateiron oxide minerals in soils and sediments: unique propertiesand contaminant scavenging mechanisms. J. Nanopart. Res.2005, 7, 409–433.

(14) Seehra, M. S.; Roy, P.; Raman, A.; Manivannan, A. Structuralinvestigations of synthetic ferrihydrite nanoparticles doped withSi. Solid State Commun. 2004, 130, 597–601.

(15) Stevens, J. G.; Khasanov, A. M.; Mabe, D. R. Mossbauer andX-ray diffraction investigations of a series of B-doped ferrihy-drites. Hyperfine Interact. 2005, 161, 83–92.

(16) Schwertmann, U. Inhibitory effect of soil organic matter oncrystallization of amorphous ferric hydroxide. Nature 1966, 212,645–646.

(17) Cornell, R. M.; Schwertmann, U. Influence of organic-anionson the crystallization of ferrihydrite. Clays Clay Miner. 1979, 27,402–410.

(18) Schwertmann, U.; Wagner, F.; Knicker, H. Ferrihydrite-humicassociations: Magnetic hyperfine interactions. Soil Sci. Soc. Am. J.2005, 69, 1009–1015.

(19) Mikutta, C.; Mikutta, R.; Bonneville, S.; Wagner, F.; Voegelin, A.;Christl, I.; Kretzschmar, R. Synthetic coprecipitates of ex-opolysaccharides and ferrihydrite. Part I: Characterization.Geochim. Cosmochim. Acta 2008, 72, 1111–1127.

(20) Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory.Preparation and Characterization, 2nd ed.; Wiley-VCH VerlagGmbH: Weinheim, 2000.

(21) Cheary, R. W.; Coelho, A. A, Programs XFIT and FOURYA,deposited in CCP14 Powder Diffraction Library, Engineeringand Physical Sciences Research Council, Daresbury Laboratory,Warrington: England , 1996; http://www.ccp14.ac.uk/tutorial/xfit-95/xfit.htm.

(22) Chiou, C. T.; Lee, J. F.; Boyd, S. A. The surface-area of soil organic-matter. Environ. Sci. Technol. 1990, 24, 1164–1166.

(23) Tipping, E. The adsorption of aquatic humic substances by iron-oxides. Geochim. Cosmochim. Acta 1981, 45, 191–199.

(24) Kaiser, K.; Mikutta, R.; Guggenberger, G. Increased stability oforganic matter sorbed to ferrihydrite and goethite on aging.Soil Sci. Soc. Am. J. 2007, 71, 711–719.

FIGURE 7. Mossbauer spectra of coprecipitates obtained fromthe solution with an initial C/Fe ratio of 0.57 (C/Fec ) 0.30)taken at different temperatures.

7896 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

(25) Eusterhues, K; Rumpel, C; Kogel-Knabner, I. Organo-mineralassociations in sandy acid forest soils: importance of specificsurface area, iron oxides and micropores. Eur. J. Soil Sci 2005,56, 753–763.

(26) Waychunas, G. A.; Rea, B. A.; Fuller, C. C.; Davis, J. A. Surface-chemistry of ferrihydrite 0.1. EXAFS studies of the geometry ofcoprecipitated and adsorbed arsenate. Geochim. Cosmochim.Acta 1993, 57, 2251–2269.

(27) Bancroft, G. Mossbauer Spectroscopy: An Introduction forInorganic Chemists and Geochemists, Halsted Press: New York,1973.

(28) Cardile, C. M. Tetrahedral Fe-3+ in ferrihydrite - Fe-57Mossbauer spectroscopic evidence. Clays Clay Miner. 1988, 36,537–539.

(29) Murad, E.; Bowen, L. H.; Long, G. J.; Quin, T. G. The influenceof crystallinity on magnetic-ordering in natural ferrihydrites.Clay Miner. 1988, 23, 161–173.

(30) Murad, E.; Schwertmann, U. The influence of aluminumsubstitution and crystallinity on the Mossbauer-spectra ofgoethite. Clay Miner. 1983, 18, 301–312.

(31) Mørup, S.; Ostenfeld, C. W. On the use of Mossbauer spec-troscopy for characterisation of iron oxides and oxyhydroxidesin soils. Hyperfine Interact. 2001, 136, 125–131.

(32) Guyodo, Y.; Banerjee, S. K.; Penn, R. L.; Burleson, D.; Berauo,T. S.; Seda, T.; Solheid, P. Magnetic properties of synthetic six-line ferrihydrite nanoparticles. Phys. Earth Planet. Inter. 2006,154, 222–233.

ES800881W

VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 7897