chemical processes for the extreme enrichment of...

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

ScienceDirect

Geochimica et Cosmochimica Acta 131 (2014) 150–163

Chemical processes for the extreme enrichment of telluriuminto marine ferromanganese oxides

Teruhiko Kashiwabara a,⇑, Yasuko Oishi b, Aya Sakaguchi b, Toshiki Sugiyama b,Akira Usui c, Yoshio Takahashi b,⇑

a Institute for Research on Earth Evolution (IFREE)/Submarine Resources Research Project (SRRP), Japan Agency for Marine-Earth

Science and Technology (JAMSTEC), 2-15 Natsushimacho, Yokosuka, Kanagawa 237-0061, Japanb Department of Earth and Planetary Systems Science, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama,

Higashi-Hiroshima, Hiroshima 739-8526, Japanc Department of Natural Environmental Science, Faculty of Science, Kochi University, Kochi-shi, Kochi 780-8520, Japan

Received 2 May 2013; accepted in revised form 15 January 2014; available online 3 February 2014

Abstract

Tellurium, an element of growing economic importance, is extremely enriched in marine ferromanganese oxides. We inves-tigated the mechanism of this enrichment using a combination of spectroscopic analysis and adsorption/coprecipitation exper-iments. X-ray Absorption Near-Edge Structure (XANES) analysis showed that in adsorption/coprecipitation systems, Te(IV)was oxidized on d-MnO2 and not oxidized on ferrihydrite. Extended X-ray Absorption Fine Structure (EXAFS) analysisshowed that both Te(IV) and Te(VI) were adsorbed on the surface of d-MnO2 and ferrihydrite via formation of inner-spherecomplexes. In addition, Te(VI) can be structurally incorporated into the linkage of Fe octahedra through a coprecipitationprocess because of its molecular geometry that is similar to the Fe octahedron. The largest distribution coefficient obtainedin the adsorption/coprecipitation experiments was for the Te(VI)/ferrihydrite coprecipitation system, and it was comparableto those calculated from the distribution between natural ferromanganese oxides and seawater. Our XAFS and micro-focusedX-ray fluorescence (l-XRF) mapping of natural ferromanganese oxides showed that Te was structurally incorporated asTe(VI) in Fe (oxyhydr)oxide phases. We conclude that the main process for the enrichment of Te in ferromanganese oxidesis structural incorporation of Te(VI) into Fe (oxyhydr)oxide phases through coprecipitation.

This mechanism can explain the unique degree of enrichment of Te compared with other oxyanions, which are mainlyenriched via adsorption on the surface of the solid structures. In particular, the great contrast in the distributions of Te andSe is caused by their oxidized species: (i) the similar geometry of the Te(VI) molecule to Fe octahedron, and (ii) quite solublenature of Se(VI). Coexisting Mn oxide phases may promote structural incorporation of Te(VI) by oxidation of Te(IV), althoughthe surface oxidation itself may not work as the critical enrichment process as in the case of some cations. This enrichment mech-anism also means that ferromanganese oxides mainly scavenge dominant Te(VI) species from seawater and do not affect its spe-cies distribution in seawater, as described in a previous model. The variation in Te abundances and the correlation of Teconcentration with the growth rate of natural ferromanganese oxides are consistent with the coprecipitation mechanism.� 2014 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.gca.2014.01.020

0016-7037/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors. Tel.: +81 46 867 9491; fax: +81 46 8679315 (T. Kashiwabara).

E-mail addresses: [email protected](T. Kashiwabara), [email protected] (Y. Takahashi).

1. INTRODUCTION

Ferromanganese oxides are important materials in themarine geochemistry of trace elements. They are aggregatesof amorphous Fe (oxyhydr)oxides and Mn oxides and are


mailto:[email protected]

mailto:[email protected]


T. Kashiwabara et al. / Geochimica et Cosmochimica Acta 131 (2014) 150–163 151

widely present in forms such as crusts, nodules, and precip-itate particles. Their large surface area and strong affinityfor many elements make them efficient scavengers of tracemetals from seawater. Some types of ferromanganese oxi-des are considered potential resources of valuable metals.It is well known that hydrogenetic ferromanganese crustshave traditionally been called cobalt (Co)-rich crusts. Theyoften cover the surface of seamounts in the form of ex-tended sheets and strongly accumulate many metals of eco-nomic interest, including Co, relative to the Earth’s crust(e.g., Halbach et al., 1989; Hein et al., 2000).

From an economic viewpoint, ferromanganese crustscould properly be called “tellurium (Te)-rich crusts” (Heinet al., 2003). Tellurium is a metalloid located between Seand Po in the p-block of the periodic table. It exhibits cer-tain metallic and non-metallic properties supporting recentdevelopment of a wide range of high-tech industries (e.g.,Brown, 2000; Wachter, 2004; Deng et al., 2007; Ikedaet al., 2009; Graf et al., 2009; Zhang et al., 2009; Heinet al., 2010). Interestingly, Te is extremely enriched inhydrogenetic ferromanganese oxides (e.g., Hein et al.,2003, 2010; Lakin et al., 1963). While other elements includ-ing Co are enriched by 100–250 times over their Earth’scrustal values, the enrichment of Te is �55,000 times, thelargest among all reported elements in ferromanganese oxi-des (Hein et al., 2000, 2003, 2010). Because Te is not minedas a primary ore anywhere in the world, its high concentra-tions in ferromanganese oxides could be a potential re-source for meeting growing Te demands.

Ferromanganese oxides can also be key materials forunderstanding the geochemical behavior of Te in the sur-face environment, which is poorly known owing to the ex-tremely low abundance and difficulty of analysis. Telluriumshows a scavenging-type vertical profile in typical seawater,with concentrations ranging from 0.4 to 1.9 pM (Lee andEdmond, 1985). Tellurium(VI) species predominate overTe(IV) species by factors of 2.0–3.5 in seawater, in whichthe dominant species are likely TeOðOHÞ�3 andTeOðOHÞ�5 (Lee and Edmond, 1985; Byrne, 2002). Someresearchers have suggested that the distribution of Te spe-cies in seawater may be controlled by preferential uptakeof Te(IV) species by ferromanganese oxides (Hein et al.,2003). Other researchers suggest that the anomalous Teabundance in ferromanganese oxides may be related tothe input of interplanetary dust particles (Li et al., 2005).Although several studies have reported such higher Te con-centrations in ferromanganese oxides (�205 ppm) than inany other geological samples (Koschinsky and Hein,2003; Hein et al., 2003; Li et al., 2005; Baturin, 2007; Batu-rin et al., 2007), unsolved mechanisms of the extreme Teenrichment can raise controversial interpretation.

Here, we have to note first that the extremely large valueof Te enrichment factor, [TeFe–Mn crust]/[TeEarth’s crust], ispartly attributed to the quite low abundance of Te inEarth’s crust. To understand the Te abundance in ferro-manganese oxides without any over/underestimation, it isnecessary to reveal the enrichment mechanism of Te intoferromanganese oxides “from seawater”. The value of[TeFe–Mn crust]/[Teseawater] is not as unique as the value of[TeFe–Mn crust]/[TeEarth’s crust], but it still falls among the

most enriched group (Hein et al., 2003, 2010). Especially,the value of [TeFe–Mn crust]/[Teseawater] (�109) is obviouslymuch larger than those of other oxyanions (�107) (e.g.,Takematsu et al., 1990). This emphasizes that an essentialproblem to be solved is the reason why Te shows uniqueenrichment into ferromanganese oxides from seawater com-pared with other oxyanions.

Tellurium is a redox-active oxyanion; thus, surface oxi-dation has been proposed as an explanation for the uniqueenrichment of Te into ferromanganese oxides from seawa-ter (Hein et al., 2003). Oxidation reaction can be a triggerfor efficient enrichment for some cationic elements such asCo, Ce, and Tl (Murray and Dillard, 1979; Manceauet al., 1997; Takahashi et al., 2007; Peacock and Moon,2012). By analogy with these elements, oxidation can alsobe a controlling factor for Te, where Te(IV) in seawater ispreferentially adsorbed onto Fe (oxyhydr)oxides and is sub-sequently oxidized to Te(VI) (Hein et al., 2003).

However, no evidence of the chemical state of Te in mar-ine ferromanganese oxides has been published, yet. Withoutthat information, it is difficult to know the chemical pro-cesses, for example, whether Te(IV) is preferentially incor-porated into ferromanganese oxides and whether theoxidation reaction of Te(IV) can actually occur. Such fun-damental information can be provided by a combinationof adsorption/coprecipitation experiments and X-rayAbsorption Fine Structure (XAFS) analysis to understandthe unique behavior of Te at the seawater/ferromanganeseoxide interface.

In this paper, we present new insights into mechanism ofthe extreme enrichment of Te into ferromanganese oxides.We investigated (i) the chemical state of Te in syntheticFe (oxyhydr)oxides, Mn oxides, and natural ferromanga-nese oxides, and (ii) the macroscopic distribution of Te(IV)and Te(VI) on Fe (oxyhydr)oxides and Mn oxides, respec-tively. The findings on the chemical process at the seawater/ferromanganese oxide interface will develop understandingof the roles of marine ferromanganese oxides in the geo-chemical cycle of trace elements.

2. MATERIALS AND METHODS

2.1. Samples

Hydrogenetic ferromanganese oxides were collectedfrom two sites in the Pacific Ocean: D535 in the SouthPacific Ocean (13.0�S, 159.2�W, 5222 m depth), andAD14 around the Marshall Islands (14.1�S, 167.2�W,1617 m depth) (Takahashi et al., 2000, 2007; Kashiwabaraet al., 2008, 2009, 2010, 2011, 2013). These samples wereair-dried, and a part of them were polished to make ca.30-lm-thick sections on glass slides for two-dimensionalmicro-focused X-ray fluorescence (l-XRF) mapping.Another part of the samples was ground into powder usingan agate mortar. Elemental concentrations were determinedby ICP-MS (Agilent technologies, Agilent 7500). Amongmajor elements, Mn was 27.9 wt% and 29.5 wt% and Fewas 16.4 wt% and 16.7 wt% for D535 and AD14, respec-tively. Concentrations of oxyanions, including Te, aresummarized in Table 1. Chemical compositions of these

Table 1Concentration of oxyanions in natural ferromanganese oxides(mg k�1).

Sample Te Se As Sb Mo W

D535 31.7 10.1 157 26.6 284 39.0AD14 45.4 11.3 194 29.3 459 24.5

152 T. Kashiwabara et al. / Geochimica et Cosmochimica Acta 131 (2014) 150–163

samples are typical of the hydrogenetic type among threemajor types; hydrogenetic, diagenetic, and hydrothermaltypes (Puteanus and Halbach, 1988; Takahashi et al.,2000; Usui, 1995).

For XAFS measurements, solutions of 2000 mg kg�1

Te(IV) and Te(VI) were used as reference samples. Eachsolution was prepared by dissolving Na2TeO3 and Na2H4-

TeO6 (Wako, Japan) in Milli-Q water. Tellurium adsorp-tion samples were prepared by adding stock solution ofTe(IV) or Te(VI) to 100 mg synthetic ferrihydrite or d-MnO2. Molar ratios of Te/Fe (or Te/Mn) were controlledto ca. 0.017 in 250 mL NaCl solution system. The ionicstrength of the experimental solution was controlled to0.70 M (NaCl), and the pH was adjusted to 8.0 with HCland NaOH solutions. Suspensions were filtered after 24 hof equilibration in a shaking bath at 25 �C. The solid phaseswere rinsed thoroughly with Milli-Q water to remove back-ground solution, and packed into polyethylene bags. Tellu-rium coprecipitation samples were also prepared by thesimultaneous addition of Te into the Fe(II) or Mn(II) solu-tions in the synthetic procedures of ferrihydrite and d-MnO2. Molar ratios of Te/Fe (or Te/Mn) in the systemswere controlled to ca. 0.017.

Ferrihydrite and d-MnO2 were synthesized following themethods of Schwertmann and Cornell (2000) and Fosteret al. (2003), respectively. Ferrihydrite was synthesized byadjusting a solution containing 36 mM Fe(NO3)3 � 9H2Oto pH 8.0 through dropwise addition of 4.0 M NaOH.The solids were recovered by filtration using a membranefilter of 0.2 lm pores, rinsed thoroughly with Milli-Q water,freeze-dried, and then collected as solids. d-MnO2 was pre-pared by rapid oxidation of a 30 mM MnCl2 solution withan equal volume of a 20 mM KMnO4 solution at pH 10.The resultant suspensions were recovered by filtration, rins-ing, and air-drying. These solids were characterized byXRD.

2.2. XAFS measurements and data analysis

Tellurium K-edge (31.811 keV) XAFS was measured atBL01B1 in SPring-8 (Hyogo, Japan). The white beam fromthe bending magnet was monochromatized using a Si(311)double crystal monochromator. Two mirrors with Rh coat-ing, placed upstream and downstream of the monochroma-tor, were used for collimation and focusing of the X-raybeam. Harmonic rejection was achieved by adjusting theglancing angle of these two mirrors. The X-ray energy wascalibrated using the peak of Na2H4TeO6 at 31.811 keV.The beam size was adjusted to ca. 1 mm (vertical,V) � 1 mm (horizontal, H) by slits. The spectra of standardsolutions were measured in the transmission mode by ioniza-tion chambers. Natural, adsorption, and coprecipitation

samples were measured in fluorescence mode using a19-element Ge solid-state detector. All measurements wereconducted under ambient conditions.

Analysis of XAFS spectra was performed usingREX2000 ver. 2.5 (Rigaku Co.) and FEFF 8.5 (Zabinskyet al., 1995). Quantitative analysis of XANES wasperformed by linear combination fitting using Te(IV) andTe(VI) solutions as end members in the range of31.800–31.835 keV. The experimental EXAFS oscillations,v(k), were extracted from the raw spectrum by a splinesmoothing method. The k3v(k) oscillations were Fourier-transformed from k space to R space to obtain radialstructural functions (RDFs), and then back-transformedto k space to analyze by a curve-fitting method. Theoreticalphase shifts and amplitude functions for simulation of thespectra were extracted from model compounds such as Fe2-

TeO6 and Mn3TeO6. The quality of the fit was given by thegoodness of fit parameter, R factor, defined as

R ¼Xfk3vobsðkÞ � k3vcalðkÞg

2=XfvobsðkÞg

2 ð1Þ

where vobs(k) and vcal(k) are the experimental and calcu-lated absorption coefficients at a given wavenumber k,respectively.

2.3. l-XRF mapping

Micro-focused X-ray fluorescence (l-XRF) mapping ofnatural ferromanganese oxides was performed at BL37XUin SPring-8 (Hyogo, Japan). The beam from the undulatorwas monochromatized at 35 keV using a Si(111) doublecrystal monochromator. The incident X-ray was focusedby a K–B mirror system to the size of 1.1 lm(V) � 1.2 lm (H). Samples were mounted on an XY samplestage placed at 10� to the incident X-ray beam, and werespatially rastered by step size of 5.0 lm (V) � 5.0 lm (H)using stepping motors. The intensity of the incident X-ray(I0) was measured using an ionization chamber filled withair. Fluorescence X-rays from samples were monitored bya Si(Li) solid-state detector placed at 90� to the incidentX-ray beam, and integrated for 0.2 s at each point. Thesemeasurements were carried out under ambient conditions.

The integrated peak intensity of the XRF line of eachelement per pixel was normalized by I0. Elemental mapsfor the measurement area were constructed using the 256gradation color scale with red denoting the highest intensi-ties and blue the lowest.

2.4. Adsorption/coprecipitation experiments

Two series of batch experiments were conducted. Thefirst series was for establishing adsorption isotherms using1 mg adsorbents (ferrihydrite or d-MnO2) and 10 mL Tesolution. The solution was adjusted to pH 8.0 with an ionicstrength of 0.70 M (NaCl). Initial Te concentrations wereadjusted to 1.0, 5.0, 10, and 20 mg L�1. The second serieswas to provide a comparison between adsorption andcoprecipitation for ferrihydrite. While the adsorption is areaction on the surface of the solid, coprecipitation isthree-dimensional incorporation in the solid structure dur-ing the formation of solid precipitates (i.e. ferrihydrite). To

Eenergy / eV31770 31795 31820 31845 31870

Nor

mal

ized

Ads

orpt

ion

(a) Natural (D535)

(f) COP Te(IV) / Fh

(e) ADS Te(IV) / Fh

(d) COP Te(VI) / Fh

(c) ADS Te(VI) / Fh

(j) COP Te(IV) / Mn

(i) ADS Te(IV) / Mn

(l) Te(IV) solution

(k) Te(VI) solution

(h) COP Te(VI) / Mn

(g) ADS Te(VI) / Mn

(b) Natural (AD14)

Fig. 1. Tellurium K-edge XANES spectra for natural ferroman-ganese oxides and synthetic analogues. FMO: ferromanganeseoxide, ADS: adsorption, COP: coprecipitation, Fh: ferrihydrite,Mn: d-MnO2. Dotted lines on each spectrum are simulation spectraobtained by linear combination fitting.

Table 2Contribution of end members determined by liner combinationfitting of XANES.

Sample Te(IV) (%) Te(VI) (%) Ra

(a) Natural (D535) 3 97 0.043(b) Narutal (AD14) 0 100 0.028(c) ADS Te(VI)/Fh 8 92 0.023(d) COP Te(VI)/Fh 8 92 0.026(e) ADS Te(IV)/Fh 100 0 0.037(f) COP Te(IV)/Fh 96 4 0.036(g) ADS Te(VI)/Mn 9 91 0.026(h) COP Te(VI)/Mn 8 92 0.023(i) ADS Te(IV)/Mn 14 86 0.017(j) COP Te(IV)/Mn 2 98 0.014

aR factor (%) defined as Eq. (1). Fitting was performed in the rangeof 31800–31835 eV. ADS: adsorption, COP: coprecipitation, Fh:ferrihydrite, Mn: d-MnO2.


distinguish these processes, the sequence of Te addition waschanged under identical conditions. First, 10 mL Fe(II)solutions of 18.7 mM were prepared by dissolvingFe(NO3)3 � 9H2O. For coprecipitation experiments, a stocksolution of Te(IV) or Te(VI) was added to obtain a molarratio Te/Fe of ca. 0.017, and then pH was adjusted to 8.0to synthesize precipitation of ferrihydrite. For adsorptionexperiments, the pH of the Fe(II) solution was first adjustedto 8.0 to synthesize ferrihydrite, and then the Te solutionwas added. The surface area of ferrihydrite precipitatesshould be as nearly identical as possible to compare theadsorption and coprecipitation systems. To evaluate this ef-fect, Se(VI) was also added to each system in the same wayas Te addition. Addition of Se(VI) did not affect the incor-poration of Te into ferrihydrite because its adsorption,through outer-sphere complexation, is quite weak (Haradaand Takahashi, 2008).

The stock solution of 1000 mg L�1 Te was prepared bydissolving Na2TeO3 and Na2H4TeO6 (Wako, Japan) inMilli-Q water. Adequate amount of the stock solutionwas added to each system obtain the desired Te concentra-tions. The pH was maintained at desired values by adding0.10 M NaOH or HCl. The stock solution of Se(VI) wasprepared by dissolving Na2SeO4 (Wako, Japan). The exper-iments were conducted in 15 mL polystyrene vessels, whichwere shaken at 135 rpm in a shaking water bath. After 24 hof equilibration at 25 �C, samples were filtered by 0.2 lmdisposable syringes. The filtrates were diluted and concen-trations of Te and Se were measured by ICP-MS. All pro-cedures were repeated three times to estimate theprecision of the experiments.

3. RESULTS

3.1. Te K-edge XANES

Tellurium K-edge XANES spectra are shown in Fig. 1.The spectra of reference materials show a chemical shift,that is, the peak energy of Te(VI) solution (Fig. 1(k)) was2.7 eV higher than that of Te(IV) solution (Fig. 1(l)). Thisrelationship can be utilized to estimate the dominant oxida-tion state of Te species in the solid phases.

The peak energies for Te adsorbed on d-MnO2 (Fig. 1(g)and (i)) were at the same position as that of the Te(VI) solu-tion (Fig. 1(k)). This means that the oxidation state of Teadsorbed on d-MnO2 was Te(VI) regardless of its oxidationstate in the solution. On the other hand, peak energies forTe adsorbed on ferrihydrite had the same positions as inthe experimental solutions (Fig. 1(c) and (e)). These resultsindicate that the oxidation reaction of Te(IV) to Te(VI) oc-curred only on the surface of d-MnO2 and not onferrihydrite.

The same trend is also shown in the case of coprecipita-tion samples. Although the peak energies for Te coprecipi-tated with d-MnO2 (Fig. 1(h) and (j)) were in the samepositions as that for the Te(VI) solution (Fig. 1(k)), peakenergies for Te coprecipitated with ferrihydrite (Fig. 1(d)and (f)) were the same as those for the experimental solu-tions. This means that Te(IV) in solution was oxidized dur-ing the precipitation of d-MnO2, but not during the

precipitation of ferrihydrite. In addition, the peak energiesfor the natural ferromanganese oxide (Fig. 1(a) and (b))were at the same position as the peak energy for the Te(VI)solution (Fig. 1(k)), indicating that Te was present as itsoxidized form, Te(VI), in natural ferromanganese oxides.

To investigate possible effects of minor components ofTe species, we also performed linear combination fitting(LCF) of these spectra using Te(IV) and Te(VI) solutionsas the end members. The fitting spectra are shown as red


dotted lines in Fig. 1, and the proportions of Te(IV) andTe(VI) are listed in Table 2. We obtained good fits, whichshows less than 10% of the Te(IV) fractions in natural andsynthetic samples of the Te(VI) system. In synthetic sam-ples of Te(IV), �14% of Te(IV) was obtained for adsorp-tion and 2% for coprecipitation on d-MnO2, whereasalmost 100% of Te(IV) was obtained for both adsorptionand coprecipitation on ferrihydrite. Although �14% ofTe(IV) adsorbed on d-MnO2 may be somewhat large, thisfraction did not affect the structural parameters obtainedby EXAFS. The fractions of Te(IV) estimated in ourTe(VI) adsorption/coprecipitation samples can be causedby the slight difference in chemical state between adsorbedand dissolved species. Thus, we conclude that these com-ponents did not affect the analyses, including XANESand EXAFS.

3.2. Te K-edge EXAFS

Fig. 2A shows k3-weighted v(k) spectra of Te K-edgeEXAFS and Fig. 2B shows their radial structural functions(RSFs; phase shift not corrected) for the Te(VI) system.The features of k-space spectra indicated by dotted lines(i)–(iii) suggest that there are more than two components.The k3v(k) spectra for ferrihydrite (Fig. 2A(a) and (b))show small shoulders at (i) and (ii), whereas those for d-MnO2 (Fig. 2A(c) and (d)) show broader negative peaksaround (iii). The RSFs (Fig. 2B) show these features in adifferent way: All the spectra have one prominent peak

(i) (ii)(iii)

(b)

(c)

(d)

(e) Te-O

(f) Te-Fe1 + Te-Fe2

(g) Te-Mn1 + Te-Mn2

(a)

Natural

Natural (AD14)

(D535)

3 5 7 9 11 13k Å-1

A

k3 χ(k

)

Fig. 2. Tellurium K-edge EXAFS spectra for Te(VI)-adsorption/copreci(phase shift not corrected). Natural: natural ferromanganese oxide, ADSAdsorption on (a) ferrihydrite and (c) d-MnO2, and coprecipitation with (filtered k3v(k) of (e) Te–O, (f) Te–Fe1 + Te–Fe2 shells in spectrum (b), andspectrum are simulation spectra obtained by curve-fitting analysis.

around R + DR = 1.5 A ascribed to Te–O bonds and fur-ther distant peaks originating from Te–Fe/Mn. The peakheight and width of the Te–O shells are essentially identicalamong Fig. 2B(a)–(d). Thus, the distant shells appear tocause the difference in k3v(k) spectra of ferrihydrite and d-MnO2 which can be indicators of the host phase of Te.

Fig. 2A(e)–(g) also show the inversely Fourier-filteredcontributions of Te–O, Te–Fe, and Te–Mn shells to thek3v(k) spectra from Fig. 2B(b) and (c). The features ofthe k3v(k) spectra at lines (i)–(iii) in Fig. 2A clearly overlapthe peaks in the Te–Fe and Te–Mn waves. For ferrihydrite,the shoulders appearing at (i) and (ii) in the spectrum ofcoprecipitation (Fig. 2A(b)) are larger than those in thespectrum of adsorption (Fig. 2A(a)). This difference corre-sponds to the difference between the peaks of the secondshell in the RSFs for coprecipitation (Fig. 2B(b)) and foradsorption (Fig. 2B(a)), with the peak for coprecipitationbeing larger. Thus, the contributions of the Te–Fe shellsare larger for coprecipitated Te species than for adsorbedTe species.

In contrast, significant differences are not observed be-tween the k3v(k) spectra of adsorption and coprecipitationsamples for d-MnO2 (Fig. 2A(c) and (d)). While the broaderoscillations around (iii) correspond to the Te–Mn waves inthe inversely Fourier-filtered spectra (Fig. 2A(g)), the mag-nitudes of their contributions seem to be almost the same(Fig. 2A(c) and (d)). This feature is also shown in theirRSFs with Te–Mn shells having similar heights and widths(Fig. 2B(c) and (d)).

(a) ADS Te(VI) / Fh

(b) COP Te(VI) / Fh

(c) ADS Te(VI) / Mn

(d) COP Te(VI) / Mn

FT m

agni

tude

B

R + ΔR / Å0 1 2 3 4 5

pitation samples: (A) k3-weighted v(k) spectra, and (B) their RSFs: adsorption, COP: coprecipitation, Fh: ferrihydrite, Mn: d-MnO2.b) ferrihydrite and (d) d-MnO2. (e)–(g) in (A) are inversely Fourier-

(g) Te–Mn1 + Te-Mn2 shells in spectrum (c). Dotted lines on each


We obtained structural parameters for the experimentalTe species by curve-fitting (Table 3), and simulation spectraobtained using these parameters are shown as red dottedlines on each spectrum in Fig. 2. All samples for the Te(VI)system show six O atoms at the same distance from the Teatom (RTe–O = 1.92 A) (Table 3). The further distant shellsin the RSFs were simulated by two Te–Fe or Te–Mn shells.The obtained interatomic distances are 3.06–3.08 A (RTe–

Fe1) and 3.47–3.50 A (RTe–Fe2) for Te(VI)/ferrihydrite sys-tems, and 3.00 A (RTe–Mn1) and 3.42–3.47 A (RTe–Mn2) forTe(VI)/d-MnO2 systems. Considering the Te(VI) moleculeas an octahedron, the shorter distance in each system(RTe–Fe1 and RTe–Mn1) likely corresponds to the edge-shar-ing linkage, and the longer distance (RTe–Fe2 and RTe–Mn2)corresponds to the double corner-sharing linkage (e.g.,Manceau and Drits, 1993; Waychunas et al., 1993; Kas-hiwabara et al., 2011). The slightly shorter distances forTe–Mn than for Te–Fe may reflect the smaller ionic radiusof Mn4+ than Fe3+. The estimates of coordination number(CN) show that edge-sharing linkages clearly predominatein the Te(VI)/ferrihydrite system (Table 3), and Te(VI)coprecipitated with ferrihydrite has larger CNTe–Fe1 andCNTe–Fe2 than Te(VI) adsorbed on ferrihydrite. The CNsof Te–Mn shells, on the other hand, are almost the samefor adsorption and coprecipitation samples, implying anabsence of coprecipitated species as well as similar ratiosof edge-sharing and double corner-sharing linkages for d-MnO2. Here, contributions of Te–Te shells, which may beobserved with surface precipitation, were not found inour experimental systems.

We also succeeded in obtaining EXAFS spectra of Tespecies in natural ferromanganese oxide samples (AD14

Table 3Structural parameters of Te species obtained for adsorption/coprecipitat

Sample Shell CN

Te(VI) adsorbed on ferrihydrite Te–O 6.0Te–Fe1 0.9Te–Fe2 0.5

Te(VI) coprecipitated with ferrihydrite Te–O 6.3Te–Fe1 2.2Te–Fe2 1.0

Te(VI) adsorbed on d-MnO2 Te–O 6.1Te–Mn1 0.9Te–Mn2 1.4

Te(VI) coprecipitated with d-MnO2 Te–O 6.4Te–Mn1 1.2Te–Mn2 1.0

Te(IV) adsorbed on ferrihydrite Te–O 4.3Te–Fe 1.4

Te(IV) coprecipitated with ferrihydrite Te–O 4.4Te–Fe 1.6

Te(IV) adsorbed on d-MnO2 Te–O 6.6Te–Mn1 1.3Te–Mn2 0.7

Te(IV) coprecipitated with d-MnO2 Te–O 6.9Te–Mn1 1.0Te–Mn2 1.5

Parameters were optimized by minimizing the R factor (%) defined as Eq.for coordination numbers.

and D535) through repeated collection of the spectra(Fig. 2A). Although their spectral qualities were still poorbecause of the low concentration of Te, the same shouldersalong lines (i) and (ii) described for ferrihydrite were pres-ent. Thus, we conclude that the host phase of Te(VI) speciesin natural ferromanganese oxides is ferrihydrite. Becausethese shoulders showed a better match to those for Te(VI)coprecipitated with ferrihydrite (Fig. 2A(b)), it seems thatthe dominant species in our natural ferromanganese oxidesamples is Te(VI) coprecipitated with ferrihydrite.

Tellurium K-edge EXAFS spectra were also analyzedfor the experimental system using the Te(IV) solution(Fig. 3 and Table 3). Their k3v(k) spectra in Fig. 3A arecomposed of Te–O and Te–Fe/Mn shells, as shown inFig. 3B. The spectra for d-MnO2 (Fig. 3A, B(c) and (d))were similar to those for the Te(VI) system (Fig. 2A, B(c)and (d)) because Te(IV) was oxidized on d-MnO2. In con-trast, the spectra for ferrihydrite (Fig. 3A, B(a) and (b))were entirely different from those of the Te(VI) system(Fig. 2A, B(a) and (b)) because of the different oxidationstate of Te. On the other hand, the spectra of adsorptionand coprecipitation samples for Te(IV) were essentiallythe same for both ferrihydrite (Fig. 3A, B(a) and (b)) andd-MnO2 (Fig. 3A, B(c) and (d)).

Quantitative analysis (Table 3) shows that Te(IV)species on ferrihydrite have four O atoms at �1.87 A.The distances of Te–Fe shells are 3.47–3.50 A, likely corre-sponding to the double corner-sharing linkage (e.g.,Manceau and Drits, 1993; Waychunas et al., 1993;Kashiwabara et al., 2011). The CNs of these shells are closeto 2, and other contributions of Te–Fe shells were not ob-served. This implies that most of the Te(IV) species was

ion systems for ferrihydrite/d-MnO2.

R (A) DE0 (eV) r2 R factor

1.92 7.5 0.004 0.7803.06 0.0043.47 0.0041.92 6.4 0.004 0.7883.08 0.0043.50 0.0041.92 6.6 0.004 0.7553.00 0.0043.42 0.0041.92 6.6 0.004 1.7023.00 0.0053.47 0.0051.87 9.4 0.004 1.1083.50 0.0091.87 8.0 0.004 0.6093.47 0.0091.92 7.0 0.005 2.1052.99 0.0053.49 0.0051.92 6.0 0.006 1.9272.99 0.0053.44 0.005

(1). Standard errors in EXAFS are ±0.02 A for distances and ±20%

(a)

(b)

(c)

(d)

(a) ADS Te(IV) / Fh

(b) COP Te(IV) / Fh

(c) ADS Te(IV) / Mn

(d) COP Te(IV) / Mn

FT m

agni

tude

B

R + ΔR / Å3 5 7 9 11 13

k Å-1

A

k3 χ(k

)

0 1 2 3 4 5

Fig. 3. Tellurium K-edge EXAFS spectra for Te(IV)-adsorption/coprecipitation samples: (A) k3-weighted v(k) spectra, and (B) their RSFs(phase shift not corrected). ADS: adsorption, COP: coprecipitation, Fh: ferrihydrite, Mn: d-MnO2. Dotted lines on each spectrum aresimulation spectra obtained by curve-fitting analysis.


in the double corner-sharing complex. In contrast, thestructural parameters of Te(IV) species on d-MnO2 weresimilar to those of Te(VI)/d-MnO2 systems, where bothedge-sharing and double corner-sharing complexes wereapparently present. No systematic differences were foundbetween coprecipitation and adsorption samples for Te(IV)systems, even though 14% of Te(IV) is possible in Te(IV)adsorption on d-MnO2 by XANES estimation. Te(IV) doesnot seem to be incorporated into ferrihydrite and d-MnO2

as a coprecipitated species.

3.3. l-XRF mapping for natural ferromanganese oxide

Elemental maps in natural ferromanganese oxide byl-XRF mapping are shown in Fig. 4. Typical columnarstructure was focused on (Fig. 4(a)). Manganese is uniformlydistributed in the columns and relatively sparse between col-umns (Fig. 4(c)). On the other hand, Fe is comparably enrichedbetween columns (Fig. 4(b)) and uniformly distribution inthe columns. Different trends are also observed in the mapsof Mn and Fe around the base of the columnar structureon the lower right corner. Intensity of the Mn Ka line in thatarea seems to be slightly lower than that in the columnarstructure, while the intensity of the Fe Kb line in that areais slightly higher than the columnar structure.

The Ka maps of Te, Sb, and Mo (Fig. 4(d)–(f)) were ob-tained by using high-energy X-ray (35 keV). Although theirhigher energy results in a deeper escape depth relative tothose of Fe (Kb: 7.058 keV) and Mn (Ka: 5.895 keV), it

does not hamper mutual comparisons of Te (Ka:27.382 keV), Sb (Ka: 26.279 keV), and Mo (Ka:17.446 keV). Molybdenum (Fig. 4(f)), which is retained inMn oxide phases, showed a distribution pattern similar toMn (Fig. 4(c)) (Kashiwabara et al. 2011). On the otherhand, Te (Fig. 4(d)) and Sb (Fig. 4(e)) both showed a sim-ilar contrasting distribution patterns to Mo. Antimony,which is considered to be retained in Fe (oxyhydr)oxidephases, showed a pattern similar to Fe (Fig. 4(d)) (Kas-hiwabara et al. 2008; Mitsunobu et al. 2010). These mapsindicate that Te is mainly retained in Fe (oxyhydr)oxidephases, which is consistent with the EXAFS analysis de-scribed in Section 3.2.

3.4. Adsorption/coprecipitation experiments for Te to

ferrihydrite and d-MnO2

Fig. 5 shows adsorption isotherms for Te on ferrihydriteand d-MnO2 at pH 8.0 and I = 0.70 M (NaCl). All adsorp-tion isotherms seem to be of Langmuir or Freundlich type,and surface precipitation does not seem to occur in thesesystems. Ferrihydrite (Fig. 5(a) and (b)) showed largeradsorption than d-MnO2 (Fig. 5(c) and (d)) for both Te(VI)and Te(IV). For both solids, Te(IV) (Fig. 5(a) and (c))showed larger adsorption than Te(VI) (Fig. 5(b) and (d)).Thus, Te(IV) adsorption on ferrihydrite (Fig. 5(a)) tendto be the largest among these adsorption systems underour experimental conditions and Te(VI) on d-MnO2

(Fig. 5(d)) tend to be the lowest.

(d) Te (e) Sb (f) Mo

(a) Optical image (b) Fe (c) Mn

High Low Normalized Intensity of XRF

150 µµµµm

Fig. 4. XRF maps of natural ferromanganese oxide (AD14). Measurement area is within the red square of (a). XRF maps: (b) Fe Kb, (c) MnKa (d) Te Ka, (e)Sb Ka, (f) Mo Ka. Red means higher concentration blue means lower concentration. Intensities are normalized by theintensity of the incident X-ray.

0

200

400

600

800

1000

1200

1400

1600

0 50 100 150 200

(a) Te(IV) on ferrihydrite

(b) Te(VI) on ferrihydrite

(c) Te(IV) on δ-MnO2

(d) Te(VI) on δ-MnO2

Equilibrium Te concentration in solution (μmol L-1)

Ads

orbe

d am

ount

of T

e (m

mol

kg-1

)

Fig. 5. Adsorption isotherms for Te on ferrihydrite and d-MnO2 atpH 8 (I = 0.70 M).

log [Kd (mL g-1)]2.5 3.0 3.5 7.5 8.5 9.58.0 9.0 10

(c) Te(IV) adsorption

(b) AD14

(a) D535SeTe

(d) Te(IV) coprecipitation

(e) Te(VI) adsorption

(f) Te(VI) coprecipitation

experimental

natural

Fig. 6. Comparison of distribution coefficients between adsorptionand coprecipitation for Te/ferrihydrite systems. Apparent distri-bution coefficients for natural ferromanganese oxides (AD14 andD535) are also plotted based on Te concentration in seawater (Leeand Edmond, 1985). Estimated errors (average values ±2r) aresmaller than symbols except in (f).


Fig. 6 shows logarithm of the distribution coefficients,Kd = Csolid/Csolution, obtained in adsorption and


coprecipitation experiments for ferrihydrite under identicalconditions (molar ratio of Te/Fe of Mn = 0.017). These Kdvalues correspond to the lower equilibrium solution con-centrations in isotherms (�0.0036 lmol L�1 in Fig. 5). Sele-nium(VI), which was added for evaluation of the effect ofsurface area, showed a similar distribution in all of theexperiments, indicating that the effective surface area of fer-rihydrite precipitates was almost the same. On the otherhand, Kd values for Te species in coprecipitation experi-ments (Fig. 6(d) and (f)) are significantly larger than thosein adsorption experiments (Fig. 6(c) and (e)). Especially,Te(VI) shows the largest distribution in ferrihydritethrough coprecipitation process (Fig. 6(f)). The Kd valuefor coprecipitation of Te(VI) with ferrihydrite (Fig. 6(f))is comparable to those (Fig. 6(a) and (b)) calculated fromTe concentrations in natural ferromanganese oxides (Ta-ble 1) and seawater (=0.4 pM; Lee and Edmond, 1985).

4. DISCUSSION

4.1. Chemical processes for Te enrichment into

ferromanganese oxides

Previous study suggested that surface oxidation ofTe(IV) on Fe (oxyhydr)oxides is the most important pro-cess for Te enrichment into ferromanganese oxides (Heinet al., 2003). Our XANES analysis showed that Te is pres-ent exclusively as Te(VI) in natural ferromanganese oxidesin spite of the presence of both Te(IV) and Te(VI) in mod-ern oxic seawater (e.g., Lee and Edmond, 1985; Byrne,2002). However, in our synthesized adsorption systems, oxi-dation of Te(IV) did not occur on ferrihydrite, although itdid occur on d-MnO2. This result indicates that other pro-cesses with Te(VI) as a final Te species are important in-stead of Te(IV) adsorption on ferrihydrite. Two pathwaysare possible: (i) oxidation of Te(IV) to Te(VI) via d-MnO2, and (ii) coprecipitation of Te(VI) into the structureof ferrihydrite or d-MnO2 without redox transformation.

EXAFS spectra of natural ferromanganese oxidesshowed that the host phase of Te is ferrihydrite, and thesimilar distributions of Te and Fe in l-XRF mapping sup-ports this finding. In addition, the EXAFS spectra of natu-ral samples showed similar features to Te(VI)coprecipitated with ferrihydrite, which reflects the strongercontribution of Te–Fe shells in coprecipitated species thanin adsorbed species. Although the spectral quality of EX-AFS were still not good enough, we consider that the mainTe species in natural ferromanganese oxides is Te(VI)coprecipitated with ferrihydrite, which is incorporated fromseawater without redox transformation.

The differences between adsorbed and coprecipitatedspecies suggested by molecular characterization are as fol-lows. Adsorption samples were prepared by adding Teion to the aqueous phase after precipitation of ferrihydrite.Thus, Te should form surface complexes on the surface ofsolid structures. We found that, in adsorption systems, bothTe(IV) and Te(VI) formed inner-sphere complexes on fer-rihydrite. It seemed that Te(IV) and Te(VI) preferred dou-ble corner-sharing and edge-sharing linkages onferrihydrite, respectively. This preference may be due to

differences in their molecular geometry (CNTe–O = 4, RTe–

O = �1.87 A for Te(IV); CNTe–O = 6, RTe–O = 1.92 A forTe(VI)).

In the coprecipitation experiments, on the other hand,Te ion was already present in solution during the formationof ferrihydrite precipitates. In this system, Te can be struc-turally incorporated in the linkages of Fe octahedra. Cer-tainly, Te(VI) seems to be isomorphously substituted tothe octahedral Fe(III) site in ferrihydrite (CNFeIII–O = 6,FeIII–O = 1.94–2.05 A; Waychunas et al., 1993; Manceauand Drits, 1993). The stronger contribution of Te–Fe shellsin coprecipitation samples compared with adsorptionsamples in the EXAFS spectra are evidence of this struc-tural incorporation. The interatomic distances of Te–Feshells (RTe–Fe1 = 3.08 A and RTe–Fe2 = 3.50 A) are also sim-ilar to those of Fe–Fe in ferrihydrite (Manceau and Drits,1993), indicating a similar attachment geometry withoutchange of the basic ferrihydrite structure. One may pointout that the CNs of TeVI–Fe are relatively smaller thanthose of Fe–Fe shells in pure ferrihydrite. This differencemay reflect (i) the increased disorder in Fe–Fe linkages inthe ferrihydrite structure due to Te(VI) incorporationand/or (ii) the average nature of EXAFS showing themixing of adsorbed species in the Te(VI) coprecipitatedwith ferrihydrite (Hansel et al., 2003; Mitsunobu et al.,2010).

In contrast, the differences in structure between theTe(IV) molecule and the Fe(III) octahedron(RTe–O = �1.87 A and/or CNTe–O = 4) may not allow itsstructural incorporation into ferrihydrite. We found nospectroscopic signature of structural incorporation ofTe(IV) into ferrihydrite in our coprecipitation experiments.The similarity of spectra in adsorption and coprecipitationsamples indicates that the uptake of Te(IV) occurs on solidsurfaces. Thus, adsorption is the main process for enrich-ment of Te(IV) in ferrihydrite. The same considerationmay be true for Te/d-MnO2 systems, which also did notshow significant differences between adsorption and copre-cipitation samples for either Te(IV) or Te(VI). Slight differ-ences in the geometry of the Mn4+ octahedron(RMn–O = 1.90 A, CNMn–O = 6; Foster et al., 2003) fromthe Fe3+ octahedron and/or an affinity for Te ions can ac-count for the smaller amount of structural incorporation ofTe into d-MnO2.

These molecular-scale insights are consistent with mac-roscopic distributions. The largest distribution was ob-tained for coprecipitation of Te(VI) with ferrihydrite incomparison among adsorption and coprecipitation experi-ments for each Te species. Moreover, this distribution coef-ficient is comparable to that between seawater and naturalferromanganese oxides (Fig. 6). This indicates that copre-cipitation of Te(VI) with ferrihydrite can occur in the natu-ral environment, supporting the EXAFS evidence fornatural samples. The experimental results also show thatTe enrichment is possible without assuming the input ofinterplanetary dust particles, as suggested by someresearchers (Li et al., 2005). Therefore, we consider thatstructural incorporation of Te(VI) in ferrihydrite throughcoprecipitation can be the most important process for Teenrichment into natural ferromanganese oxides.


A large number of cations are known to be structurallyincorporated into Fe(III) (oxyhydr)oxides (e.g., NiII, CuII,ZnII, CdII, CrIII, VIII, GaIII, CoIII, TiIV, GeIV, PV, AsV,and SbV; e.g., Giovanoli and Cornell, 1992; Martin et al.,1997; Manceau et al., 2000; Martinez and McBride, 2001;Carvalho-E-Silva et al., 2003; Wells et al., 2006; and refer-ences cited therein). Although factors affecting the extent ofincorporation are not conclusive (e.g., Manceau et al., 2000;Klas et al., 2011; Liu et al., 2012), they include the valenceof the ion, the similarity of ionic radii, and the geometry ofmolecules (e.g., Fuller et al., 1993; Waychunas et al., 1993;Manceau et al., 2000; Mitsunobu et al., 2010). In the case ofTe(VI), the structure of the Te(VI) octahedron closelymatches that of the Fe octahedron in ferrihydrite. Thus,we consider that structural incorporation of Te(VI) into fer-rihydrite is possible. The large difference in charge betweenTe6+ and Fe3+ might inhibit the extent of Te(VI) substitu-tion into ferrihydrite; however, this factor can be compen-sated to an extent by adsorption of other anions or watermolecules, considering that the absolute concentration ofTe in natural ferromanganese oxides is very low (molar ra-tio of Te/Fe = �2.0 � 10�4).

4.2. Molecular-scale insights into the unique behaviors of Te

compared with other oxyanions at the seawater/

ferromanganese oxide interface

One of our interests is the reason why Te shows a uniqueenrichment in ferromanganese oxides compared with otheroxyanions. Although additional EXAFS data of naturalsamples with higher quality are still desirable, we can pro-vide molecular explanations at present for the uniquebehavior of Te at the seawater/ferromanganese oxide inter-face considering the varieties of chemical reactions foroxyanions.

Enrichment of Te in natural ferromanganese oxidesseems to result from the structural incorporation of Te(VI)into Fe (oxyhydr)oxide phases. This type of reaction is wellknown as a process for accumulating cationic elements innatural ferromanganese oxides (Cronan, 1980). It is typicalthat Co(III) is substituted isomorphously into the Mn sitein hydrogenetic phyllomanganates (Manceau et al., 1987,1997). Lithium (Li+), Ni2+, Cu2+, and Zn2+ are also knownto be incorporated in the crystal structures of diageneticallyformed Mn oxides, such as todorokite (Jiang et al., 2007).These cations show high average concentrations in ferro-manganese oxides, in spite of their small degrees of adsorp-tion, because three-dimensional incorporation in solidstructures is a more effective uptake process than surfaceadsorption (e.g., Crawford et al., 1993a,b; Fuller et al.,1993; Karthikeyan et al., 1997; Mitsunobu et al., 2010).This mechanism can be explained by the similarity of theirionic radii to that of Mn4+ and/or other sites, such as inter-layer sites in phyllomanganates.

In contrast, structural incorporation of oxyanions isgenerally difficult because of their large molecular sizes(Takematsu et al., 1985, 1990). They are mainly adsorbedon ferromanganese oxides. The enrichment factors of oxya-nions relative to seawater are �107, which is relatively lowerthan those of the cations described above (Takematsu et al.,

1990; Hein et al., 2003). In this case, the concentrations ofoxyanions in ferromanganese oxides are inversely related tothe first or second dissociation constants of their acids, pKa(Li, 1981, 1991). On the other hand, structural incorpora-tion is possible for Te because of the geometric similarityof the Te(VI) molecule and the Fe octahedron. The Kd incoprecipitation experiment for the Te(VI)/ferrihydritesystem (�109) is significantly larger than those in adsorp-tion experiments (�107), which are themselves large dueto the large pKa of H6TeO6 (pKa1 = 7.70; McPhail,1995). Thus, we found persuasive evidence that the uniqueTe enrichment in ferromanganese oxides can take placethrough structural incorporation into Fe (oxyhydr)oxides,unlike other oxyanions that are mainly enriched viaadsorption.

Interestingly, Se is one of the least enriched oxyanions inferromanganese oxides although it is regarded as a geo-chemical twin of Te (Takematsu et al., 1990; Hein et al.,2003). In seawater, the redox speciations of Te and Se arevery similar, and the ratios of Te(VI)/Te(IV) and Se(VI)/Se(IV) in seawater are also quite similar (Measures et al.,1980; Lee and Edmond, 1985; Cutter and Cutter, 2001).On the other hand, the uptake of Se is controlled byadsorption as is the case for other oxyanions. Harada andTakahashi (2008) reported that the contrasting adsorptionsof Te and Se in the soil–water system are caused by thehighly soluble nature of Se(VI). This is due to the formationof an outer-sphere complex of Se(VI) on ferrihydritewhereas Te(IV), Te(VI), and Se(IV) form inner-sphere com-plexes (Harada and Takahashi, 2008; this study). There-fore, the great contrast in the distributions of Te and Seat the seawater/ferromanganese oxide interface is causedby their oxidized species: (i) the similarity in geometry ofthe Te(VI) molecule to the Fe octahedron, and (ii) quite sol-uble nature of Se(VI).

Previously, surface oxidation has been considered as avery efficient scavenging mechanism, and enrichments ofCo, Ce, and Tl are attributed to this process (Murray andDillard, 1979; Bidoglio et al., 1993; Takahashi et al.,2000; Peacock and Moon, 2012). Since the oxidized formsof Co(III), Ce(IV), and Tl(III) are sparingly soluble, theyare removed from the equilibrium reaction system after oxi-dation, leading to their progressive enrichment. This mech-anism has also been proposed for the extreme enrichmentof Te in natural ferromanganese oxides (Hein et al.,2003). However, oxidized Te(VI) and Se(VI) are still solu-ble, meaning that the oxidation does not necessarily pro-mote their enrichment through a change of solubility asthe case of cations above. Rather, it may facilitate theirfractionation on ferromanganese oxides through structuralincorporation of Te(VI) and the distribution of Se(VI) tothe aqueous phase.

The potential roles of Mn oxides on uniqueness of Teare also implied in some other cases. Previous laboratoryexperiments reported that some oxyanions, such as V(III)and Cr(III), are structurally incorporated into Fe (oxy-hydr)oxides because of their similarity in molecular geome-try to the Fe octahedron (e.g., Schwertmann and Pfab,1996; Sileo et al., 2004; Kaur et al., 2009). However, theirstructural incorporation in ferromanganese oxides is


unlikely, in part because their oxidation on coexisting Mnoxides could preclude incorporation through the changeof their geometry from Oh (for V(III) and Cr(III)) to Td

(for V(V) and Cr(VI)). This is the effect contrasting to thecase of Te oxidation where geometric change to Oh occursinstead. In addition, Mn oxides themselves are dominantphases attracting Mo and W to their surfaces by adsorp-tion. Larger enrichment of W than of Mo originates fromthe different stabilities of their surface complexes on Mnoxides (Kashiwabara et al., 2011, 2013). In the case of Te,redox transformation occurs on Mn oxides, causing achange in the Kd for Te adsorption on Mn oxides. TheKd of Te(VI) on d-MnO2 is the smallest among the adsorp-tion systems we examined (Fig. 5). Thus, the oxidation ofTe(IV) to Te(VI) on Mn oxides may provide “extra” Tefraction to coexisting Fe (oxyhydr)oxides. A similar processhas been suggested for As: As is finally adsorbed on ferrihy-drite after its oxidation on the surface of coexisting Mnoxides (Ying et al., 2012).

For many other oxyanions, positively-charged Fe (oxy-hydr)oxides are the most important phases, where variablechemistry can also occur to emphasize the uniqueness of Te.Some researchers have reported that As(V) and Sb(V) canbe incorporated as coprecipitated species in synthetic Fe(oxyhydr)oxides (e.g., Fuller et al., 1993; Violante et al.,2007; Mitsunobu et al., 2010; Tokoro et al., 2010). How-ever, As(V) is excluded from linkages of the Fe octahedraafter long aging time because of its Td geometry (Fulleret al., 1993; Waychunas et al., 1993). Thus, the slow growthrate of ferromanganese oxides would prefer adsorptionrather than incorporation for As(V), and further, for someother oxyanions with similar molecular geometries in natu-ral systems. As for Sb(V), although the Sb(V) molecule hasa similar geometry to that of the Fe octahedron, its distri-bution factor is not as large as that of Te. Consideringthe similar attachment geometries between adsorbed andcoprecipitated species, the low pKa of HSbO(OH)4

(pKa1 = 2.72; Smith and Martell, 1976) may be related toits smaller degree of enrichment in ferromanganese oxides.

Although further investigations are still necessary forsystematic understanding of variable chemistry at the sea-water/ferromanganese oxide interface, we argue that struc-tural incorporation of Te(VI) into Fe (oxyhydr)oxides canexplain the unique behavior of Te among oxyanions, whichmay also be emphasized by other factors including the co-existence of Mn oxides, the slow growth rate of ferroman-ganese oxides, and the large pKa of H6TeO6.

4.3. Difference from previous model and its relationship with

Te in seawater

The mechanism of extreme Te enrichment into ferroman-ganese oxides proposed in this study offers insights into therelationship between ferromanganese oxides and the abun-dance of Te species in seawater. The previous model forTe enrichment is based on an attempt to explain the discrep-ancy between the greater thermodynamic stability of Te(IV)over Te(VI) species (McPhail, 1995) and natural observa-tions of the lack of Te(IV) species in seawater (Lee andEdmond, 1985). In that model, marine ferromanganese

oxides cause the lack of Te(IV) species in seawater (Heinet al., 2003). That is, the preferential removal of Te(IV) byferromanganese oxides can make Te(VI) predominant ifthe redox reaction between Te(IV) and Te(VI) species is veryslow in seawater (McPhail, 1995).

However, our coprecipitation model suggests that ferro-manganese oxides do not affect the relative abundances ofTe species in seawater. In our model, the main pathway isstructural incorporation of Te(VI) into Fe (oxyhydr)oxidephases without redox transformation. In addition, we arguethat the reaction between aqueous Te(IV) and Te(VI) spe-cies should reach equilibrium considering the geologicaltimescale for the slow growth rate of hydrogenetic ferro-manganese crust (�6 mm/Ma). Although some compila-tions of thermodynamic data (e.g., Mishra et al., 1990;McPhail, 1995) are often cited in support of the idea thatequilibrium between Te(IV) and Te(VI) is not achieved innatural seawater, we should note that they are based onlaboratory experiments in much shorter timescales. Ther-modynamic stability of Te(VI) species has not yet beenproperly evaluated because of a lack of experimental datasets (McPhail, 1995). Actually, other experimental and the-oretical studies have reported that Te(VI) can be predomi-nant in oxic solutions such as seawater (Brookins, 1988;Nolan et al., 1991; Harada and Takahashi, 2008).

The validity of our model is supported by observationsof Te concentration in natural ferromanganese oxides. Ifthe incorporation of Te is dominated by adsorption, thenTe concentration in ferromanganese oxides is mainlydependent on its concentration in seawater. However, meanTe concentrations significantly vary from 0.06 to 205 ppmrelative to the small variations in seawater from various re-gions of the global ocean (0.4 – 1.9 pM; Lee and Edmond,1985; Hein et al., 2003). In addition, Te concentrations innatural ferromanganese oxides are well correlated withtheir growth rates (Hein et al., 2003). These facts imply thatincorporation of Te is a dynamic process related to Fe andMn chemistry, including coprecipitation rather thanadsorption–desorption equilibrium between seawater andsolid surfaces. We conclude that Te enrichment model viacoprecipitation is chemically consistent with the trend inTe concentrations observed in a number of natural ferro-manganese oxides.

5. CONCLUSION

We investigated the chemical factors for the extremeenrichment of Te into marine ferromanganese oxides usinga combination of XAFS analysis and adsorption/coprecip-itation experiments. Our findings are as follows:

(i) The oxidation state of Te species in natural ferroman-ganese oxides is hexavalent in spite of the coexistence ofTe(IV) and Te(VI) in seawater. The host phase of Te in nat-ural ferromanganese oxides is ferrihydrite, and it is mainlya coprecipitated species.

(ii) Tellurium(IV) is oxidized on d-MnO2 during adsorp-tion and coprecipitation reactions, whereas it is not oxi-dized on ferrihydrite.

(iii) Tellurium(IV) and Te(VI) are adsorbed on thesurface of both ferrihydrite and d-MnO2 via formation of


inner-sphere complexes. Further, only Te(VI) species can bestructurally incorporated into ferrihydrite throughcoprecipitation.

(iv) Although adsorption of Te(IV) species on ferrihy-drite has a larger distribution coefficient than other adsorp-tion systems, coprecipitation of Te(VI) species intoferrihydrite has the largest one of all the experimental sys-tems. The Kd for Te(VI) coprecipitation can explain thenatural distribution between seawater and ferromanganeseoxides.

While other oxyanions are mainly incorporated viaadsorption on the surface of solid structures, Te(VI) canbe structurally incorporated into Fe (oxyhydr)oxide phasesdue to the similarity of its molecular geometry to that of theFe(III) octahedron. This process explains the uniqueenrichment of Te compared with other oxyanions, and itis also responsible for the great difference between Te andSe, which is highly soluble in its oxidized Se(VI) form.Coexisting Mn oxide phases may also promote structuralincorporation of Te(VI) by oxidation of Te(IV), althoughsurface oxidation itself may not work as the critical enrich-ment process for Te, as is the case for some cations such asCo, Ce, and Tl. This mechanism also means that ferroman-ganese oxides mainly scavenge Te(VI) species from seawa-ter without redox transformation and do not affect thespecies distribution of Te in seawater, as previously pro-posed by Hein et al. (2003). The variation in abundanceof Te and the correlation of Te concentration with thegrowth rate of natural ferromanganese oxides are consis-tent with an enrichment mechanism dominated bycoprecipitation.

ACKNOWLEDGMENTS

We thank Dr. Yoko Shimamoto (Hiroshima University) andDr. Satoshi Mitsunobu (Shizuoka-kenritsu University) for theircontribution to this work. This study was supported by grants-in-aid for scientific research on Innovative Areas including theTrans-crustal Advection and In situ reaction of Global sub-seafloorAquifer (TAIGA) project from the Ministry of Education, Culture,Sports, Science and Technology of Japan (MEXT), a ResearchActivity Start-up grant from the Japan Society for the Promotionof Science (JSPS KAKENHI Grant no. 23840058), and the Sasak-awa Scientific Research Grant from the Japan Science Society. Thiswork was conducted with approvals from JASRI (Proposal Nos.2010A1612, 2010B1664, 2011B1400, 2012A1767, and 2012B1672)and KEK (Proposal Nos. 2009G585 and 2011G635).

REFERENCES

Baturin G. N. (2007) Tellurium and thallium in ferromanganesecrusts and phosphates on oceanic seamounts. Dokl. Earth Sci.

413, 331–335.Baturin G. N., Dubinchuk V. T., Azarnova L. A. and Mel’nikov

M. E. (2007) Species of molybdenum, thallium, and tellurium inferromanganese crusts of oceanic seamounts. Dokl. Earth Sci.

47, 415–422.Bidoglio G., Gibson P. N., O’Gorman M. and Roberts K. J. (1993)

X-ray absorption spectroscopy investigation of surface redoxtrans-1126 J. R. Hein, A. Koschinsky, and A. N. Hallidayformations of thallium and chromium on colloidal mineraloxides. Geochim. Cosmochim. Acta 57, 2389–2394.

Brookins D. G. (1988) Eh–pH Diagrams for Geochemistry.Springer-Verlag, Berlin.

Brown R. D. (2000) Selenium, tellurium, and bismuth: supply,demand, and progress. In Minor Elements 2000: Processing andEnvironmental Aspects of As, Sb, Se, Te, and Bi (ed. C.Young). Society for Mining, Metallurgy and Exploration. pp.307–321.

Byrne R. H. (2002) Inorganic speciation of dissolved elements inseawater: the influence of pH on concentration ratios. Geochem.

Trans. 3, 11–16.Carvalho-E-Silva M. L., Ramos A. Y., Tolentino H. C. N.,

Enzweiler J., Netto S. M. and Alves M. D. C. M. (2003)Incorporation of Ni into natural goethite: an investigation byX-ray absorption spectroscopy. Am. Mineral. 88, 876–882.

Crawford R. J., Harding I. H. and Mainwaring D. E. (1993a)Adsorption and coprecipitation of single heavy metal ions ontothe hydrated oxides of iron and chromium. Langmuir 9, 3050–3056.

Crawford R. J., Harding I. H. and Mainwaring D. E. (1993b)Adsorption and coprecipitation of multiple heavy metal ionsonto the hydrated oxides of iron and chromium. Langmuir 9,3057–3062.

Cronan D. S. (1980) Underwater Minerals. Academic Press,London, p. 362.

Cutter G. A. and Cutter L. S. (2001) Sources and cycling ofselenium in the western and equatorial Atlantic Ocean. Deep-

Sea Res. II 48, 2917–2931.Deng Z. T., Zhang Y., Yue J. C., Tang F. Q. and Wei Q. (2007)

Green and orange CdTe quantum dots as effective pH-sensitive fluorescent probes for dual simultaneous and inde-pendent detection of viruses. J. Phys. Chem. B 111, 12024–12031.

Foster A. L., Brown G. E. and Parks G. A. (2003) X-rayabsorption fine structure study of As(V) and Se(IV) sorptioncomplexes on hydrous Mn oxides. Geochim. Cosmochim. Acta

67, 1937–1953.Fuller C. C., Davis J. A. and Waychunas G. A. (1993) Surface

chemistry of ferrihydrite: part 2. Kinetics of arsenate adsorp-tion and coprecipitation. Geochim. Cosmochim. Acta 57, 2271–2282.

Graf C., Assoud A., Mayasree O. and Kleinke H. (2009) Solid statepolyselenides and polytellurides: a large variety of Se–Se andTe–Te interactions. Molecules 14, 3115–3131.

Giovanoli R. and Cornell R. M. (1992) Crystallization of metalsubstituted ferrihydrites. Pflanzenernaehr. Bodenkd. 155, 455–460.

Halbach P., Sattler C., Teichmann F. and Wahsner M. (1989)Cobalt-rich and platinum-bearing manganese crust deposits onseamounts: nature, formation, and metal potential. Marine

Mining 8, 23–39.Hansel C. M., Wielinga B. W. and Fendorf S. R. (2003) Structural

and compositional evolution of Cr/Fe solids after indirectchromate reduction by dissimilatory iron-reducing bacteria.Geochim. Cosmochim. Acta 67, 401–412.

Harada T. and Takahashi Y. (2008) Origin of the difference in thedistribution behavior of tellurium and selenium in a soil–watersystem. Geochim. Cosmochim. Acta 72, 1281–1294.

Hein J. R., Koschinsky A., Halbach P., Manheim F. T., Bau M.,Kang J.-K. and Roberts L. (2000) Cobalt-rich ferromanganesecrusts in the Pacific. In Handbook of Marine Mineral Deposits

(ed. D. S. Cronan). CRC Press, Boca Raton, Florida, pp. 239–279.

Hein J. R., Koschinsky A. and Halliday A. N. (2003) Globaloccurrence of tellurium-rich ferromanganese crusts and a modelfor the enrichment of tellurium. Geochim. Cosmochim. Acta 67,1117–1127.

http://refhub.elsevier.com/S0016-7037(14)00064-7/h0005










































































Hein J. R., Conrad T. A. and Staudigel H. (2010) Seamountmineral deposits a source of rare metals for high-technologyindustries. Oceanography 23, 184–189.

Ikeda H., Fujino S. and Kajiwara T. (2009) Preparation andcharacterization of BaO–TeO2 thin films obtained from tellu-rium(VI) alkoxide by a Sol-Gel method. J. Am. Ceram. Soc. 92,2619–2622.

Jiang X., Lin X., Zhai S. and Guo W. (2007) Geochemistry oflithium in marine ferromanganese oxide deposits. Deep Sea Res.

Part 1 54, 85–98.Kashiwabara T., Mitsunobu S., Das A., Itai T., Tanimizu M. and

Takahashi Y. (2008) Oxidation state of antimony and arsenic inmarine ferromanganese oxides related to their fractionation inoxic marine environment. Chem. Lett. 37, 756–757.

Kashiwabara T., Takahashi Y. and Tanimizu M. (2009) A XAFSstudy on the mechanism of isotopic fractionation of molybde-num during its adsorption on ferromanganese oxides. Geochem.

J. 43, e31–e36.Kashiwabara T., Takahashi Y., Uruga T., Tanida H. and Terada

Y. (2010) Speciation of tungsten in natural ferromanganeseoxides using wavelength dispersive XAFS. Chem. Lett. 39, 870–871.

Kashiwabara T., Takahashi Y., Tanimizu M. and Usui A. (2011)Molecular-scale mechanisms of distribution and isotopic frac-tionation of molybdenum between seawater and ferromanga-nese oxides. Geochim. Cosmochim. Acta 75, 5762–5784.

Kashiwabara T., Takahashi Y., Marcus M. A., Uruga T., TanidaH., Terada Y. and Usui A. (2013) Tungsten species in naturalferromanganese oxides related to its different behavior frommolybdenum in oxic ocean. Geochim. Cosmochim. Acta 106,364–378.

Karthikeyan K. G., Elliott H. A. and Cannon F. S. (1997)Adsorption and coprecipitation of copper with the hydrousoxides of iron and aluminium. Environ. Sci. Technol. 31, 2721–2725.

Kaur N., Singh B., Kennedy B. J. and Graefe M. (2009) Thepreparation and characterization of vanadium-substituted goe-thite: the importance of temperature. Geochim. Cosmochim.

Acta 73, 582–593.Klas S., Dubowski Y., Pritosiwi G., Calmano W. and Lahav O.

(2011) Extent and mechanism of metal ion incorporation intoprecipitated ferrites. J. Colloid Incerf. Sci. 358, 129–135.

Koschinsky A. and Hein J. R. (2003) Uptake of elements fromseawater by ferromanganese crusts: solid-phase associationsand seawater speciation. Mar. Geol. 198, 331–351.

Lakin H. W., Thompson C. E. and Davidson D. F. (1963)Tellurium content of marine manganese oxides and othermanganese oxides. Science 142, 1568–1569.

Lee D. S. and Edmond J. M. (1985) Tellurium species in seawater.Nature 313, 782–785.

Li Y. H. (1981) Ultimate removal mechanisms of elements from theocean. Geochim. Cosmochim. Acta 45, 1659–1664.

Li Y. H. (1991) Distribution patterns of the elements in the ocean:a synthesis. Geochim. Cosmochim. Acta 55, 3223–3240.

Li Y. H., Wang Y. M., Song H. B. and Yue G. L. (2005) Extremeenrichment of tellurium in deep-sea sediments. Acta Geol.

Sinica 79, 547–551.Liu H., Chen T., Frost R. L., Chang D., Qing C. and Xie Q. (2012)

Effect of aging time and Al substitution on the morphology ofaluminous goethite. J. Colloid Interface Sci. 385, 81–86.

McPhail D. C. (1995) Thermodynamic properties of aqueoustellurium species between 25 and 350�C. Geochim. Cosmochim.

Acta 59, 851–866.Manceau A., Llorca S. and Calas G. (1987) Crystal chemistry of

cobalt and nickel in lithiophorite and asbolane from NewCaledonia. Geochim. Cosmochim. Acta 51, 105–113.

Manceau A. and Drits V. A. (1993) Local structure of ferrihydriteand feroxyhite by EXAFS spectroscopy. Clay Miner. 28, 165–184.

Manceau A., Drits V. A., Silverster E., Bartoli C. and Lanson B.(1997) Structural mechanism of CO2+ oxidation by the phyllo-manganate buserite. Am. Mineral. 82, 1150–1175.

Manceau A., Schlegel M. L., Musso M., Sole V. A., Gauthier C.,Petit P. E. and Trolard F. (2000) Crystal chemistry of traceelements in natural and synthetic goethite. Geochim. Cosmo-

chim. Acta 64, 3643–3661.Martinez C. E. and McBride M. B. (2001) Cd, Cu, Pb, and Zn

coprecipitates in Fe oxides formed at different pH: aging effectson metal solubility and extractability by citrate. Environ. Chem.

20, 122–126.Martin F., Ildefonse P., Hazeamnn J. L., Mathe P. E., Noack Y.,

Grauby O., Beziat D. and Parseval P. (1997) Gallium crystalchemistry in synthetic goethites. J. Phys. IV France 7.

Measures C. I., McDuff R. E. and Edmond J. M. (1980) Seleniumredox chemistry at GEOSEC I re-occupation. Earth Planet. Sci.

Lett. 49, 102–109.Mishra K. K., Ham D. and Rajeshwar K. (1990) Anodic oxidation

of telluride ions in aqueous base: a rotating ring-disk electrodestudy. J. Electrochem. Sot. 137, 3438–3441.

Mitsunobu S., Takahashi Y., Terada Y. and Sakata M. (2010)Antimony (V) incorporation into synthetic ferrihydrite, goe-thite, and natural iron oxyhydroxides. Environ. Sci. Technol. 44,3712–3718.

Murray J. W. and Dillard J. G. (1979) The oxidation of cobalt(II)adsorbed on manganese dioxide. Geochim. Cosmochim. Acta 43,781–787.

Nolan C., Whitehead N. and Teyssie J.-L. (1991) Tellurium –speciation in seawater and accumulation by marine phyto-plankton and crustaceans. J. Environ. Radioact. 13, 217–233.

Peacock C. and Moon E. M. (2012) Oxidative scavenging ofthallium by birnessite: explanation for thallium enrichment andstable isotope fractionation in marine ferromanganese precip-itates. Geochim. Cosmochim. Acta 84, 297–313.

Puteanus D. and Halbach P. (1988) Correlation of Co concentra-tion and growth rate – a method for age determination offerromanganese crusts. Chem. Geol. 69, 73–85.

Schwertmann U. and Cornell R. M. (2000) Iron Oxides in the

Laboratory, 2nd ed. Wiley-VCH, p. 103.Schwertmann U. and Pfab G. (1996) Structural vanadium and

chromium in lateritic iron oxides: genetic implications. Geo-

chim. Cosmochim. Acta 60, 4279–4283.Sileo E. E., Ramos A. Y., Magaz G. E. and Blesa M. A. (2004)

Long-range vs. short-range ordering in synthetic Cr-substitutedgoethite. Geochim. Cosmochim. Acta 68, 3053–3063.

Takahashi Y., Shimizu H., Usui A., Kagi H. and Nomura M.(2000) Direct observation of tetravalent cerium in ferromanga-nese nodules and crust by X-ray-absorption near-edge structure(XANES). Geochim. Cosmochim. Acta 64, 2929–2935.

Takahashi Y., Manceau A., Geoffroy N., Marcus M. A. and UsuiA. (2007) Chemical and structural control of the partitioning ofCo, Ce, and Pb in marine ferromanganese oxides. Geochim.

Cosmochim. Acta 71, 984–1008.Takematsu N., Sato Y., Okabe S. and Nakayama E. (1985) The

partition of vanadium and molybdenum between manganeseoxides and seawater. Geochim. Cosmochim. Acta 49, 2395–2399.

Takematsu N., Sato Y., Okabe S. and Usui A. (1990) Uptake ofselenium and other oxyanionic elements in marine ferroman-ganese concretions of different origins. Mar. Chem. 31,271–283.

Tokoro C., Yatsugi Y., Koga H. and Owada S. (2010) Sorptionmechanisms of arsenate during coprecipitation with ferrihydritein aqueous solution. Environ. Sci. Technol. 44, 638–643.








































































































































Usui A. (1995) Studies of marine manganese deposits: review andperspectives. Chishitsu News 493, 30–41 (in Japanese).

Violante A., Gaudio S. D., Pigna M., Ricciardella M. and BanerjeeD. (2007) Coprecipitation of arsenate with metal oxides. 2.nature, mineralogy, and reactivity of iron(III) precipitates.Environ. Sci. Technol. 41, 8275–8280.

Wachter J. (2004) Metal telluride clusters – from small molecules topolyhedral structures. Eur. J. Inorg. Chem., 1367–1378.

Waychunas G. A., Rea B. A., Fuller C. C. and Davis J. A. (1993)Surface-chemistry of ferrihydrite: part 1. EXAFS studies of thegeometry of coprecipitated and adsorbed arsenate. Geochim.

Cosmochim. Acta 57, 2251–2269.Wells M. A., Fitzpatrick R. W. and Gilkes R. J. (2006) Thermal

and mineral properties of Al-, Cr-, Mn-, Ni- and Ti-substitutedgoethite. Clays Clay Miner. 54, 176–194.

Ying S. C., Kocar B. D. and Fendorf S. (2012) Oxidation andcompetitive retention of arsenic between iron- and manganeseoxides. Geochim. Cosmochim. Acta 96, 294–303.

Zabinsky S. I., Rehr J. J., Ankudinov A., Albers R. C. and Eller M.J. (1995) Multiple-scattering calculations of X-ray-absorptionspectra. Phys. Rev. B 52, 2995–3009.

Zhang L., Wang C. and Wen D. Y. (2009) Preparation byhydrothermal techniques in a tungstosilicate acid solutionsystem and optical properties of tellurium nanotubes. Eur. J.

Inorg. Chem., 3291–3297.

Associate editor: Karen Johannesson


























chemical processes for the extreme enrichment of...

Documents