groer test - dynamic-eye

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

An atomic level study of rhenium and radiogenicosmium in molybdenite

Yoshio Takahashi a,b,*, Tomoya Uruga c, Katsuhiko Suzuki d, Hajime Tanida c,Yasuko Terada c, Keiko H. Hattori e

a Department of Earth and Planetary Systems Science, Graduate School of Science, Hiroshima University,

Higashi-Hiroshima, Hiroshima 739-8526, Japanb Laboratory for Multiple Isotope Research for Astro- and Geochemical Evolution (MIRAGE), Hiroshima University,

Hiroshima 739-8526, Japanc SPring-8, Japan Synchrotron Radiation Research Institute (JASRI), Sayo-cho, Hyogo 679-5198, Japan

d Institute for Research on Earth Evolution (IFREE), Japan Agency for Marine-Earth Science and Technology (JAMSTEC),

Yokosuka 237-0061, Japane Department of Earth Sciences, University of Ottawa, Ottawa, Ont., Canada K1N 6N5

Received 28 December 2006; accepted in revised form 3 August 2007; available online 7 September 2007

Abstract

Local atomic structures of Re and radiogenic Os in molybdenite from the Onganja mine, Namibia, were examined usingX-ray absorption fine structure (XAFS). Rhenium LIII-edge X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) show that the oxidation state of Re, the interatomic distances between Re and theneighboring atoms, and the coordination number of Re to S are very similar to those of Mo in molybdenite. The results con-firm that Re is present as Re(IV) in the Mo site in molybdenite.

Measurement of LIII-edge XANES and EXAFS of a minor concentration (8.55 ppm) of radiogenic Os was accomplished influorescence mode by removing the interfering X-rays from Re and other elements using a crystal analyzer system. The dataindicate that the oxidation state of radiogenic Os is Os(III) and Os(IV) and clearly different from Os(II) in natural sulfideminerals, such as OsS2 (erlichmanite). XANES data also suggest that radiogenic Os does not form a secondary Os phase, suchas OsS2 or Os metal, in molybdenite.

EXAFS of radiogenic Os was successfully simulated assuming that Os is present in the Mo site in molybdenite. The dataare consistent with the XANES data; Os does not form Os phases in molybdenite. The EXAFS simulation showed that theinteratomic distance between Os and S is 2.27 A, which is 0.12 A smaller than the distances of Re–S and Mo–S (2.39 A) inmolybdenite. Similar valences and ionic sizes of Re and Mo in molybdenite support the fact that large amounts of Re canbe incorporated into the Mo site as has been observed in previous studies, whereas the different properties of Os comparedto Mo and Re suggested here support much lower abundance of common Os in molybdenite. This makes molybdenite an idealmineral for the Re–Os geochronometer as shown in many studies. However, the shorter distance between radiogenic Os and Scompared to those of Re–S and Mo–S in molybdenite suggests that the radiogenic Os has a smaller ionic size than Re(IV) andMo(IV). Furthermore, Os may be partly present as Os(III). Smaller and lower charge Os can diffuse faster than larger andhigher charge Re in molybdenite at a given set of conditions. Hence, our study provides an atomic-level explanation forthe high mobility of Os compared to Re, which has been suggested by earlier workers using laser ablation ICP-MS.� 2007 Published by Elsevier Ltd.

0016-7037/$ - see front matter � 2007 Published by Elsevier Ltd.

doi:10.1016/j.gca.2007.08.007

* Corresponding author. Fax: +81 82 424 0735.E-mail address: [email protected] (Y. Takahashi).

www.elsevier.com/locate/gca

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 71 (2007) 5180–5190


1. INTRODUCTION

Radiometric dating has been widely applied to variousterrestrial and extraterrestrial materials (Faure and Mensing,2004). Such dating assumes the retention of both parent anddaughter isotopes in the samples since formation. The chem-ical properties of the daughter commonly differ from those ofthe parent, which may result in their separation. Thus, theirretention in the system may depend on the chemical environ-ment of both elements, which determines their stabilitiesin situations of various processes such as metamorphism,alteration, and diffusion within the crystal. Therefore, thechemical states of parent and daughter isotopes in the min-eral or rock samples may be important for precise dating.

Since the first application of the 187Re–187Os system fordating of molybdenite by Herr and Merz (1955), the datingof molybdenite has been established as a robust geochronom-eter resistant to regional deformation and metamorphism(e.g., Stein et al., 1997, 2001, 2003; Raith and Stein, 2000;Watanabe and Stein, 2000; Selby and Creaser, 2001a,b,2004). In particular, the method has become an important toolin dating the timing and duration of sulfide mineralization inore-forming systems (e.g., Suzuki et al., 1996; Stein et al., 1997,1998). However, decoupling of 187Os and Re possibly due tohigh mobility of Os relative to Re has been observed withinmolybdenite crystals by micrometer scale analysis using the la-ser ablation ICP-MS technique (Stein et al., 2001, 2003; Kos-ler et al., 2003; Selby and Creaser, 2004). Hence, it has beenindicted that successful dating requires careful selection of rep-resentative samples from a specific geologic occurrence with asufficient volume of the sample for the analysis (Stein et al.,2001, 2003; Selby and Creaser, 2004; Stein, 2006). In thisstudy, we examine the local structures of Re and radiogenic187Os in molybdenite using their X-ray absorption fine struc-ture (XAFS) to evaluate the atomic scale environment forthe two elements and the cause of the decoupling. Since theabundance of common Os is negligible in molybdenite (Herrand Merz, 1955; Hirt et al., 1963; Morgan et al., 1968; Steinet al., 2001), the XAFS data of Os in molybdenite providesthe information on the decay product, 187Os.

For our study, fluorescence XAFS spectra of Re and Oswere employed because they have low detection limits andhigh selectivity of any trace elements in a variety of samples(e.g., Takahashi et al., 2000, 2004). XAFS consists of two en-ergy regions, X-ray absorption near-edge structure(XANES) and extended X-ray absorption fine structure(EXAFS). We used both sets of data to evaluate the oxida-tion state and local atomic structure of Re and radiogenicOs. Retrieving the data from trace Os was an analytical chal-lenge because its spectrum was obscured by the intense fluo-rescence X-ray emitted from Re and other scattering X-rays.This was overcome by employing a crystal analyzer in Lauegeometry to extract Os La emission (Takahashi et al., 2006).

2. SAMPLES AND METHODS

2.1. Samples and materials

Natural molybdenite (MoS2; polytype: 2H) wasobtained from the Onganja mine in the Damara Facies,

70 km northeast of Windhoek, in central Namibia. Themolybdenite was collected from a post-metamorphicquartz–biotite vein in the schist of the Late ProterozoicKuiseb Formation. The concentrations of Re and Os inthe sample measured by inductively coupled plasma-massspectrometer (ICP-MS) were 1610 ± 10 and 8.55 ±0.04 ppm, respectively, following the procedures reportedin Suzuki et al. (1992). The sample was chosen for thisstudy due to its high concentrations of Re and Os; thegreater the concentrations, the more clearly is the XAFSspectrum. The isotopic composition of Os in the Onganjamolybdenite shows that 99.2% of Os was 187Os (Schiederet al., 1981), implying that more than 99% of 187Os is a de-cay product of 187Re.

Reference materials used for Os LIII-edge XAFS wereOs metal, OsIIS2, OsIIICl3, and OsVIIIO4 as reported bySakakibara et al. (2005). Osmium metal and OsCl3 werepurchased from Rare Metallic Co. Ltd., whereas OsO4 isfrom Kanto Chemical Co. Ltd. Osmium disulfide (OsS2,erlichmanite) was synthesized by following the procedurereported in Hayashi et al. (2000). Reference materials forRe were Re metal, ReIVS2, ReIVO2, and ReVIO3 obtainedfrom Rare Metallic Co. Ltd. and ReO4

� solution fromKanto Chemical Co. Ltd. All purchased chemicals wereat reagent grades used without further purification for theXAFS measurements.

2.2. XAFS measurements

Rhenium LIII-edge XAFS was measured at BL-12C(bending magnet) of the Photon Factory (PF) at High En-ergy Accelerator Research Organization (KEK), Tsukuba,Japan (Nomura and Koyama, 1996). In the beamline,X-rays from a synchrotron operated at 2.5 GeV (current:450–300 mA) were monochromatized with a Si(111) dou-ble-crystal monochromator, and focused to an area of 1(horizontal) · 0.5 (vertical) mm2 with a bent cylindricalmirror, which also reduces the higher order. XAFS spectrawere obtained in transmission mode for reference materials,and in fluorescence mode for Re in molybdenite. The fluo-rescence yield was measured using a 19-element Ge semi-conductor detector (SSD). The total counts weremaintained at less than 200 kcps by adjusting the distancebetween the SSD and the sample. Counts for high intensi-ties of fluorescence and scattering X-rays are corrected fordead counting time (Nomura, 1998).

Osmium LIII-edge XAFS of Os in molybdenite was mea-sured in the fluorescence mode using a crystal analyzer atan undulator beamline, BL37XU, at SPring-8, Japan(Terada et al., 2004). The beamline consists of Si(111) dou-ble-crystal monochromator and Rh-coated mirror for sup-pression of higher harmonics. The beam size and themonochromatic photon flux at the sample position were0.2 (horizontal) · 0.8 (vertical) mm2 and 4.4 · 1012 pho-tons/s, respectively. The sample was placed at 45� to theincident X-ray beam. The extraction of Os La fluorescenceused a commercially available fluorescence analyzer(BCLA: Bent Crystal Laue Analyzer) made by OxfordDanfysik (Zhong et al., 1999). The analyzer, DCA 0950,used here was specially designed to determine the intensity

Local atomic structures of Re and radiogenic Os in molybdenite 5181


of fluorescence X-rays ranging from 8.5 to 10.5 keV, whichcovers Os La fluorescence X-rays. The analyzer consists of alog spiral Si(111) crystal and slits made of Mo metal. A 19-element Ge semiconductor detector (Ortec IGLET-X-11160) was used for detection. The detector was wellshielded by a Pb sheet to collect X-rays only from the crys-tal analyzer. XAFS spectra of Os in reference materialswere measured in transmission mode.

2.3. XAFS analysis

EXAFS data were analyzed using REX2000 ver.2.3(Rigaku Co.) with parameters generated by the computerprogram FEFF7.02 (an automated program for ab initio

multiple scattering calculations of XAFS; Zavinsky et al.,1995; Ankudinov and Rehr, 1997). Initial structural datafor the FEFF calculations were provided by ATOMS(Ravel, 2001) using crystallographic data listed in Table 1.For the simulation of Re in molybdenite, initial structurefor the FEFF calculation assumed that Re replaces Mo inthe molybdenite structure reported in Bronsema et al.(1986). The same procedure was also applied to Os inmolybdenite, since b-decay without emission of c-rays, suchas the decay of 187Re to 187Os (energy of b-ray: 2.6 keV;Parrington et al., 1996), is not accompanied by a large aver-age recoil that can rupture the bonds of solid matrices (e.g.,Maddock et al., 1992). Therefore, Os most likely remains inthe same site as Re in molybdenite after the decay. Thevalidity of this assumption for the EXAFS simulation isfurther discussed in Section 4. For the analysis of ReS2

and OsS2 spectra, the initial structure data to run FEFFwere obtained from their crystallographic data in Murrayet al. (1994) and Stingl et al. (1992), respectively.

After background subtraction from the observed spec-tra and normalization, the smooth Re or Os LIII-edgeabsorption of the free atom was removed using a cubicspline curve. The energy unit was transformed from keVto A�1 to produce the EXAFS function v(k), where k

(A�1) is the photoelectron wave vector given byffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2mðE � E0=h2Þ

q(E: the energy of the incident X-ray; E0:

the threshold energy). The E0 value was determined fromthe maximum value in the first derivative of v(k) in theabsorption edge region. The k3-weighted v(k) functionwas Fourier transformed from k (1/A) space into R (A)space to give a radial structural function (RSF). The the-oretical EXAFS function was fitted to the back trans-formed k3-weighted v(k) functions using parametersgenerated by FEFF 7.02. During the simulation, it wasfound that multiple scattering is not important in all sys-tems examined in this study. The EXAFS analysis givesthe coordination number (CN), the interatomic distancebetween absorber and scatterer atoms (R), the Debye–

Waller term (r2), and the energy offset (DE0). The S20

(dumping factor) was fixed at unity in our simulation.The maximum number of parameters for EXAFS fittings

was restricted below Nfree defined as N free ¼ 2Dk�Drp þ 2

(Stern, 1993), where Dk and Dr are k region (1/A) of Fou-rier transformation and R region (A) of back Fouriertransformation, respectively.

3. RESULTS

3.1. Characterization of the Onganja molybdenite

The age of the Onganja molybdenite is calculated to be505 ± 6 Ma, assuming that all Os is a decay product of187Re. The uncertainty in the age includes that of the decayconstant, k = (1.666 ± 0.017) · 10�11 years�1 (Smoliaret al., 1996). This is in accordance with the Re–Os age of510 ± 40 Ma for molybdenite in the Onganja Mine re-ported by Schieder et al. (1981) (recalculated usingk = 1.666 · 10�11 years�1), and also with the K–Ar age of480 ± 25 Ma for a biotite in the molybdenite-bearing vein(Clifford, 1967). Similar Re–Os and K–Ar ages suggest thatthe Re–Os and K–Ar isotope systems were not disturbed bylater hydrothermal or magmatic activity.

3.2. Rhenium LIII-edge XAFS of Onganja molybdenite

Rhenium LIII-edge XANES spectra of reference materi-als and Re in the molybdenite are shown in Fig. 1. Themolybdenite sample shows a peak of XANES close to thoseof ReIVS2 and Re metal. Its shoulder at about 10.534 keV issimilar to that of ReS2, but dissimilar to that of Re metal.In addition, the intensity of the ‘‘white line’’ (= main peakin the XANES spectrum) of Re in molybdenite is similar tothat of ReS2. These results suggest that the local atomicenvironment of Re in molybdenite is similar to that ofReS2 and that the oxidation state of Re in molybdenite isRe(IV), as suggested in previous studies (Fleischer, 1959;Newberry, 1979; Stein et al., 2001). The results confirm thatMo(IV) in molybdenite can be readily substituted for byRe(IV).

The EXAFS data suggest that local structures of Re inmolybdenite and ReS2 are similar to that of Mo in molyb-denite (Figs. 2 and 3). The k3-weighted Re LIII-edge EX-AFS spectra v(k) and the radial structural function (RSF)obtained by Fourier transforms from k3-weighted v(k) weresimulated by the parameters generated by FEFF 7.02. Thestructure for the FEFF calculation was created based onthe molybdenite structure, assuming that Re is in the Mosite. In the EXAFS simulation of Re in molybdenite, thecoordination numbers (CN) of S and Mo were fixed at 6based on the crystallographic information. In the k3 v(k)pattern, the EXAFS oscillation for Re in the molybdenitesample is similar to Re in ReS2 at k below 6 A�1, since asimilar Re–S shell is mainly responsible for the oscillationin the k region (Fig. 2a). However, the oscillation patternsfor Re in molybdenite and ReS2 become different in the lar-ger k region, suggesting that the second neighboring atomsfor the two are different. The parameters obtained by thesimulation are tabulated in Table 1. In the molybdenitesample, the distance between Re and S is 2.389 A (Table1), which is similar to the distance, 2.401 A, between Moand S in molybdenite (polytype 2H), based on crystallo-graphic studies (Table 1). The interatomic distance of thesecond shell, Re–Mo, in the molybdenite sample was calcu-lated to be 3.157 A, which is also very close to the Mo–Modistance (3.154 A) in molybdenite in the literature (Hassel,1925; Kalikhman, 1983; Schoenfeld et al., 1983; Bronsema

5182 Y. Takahashi et al. / Geochimica et Cosmochimica Acta 71 (2007) 5180–5190


Table 1EXAFS analyses results obtained by the simulation using parameters generated by FEFF 7.02

Sample Edge Analyzed shells Shell CN R (A) DE0 (eV) r2 (· 103) Nfree Literature values

Compounds Shell CN R (A)

ReS2 Re LIII 2 shells Re–S 5.9 ± 0.3 2.388 ± 0.003 6.0 ± 0.5 6.1 ± 0.04 12 ReS2 Re–S 6 2.312–2.504c

Re–Re 3.0a 2.777 ± 0.007 — 7.1 ± 0.01 Re–Re 3 2.790–2.896d

Molybdenite Re LIII 2 shells Re–S 6.0a 2.389 ± 0.002 6.8 ± 0.4 0.44 ± 0.04 13 MoS2 Mo–S 6 2.401 ± 0.026e

Re–Mo 6.0a 3.157 ± 0.018 — 5.3 ± 0.01 Mo–Mo 6 3.154 ± 0.010e

OsS2 Os LIII 3 shells Os–S 5.7 ± 0.1 2.353 ± 0.001 8.6 ± 0.3 0.63 ± 0.04 18 OsS2 Os–S 6 2.352 ± 0.001f

Os–S 6.0a 3.556 ± 0.007 — 2.8 ± 0.04 Os–S 6 3.565 ± 0.001f

Os–Os 12.0a 3.974 ± 0.001 — 2.5 ± 0.12 Os–Os 12 3.974f

Molybdenite Os LIII 1 shellb Os–S 6.0a 2.271 ± 0.022 �2.9 ± 3.5 0.40 ± 5.3 6Molybdenite Os LIII 2 shellsb Os–S 6.0a 2.272 ± 0.021 �5.1 ± 0.2 0.23 ± 8.3 8

Os–Mo 6.0a 3.18 ± 0.41 — 1.2 ± 0.01

Corresponding literature values were also listed. CN, coordination number; R, interatomic distance; DE0, threshold E0 shift; r, Debye–Waller term; Nfree, maximum number of parameters.Least squares precisions are given to each value.* Errors in the fitted parameters were estimated to be generally ±0.02 A for R, ±20% for N, and 20% for r2 (O’Day et al., 1994).

a Fixed in the simulation.b For Os in the molybdenite sample, two simulations were performed; one considers only the first shell and the other the first and second shells in RSF.c In ReS2, the distances between Re and neighboring six S atoms are different (Murray et al., 1994).d In ReS2, the distances between Re and neighboring three Re atoms are different (Murray et al., 1994).e Average and 1r of four reported values (Hassel, 1925; Kalikhman, 1983; Schoenfeld et al., 1983; Bronsema et al., 1986).f Average and 1r of 2 reported values (Knop and Reid, 1967; Stingl et al., 1992). Two reported values were idential for the Os–Os.

Lo

calato

mic

structu

reso

fR

ean

drad

iogen

icO

sin

mo

lybd

enite

5183


et al., 1986). These results clearly show that Re resides inthe Mo site of molybdenite.

Rhenium in ReS2 was also examined by EXAFS(Fig. 3). The CN of Re–S was optimized by the simulation,while CN for the second shell, Re–Re, was fixed at 3 as dis-cussed in Murray et al. (1994). The interatomic distanceand the CN of the first shell, Re–S, were calculated to be2.388 A and 5.9, respectively (Table 1). Although the crys-tallographic data of ReS2 suggest different distances be-tween Re and six S, our EXAFS and XANES datashowed that the local atomic environment of Re in molyb-denite is very similar to that of ReS2, which allows a signif-icant substitution of Mo by Re in molybdenite, aspreviously noted by many workers (e.g., Fleischer, 1959;Terada et al., 1971; Stein et al., 2001).

It has been suggested that Re may form its own phase,ReS2 in molybdenite (Newberry, 1979). However, this pos-sibility is discounted because the position of the secondpeak in RSF is different between Re in molybdenite andReS2. In addition, the amplitude of the k3v(k) function atk around 7 A�1 and the oscillation period of EXAFS areclearly different between Re in molybdenite and ReS2,which can be explained by contributions of Re–Mo in

10.50 10.52 10.54 10.56

Abs

orpt

ion

Energy (keV)

ReO4-

ReO3

ReO2

ReS2

Molybdenite

Re metal

Re2S7

Fig. 1. Rhenium LIII-edge XANES for Re in molybdenite and Rereference materials such as Re metal, ReIVS2, ReIVO2, ReVIO3,ReVII

2S7, and ReVIIO4� in aqueous solution.

0 1 2 3 4 5

SampleFit

0

5

10

15

20

Inte

nsity

R + ΔR (A)

-20

-10

0

10

20

3 4 5 6 7 8 9 10

Sample

Fit

k3χ

k (A-1)o

R e-S

R e-Mo

R e-S

R e-Mo

a

b

o

Fig. 2. EXAFS results for Re in molybdenite including (a) k3-weighted Re LIII-edge EXAFS spectra v(k) with the simulated spectrum,including the contribution of the first shell (Re–S) and second shell (Re–Mo) and (b) Radial structure functions obtained by Fouriertransforms from k3-weighted v(k) for Re with the simulation curve; an arrow shows the contribution of the Re–Mo shell at around 7 A, whichis different from that of the Re–Re shell in Fig. 3.



molybdenite and Re–Re in ReS2 (arrows in Figs. 2a and3a). The simulation of the EXAFS, assuming Re as the sec-ond neighboring atom in molybdenite, yielded a Re–Reinteractomic distance of 3.06 A. This distance is impossiblylonger than the interatomic distance of Re–Re in ReS2,2.77 A (Table 1). These results confirm that Re does notform its own phase, ReS2, in molybdenite.

3.3. Fluorescence XAFS at Os LIII-edge using a crystal

analyzer

It was impossible to record Os LIII-edge XAFS for radio-genic Os in the molybdenite by normal fluorescence modewith SSD, since intense fluorescence X-rays from Re andother elements conceal a small Os La peak (Fig. 4). In addi-tion, large scattering X-rays around 11 keV and fluorescenceX-rays below 7.2 keV (mainly from Fe) saturates SSDcounting, which makes the detection of a small Os La peakimpossible. The use of a crystal analyzer placed between thesample and the detector allowed us to extract the Os La

peak. Since some interferences on Os La fluorescence fromRe was still found in the XRF spectrum even with the use

of the analyzer, the quality of XAFS spectra of Os was fur-ther improved by the use of a Ge semiconductor detectorcoupled with the crystal analyzer system.

3.4. Osmium LIII-edge XAFS of Onganja molybdenite

Osmium LIII-edge XANES for Os in molybdenite weresuccessfully measured by the fluorescence mode using thecrystal analyzer system (Fig. 5). The main peak of Os inmolybdenite is close to those for Os metal and OsS2,whereas a shoulder of about 10.877 keV is present bothfor OsS2 and Os in molybdenite. However, the intensityof the main peak for Os in molybdenite is clearly largerthan those of Os metal and OsS2 (Fig. 5). It is known fromthe studies of platinum group elements that the intensity ofthe main peak of the electron transition from 2p to 5d isproportional to the hole of the 5d electron orbitals, or theoxidation state (Lytle et al., 1979; Sinfelt et al., 1981; Man-sour et al., 1984; Santiago et al., 2003; Rhee et al., 2003).This is confirmed by a linear positive correlation betweenthe main peak intensity and the oxidation state of Os forthe reference materials in this study (Fig. 6). The intensity

0 1 2 3 4 5

SampleFit

0

5

10

15

Inte

nsit

y

R + ΔR (A)

-10

0

10

4 6 8 10 12

Sample

Fit

k3χ

k (A-1)o

Re-S

Re-ReRe-S

Re-Re

a

b

o

Fig. 3. EXAFS results for Re in ReS2 including (a) k3-weighted v(k) with the simulated spectrum with the contribution of Re–S and Re–Reshells considered in the simulation and (b) RSF for Re with the simulation curve; an arrow at around 7 A shows the contribution of Re–Reshell.



was obtained as the area of the Lorentzian function, whichwas used to simulate XANES spectrum with the combina-tion of arc-tangent function. In the simulation, the relativeposition of Lorentzian and arc-tangent functions were fixedas constant (Takahashi et al., 2005). The simulation forOsO4 is shown as an example in Fig. 5. The linear relationand the area of Lorentzian function in the XANES spec-trum suggest that the average oxidation state of Os inmolybdenite is 3.6 ± 0.1, i.e., a mixture of Os(III) andOs(IV). This result is unexpected, since Os is divalent innatural sulfides and arsenides, such as erlichmanite (OsIIS2),anduoite (OsIIAs2), and ruarsite (Ru,Os)IIAsS. The pres-ence of Os at a higher oxidation state is possibly due tothe oxidation state of the parent, Re(IV), in molybdenite.The b-decay of Re(IV), which is accompanied by the emis-sion of an electron, may result in the formation of Os(V).The observed oxidation state of Os suggests that Os recap-tures electrons to form Os(IV), as suggested by Stein et al.(2003). In addition, electrons may freely move within elec-trically conductive molybdenite (Macintyre, 1992), whichallows further reduction of Os(IV) to Os(III). However, fur-ther study is necessary to evaluate the mechanism control-ling the oxidation state of radiogenic Os, or moregenerally the oxidation state of radiogenic isotopes in solidmedia.

EXAFS results of radiogenic Os in molybdenite andOsS2 are shown in Figs. 7 and 8, respectively. As shownin the previous section, Re is in the Mo site in molybdenite.It is expected that radiogenic Os remains in the Mo siteshortly after decay of 187Re, because b-decay of 187Rewould have little radiation effect in the matrices. Thus,the initial structure of Os in the molybdenite to run FEFFcalculation was obtained assuming that the Os is at the Mosite in molybdenite. The CNs of the first shell (Os–S) andthe second shell (Os–Mo) were fixed at six during the sim-ulation. The simulation was successfully completed yieldingthe parameters listed in Table 1. The calculated interatomicdistance for Os–S is 2.271 A; a similar value of 2.272 A is

obtained in the simulation that considers only the first shell.The distance is much shorter than those for Mo–S and Re–S (ca. 2.39 A) in molybdenite, whereas the distance of the

0

1

10

100

1000

4 5 6 7 8 9 10 11

Without crystal analyzerWith crystal analyzer

Inte

nsity

(cp

s)

Energy (keV)

Fe KαFe Kβ

Ni KαCu Kα

Zn KαR e Lα

Zn KβOs Lα

Molybdenite XRF

Fig. 4. X-ray fluorescence spectra of Onganja molybdenite irradi-ated at 11 keV with (solid line) and without (dashed line) crystalanalyzer system.

10.84 10.86 10.88 10.90

Abs

orpt

ion

OsCl3

OsO4

Os metal

Energy (keV)

OsS2

Molybdenite

OsO2

Fig. 5. Osmium LIII-edge XANES for Os in molybdenite andreference materials including Os metal, OsIIS2, OsIIICl3, OsIVO2,and OsVIIIO4. The simulation of the ‘‘white line’’ is shown for OsO4

where the results in bold combine the Lorentzian and arctangentfunctions (both in broken curves). The measured values, shown asopen circles, are similar to the simulated results, confirming thevalidity of the latter.

1

1.5

2

2 3 4 5 6 7 8

Rel

ativ

e ar

ea

Oxidation state

Os in molybdenite

Fig. 6. Relationship between the oxidation state of Os and the areaof Lorentzian function simulating the ‘‘white line’’ in Os LIII-edgeXANES. The simulation was conducted coupling the Lorentzianand arctangent functions as shown in Fig. 5. Circles are data of thereference materials used for the calibration, while the triangleshows the value for Os in molybdenite. The uncertainties of theestimated values are within the symbol size.



second shell (Os–Mo, 3.18 A) is similar to those of Re–Moand Mo–Mo (3.16 A).

The distance of Os–S in molybdenite is shorter than thedistance of OsII–S, 2.35 A, in erlichmanite (OsS2), which isexplained by the smaller sizes of Os(III) and Os(IV) in molyb-denite than Os(II). Within molybdenite, the distance of Os–Sis shorter than the distances of Re–S and Mo–S. The shorterdistance of Os–S suggests that the presence of Os produces adistortion in the local structure of molybdenite.

4. DISCUSSION

Other Os phases, such as metallic Os and OsIIS2, maypossibly form as a consequence of the redistribution ofOs in molybdenite. However, the EXAFS spectra indicatethat Os in molybdenite is not in the metallic form, sincethe first peak in the RSF is located at about 1.8 A (phaseshift uncorrected), which is much shorter than that of Osmetal, 2.2 A (Sakakibara et al., 2005). The possible forma-tion of OsS2 in molybdenite is also discounted, since (i) thesimulation assuming the structure of OsS2 (six S atoms wereassigned to the first and second shells each; Table 1) did notconverge, and (ii) a small peak of OsS2 at k = 6.5 (A)(Fig. 7) was not found in k3 v(k) spectrum for Os in molyb-denite (Fig. 6). These results are also supported by theXANES spectra of OsS2 and Os metal, which are dissimilarto that of Os in molybdenite, particularly in terms of the

magnitude of the ‘‘white line’’ and the position of the shoul-der at about 10.87–10.88 keV (Fig. 5).

One may think that the assumption of the initial struc-ture for the EXAFS analysis that Os is at the Re site maynot be appropriate when Os has already migrated withinthe sample. However, no secondary minerals hosting Os,such as Os metal or OsS2, were identified in the molybdeniteof our study. Thus, we conclude that Os is present at thecation site of molybdenite even after Os has migrated inthe sample. In the simulation of EXAFS above, we fixedthe CN value at 6, but this value can be also determinedby the simulation of the EXAFS spectra. As a result ofthe optimization, it was revealed that the CN value forOs in molybdenite is 6.5 (and Os–S distance is 2.255 A) inboth cases where one-shell and two-shells are consideredfor the simulation. This confirms that Os is surroundedby six S atoms, as is Mo in MoS2. Thus, the assumptionthat Os is at the cation site of molybdenite may be valid.

When MoS2 is compared to ReS2 and OsS2, the inter-atomic distance and the oxidation state of Re in ReS2 areidentical to Mo in molybdenite, but they are different fromOs in OsS2 (Table 1). In addition, S forms dianion S2

2� incubic OsS2 and S2� in hexagonal molybdenite and triclinicReS2. Hence, the chemical environment of Re in molybde-nite is almost identical to that of Re in ReS2, whereas thechemical state of Os in molybdenite is not similar to thatof Os in OsS2. The atomic-level environments of Re and

0 1 2 3 4 5

SampleFit

0

1

2

3

Inte

nsity

R + ΔR (A)

-15

-10

-5

0

5

3 4 5 6 7

SampleFit

k3 χ(k)

k (A-1)

Os-S

Os-Mo

o

o

Os-S

Os-Mo

a

b

Fig. 7. EXAFS results for Os in molybdenite including (a) k3-weighted v(k) with the simulated spectrum, including contribution of Os–S andOs–Mo shells considered in the simulation. (b) RSF for Os with the simulation curve.



Os explain the much larger initial abundance of Re than Osin molybdenite. Although the high Re/Os ratio at the timeof the mineral formation was already suggested in previousstudies (e.g., Morgan et al., 1968; Stein et al., 2001), the factwas supported here by the EXAFS analyses at the atomiclevel. As a result, molybdenite has been utilized as a suitablemineral for the Re–Os dating.

It has been suggested that radiogenic Os may be mobilein molybdenite, which causes the decoupling of Re–187Os(Stein et al., 2001, 2003; Kosler et al., 2003; Selby and Crea-ser, 2004). Our data show that radiogenic Os is smaller thanRe and that the oxidation state of Os, trivalent or tetrava-lent, is equal to or smaller than that of Re(IV) in molybde-nite. These results suggest that the decoupling is caused bythe faster diffusion of radiogenic Os relative to Re in molyb-denite, since a smaller cation with lesser charge diffuses fas-ter than a larger one with greater charge (e.g., Cherniak,1997; Van Orman et al., 2001; Cherniak, 2003). In addition,accumulation of radiogenic Os in defects in molybdenitealso suggested in these studies (Stein et al., 2001, 2003;

Selby and Creaser, 2004) is compatible with this diffusionmodel for the Re–187Os decoupling, since ions in crystalstend to diffuse toward defects present in the crystal (e.g.,Watson and Baxter, 2007).

As shown in the parabolic relationship between ionic ra-dii and partition coefficients of trace elements in minerals(Onuma et al., 1968), the partition coefficients of largerand smaller ions than the size of a cation at a given siteare smaller than for the ion in the site. This relationshipshows that crystal structure provides a certain size for eachsite and smaller or larger ions are not well accommodatedinto the site. Hence, Os does not physically fit as well asRe in the Mo site of molybdenite and the presence of Oslikely creates a minor distortion in the local crystal struc-ture of MoS2. Structural integrity plays an important rolein diffusivity of elements in the crystal with diffusivities ofelements being faster around distorted crystals (e.g.,Watson and Baxter, 2007). Consequently, Re–187Os ratiosmay be locally modified in an individual crystal (Steinet al., 2003; Kosler et al., 2003; Selby and Creaser, 2004).

0 1 2 3 4 5

SampleFit

0

5

10

15

20

25

Inte

nsit

y

R + ΔR (A)

-60

-50

-40

-30

-20

-10

0

10

20

2 4 6 8 10 12

SampleFit

k3 χ(k)

k (A-1)

Os-S (1st)

Os-Os

o

o

Os-S (2nd)

a

bOs-S

Os-Os

Os-S

o

Fig. 8. EXAFS results for Os in OsS2 including (a) k3-weighted v(k) with the simulated spectrum, including the contribution of two Os–Sshells and Os–Os shell considered in the simulation. (b) RSF for Os with the simulation curve.



It is well established that the 187Re–187Os geochronometerin molybdenite yields reliable ages when using the whole-molybdenite approach instead of a micro-sampling tech-nique, such as a laser ablation technique (Markey et al.,1998; Stein et al., 2001, 2003; Selby and Creaser, 2004; Stein,2006). Fast diffusion of 187Os may cause the decoupling ofRe–187Os in a small area in molybdenite, but the Os maylikely remain within the crystal. Our results confirm that thatRe–Os ages determined using micro-analytical techniquesmay not be reliable; dating of whole crystals or aggregatesof representative fine grains of molybdenite is recommended.

5. SUMMARY

XAFS spectra were successfully measured for Re andradiogenic 187Os in molybdenite using a crystal analyzer.The XANES and EXAFS results of Re and Os in molybde-nite show that (i) Re and radiogenic 187Os reside in the Mosite in molybdenite, (ii) the interatomic distance of Re toneighboring S and the oxidation state of Re in molybdeniteare similar to those for Mo in molybdenite and Re in ReS2,(iii) the interatomic distance of radiogenic 187Os to S issmaller by 0.12 A than those of Re–S and Mo–S in molyb-denite, and (iv) the oxidation state of radiogenic 187Os isOs(III) and Os(IV), similar to Re(IV) and Mo(IV) inmolybdenite. These results suggest that the oxidation stateof 187Os is related to the parent Re(IV), which induces theshorter interatomic distance of OsIV-S or OsIII–S in molyb-denite compared to that of OsII–S in erlichmanite. Due tothe small size and lower charge, radiogenic 187Os diffusesfaster than Re in molybdenite. Our results provide struc-tural chemical evidence for interpreting the geochemicalobservations reported in other studies; these include (i)the abundance of Os in molybdenite is much less than thatof Re, and (ii) Os has a higher mobility than Re, promotingthe decoupling of Re and 187Os in molybdenite.

ACKNOWLEDGMENTS

We thank Mrs. M. Fukukawa and Y. Yamamoto for their helpin XAFS experiments. Dr. McG. Miller is greatly acknowledgedfor providing the molybdenite sample. We are grateful to Prof.K. Hayashi (Okayama University of Science) for synthesizing theOsS2 sample. Finally, we thank an anonymous journal reviewerand Dr. H. Stein (Colorado State University) for their helpful com-ments, which have greatly improved this paper. This study was sup-ported by a Grant-in-Aid for Scientific Research from the JapanSociety for the Promotion of Science and funds from Shimadzu Sci-ence Foundation and Nippon Seimei Foundation. This work hasbeen performed with the approval of SPring-8 (Proposal No.2001B0393, R04A38B1-0023N, 2004B0169, and 2005B0181) andPhoton Factory, KEK (Proposal No. 2004G119 and 2004G334).

REFERENCES

Ankudinov A. L. and Rehr J. J. (1997) Relativistic calculations ofspin-dependent X-ray absorption spectra. Phys. Rev. B56,

1712–1716.

Bronsema K. D., de Boer J. L. and Jellinek F. (1986) On thestructure of molybdenum diselenide and disulfide. Zeitschrift

fur Anorganische und Allgemeine Chemie 540, 15–17.

Cherniak D. J. (1997) Rare-earth diffusion in zircon. Chem. Geol.

134, 289–301.

Cherniak D. J. (2003) REE diffusion in feldspar. Chem. Geol. 193,

25–41.

Clifford T. N. (1967) The Damaran Episode in the UpperProterozoic-Lower Paleozoic structural history of SouthernAfrica. Spec. Pap. Geol. Soc. Am. 92, 100.

Faure G. and Mensing T. M. (2004) Isotopes: Principles and

Applications. Wiley, New York.Fleischer M. (1959) The geochemistry of rhenium, with special

reference to its occurrence in molybdenite. Econ. Geol. 54,

1406–1413.

Hassel O. (1925) Ueber die Kristallstruktur des Molybdaenglanzes.Zeitschrift fuer Kristallographie, Kristallgeometrie, Kristallphys-

ik, Kristallchemie 61, 92–99.

Hayashi K., Serikawa D., Maeda N., Okamoto A. and Ikeuchi T.(2000) Microwave absorption materials-IV Preparation of themixed layer structure phases, RexTa1-xS2 and OsxTa1-xS2, andsome electrical and magnetic properties. J. Eur. Ceram. Soc. 20,

2735–2742.

Herr W. and Merz E. (1955) Eine neue Methode zur Alterbestim-mung von rhenium-haltigen mineralien mittels neutronaktivie-rung. Z. Naturforsch. 10a, 613–615.

Hirt B., Herr W. and Hoffmeister W. (1963) Age determinations bythe rhenium–osmium method. In Radioactive Dating. Interna-

tional Atomic Energy Agency, Vienna, pp. 35–43.

Kalikhman V. L. (1983) Characteristics of the crystal structure,electrophysical properties, and model of the valence bandspectrum of laminar compounds of molybdenum disulfide type.Acta Crystallogr. B 39, 404–407.

Knop O. and Reid K. I. G. (1967) Chalcogenides of thetransition elements. V. Crystal structures of the disulfidesand ditellurides of ruthenium and osmium. Can. J. Chem. 45,

1391–1400.

Kosler J., Simonetti A., Sylvester P., Cox R., Tubrett M. N. andWilton D. (2003) Laser ablation ICP-MS measurements of Re/Os in molybdenites and implications for Re–Os geochronology.Can. Mineral. 41, 307–320.

Lytle F. W., Wei P. S. P., Greegor R. B., Via G. H. and Sinfelt J. H.(1979) Effect of chemical environment on magnitude of X-rayabsorption resonance at LIII edges. Studies on metallic ele-ments, compounds, and catalysts. J. Chem. Phys. 70, 4849–

4855.

Maddock A. G., Sakanoue M., Willard J. E. and Hoffma D.C. (1992) Historical development of hot atom chemistryand prospects. In Handbook of Hot Atom Chemistry (eds. J.-P.Adloff, P. P. Gaspar, M. Imamura, A. G. Maddock, T.Matsuura, H. Sano and K. Yoshihara). VCH, Tokyo, pp. 3–31.

Mansour A. N., Cook J. W. and Sayers D. E. (1984) Quantitativetechnique for the determination of the number of unoccupiedd-electron states in a platinum catalyst using the L2,3 X-rayabsorption-edge spectra. J. Phys. Chem. 88, 2330–2334.

Markey R., Stein H. and Morgan J. (1998) Highly precise Re–Osdating for molybdenite using alkaline fusion and NTIMS.Talanta 45, 935–946.

Macintyre, J. E. (ed.) (1992) Dictionary of Inorganic Compounds,vol. 3. Chapman & Hall, London, pp. 3594.

Morgan J. W., Lovering J. F. and Ford R. J. (1968) Rhenium andnon-radiogenic osmium in Australian molybdenites and othersulphide minerals by neutron activation analysis. J. Geol. Soc.

Aust. 15, 189–194.

Murray H. H., Kelty S. P., Chianelli R. R. and Day C. S. (1994)Structure of rhenium disulfide. Inorg. Chem. 33, 4418–4420.

Newberry R. J. J. (1979) Polytypism in molybdenite (II): relation-ships between polytypism, ore deposition/alteration stages andrhenium contents. Am. Mineral. 64, 768–775.



Nomura M. (1998) Dead-time correction of a multi-element SSDfor fluorescent XAFS. J. Synchrotron Rad. 5, 851–853.

Nomura M. and Koyama A. (1996) Design and performance of anew XAFS beamline at the photon factory; BL-12C. KEK Rep.,15–95.

O’Day P. A., Rehr J. J., Zabinsky S. I. and Brown G. E. (1994)Extended X-ray absorption fine structure (EXAFS) analysis ofdisorder and multiple scattering in complex crystalline solids.J. Am. Chem. Soc. 116, 2938–2949.

Onuma N., Higuchi H., Wakita H. and Nagasawa H. (1968) Traceelements partition between two pyroxenes and the host lava.Earth Planet. Sci. Lett. 5, 47–51.

Parrington J. R., Knox H. D., Breneman S. L., Baum E. M. andFeiner F. (1996) Nuclides and Isotopes, 15th ed. GE NuclearEnergy, San Jose.

Raith J. G. and Stein H. J. (2000) Re–Os dating and sulfur isotopecomposition of molybdenite from tungsten deposits in westernNamaqualand, South Africa: implications for ore genesis andthe timing of metamorphism. Mineral. Dep. 35, 741–753.

Ravel B. (2001) ATOMS: crystallography for the X-ray absorptionspectroscopy. J. Synchrotron Rad. 8, 314–316.

Rhee C. K., Wakisaka M., Tolmachev Y. V., Johnston C. M.,Haasch R., Attenkofer K., Lu G. Q., You H. and WieckowskiA. (2003) Osmium nanoislands spontaneously deposited on aPt(111) electrode: an XPS, STM and GIF-XAS study.J. Electroanal. Chem. 554, 367–378.

Sakakibara N., Takahashi Y., Okumura K., Hattori K. H., Yaita T.,Suzuki K. and Shimizu H. (2005) Speciation of osmium in an ironmeteorite and a platinum ore specimen based on X-ray absorptionfine-structure spectroscopy. Geochem. J. 39, 383–389.

Santiago E. I., Giz M. J. and Ticianelli E. A. (2003) Studies ofcarbon monoxide oxidation on carbon-supported platinum–osmium electrocatalysts. J. Solid State Electrochem. 7, 607–613.

Schieder R., Sorgalla K. H. and Herr W. (1981) The determinationof osmium isotope abundances by high resolution laser spec-troscopy. Radiochim. Acta 28, 109–113.

Selby D. and Creaser R. A. (2001a) Re–Os geochronology andsystematics in molybdenite from the Endako porphyry molybde-num deposit, British Columbia, Canada. Econ. Geol. 96, 197–204.

Selby D. and Creaser R. A. (2001b) Late and mid-Cretaceousmineralization in the Northern Canadian Cordillera: constraintsfrom Re–Os molybdenite dates. Econ. Geol. 96, 1461–1467.

Selby D. and Creaser R. A. (2004) Macroscale NTIMS andmicroscale LA-MC-ICP-MS Re–Os isotopic analysis of molyb-denite: testing spatial restrictions for reliable Re–Os age deter-minations, and implications for the decoupling of Re and Oswithin molybdenite. Geochim. Cosmochim. Acta 68, 3897–3908.

Schoenfeld B., Huang J. J. and Moss S. C. (1983) Anisotropicmean-square displacement (MSD) in single crystals of 2H- and3R-MoS2. Acta Crystallogr. B 39, 404–407.

Sinfelt J. H., Via G. H., Lytle F. W. and Greegore R. B. (1981)Structure of bimetallic clusters. Extended X-ray absorptionfine-structure (EXAFS) studies of Os–Cu clusters. J. Chem.

Phys. 75, 5527–5537.

Smoliar M. L., Walker R. J. and Morgan J. W. (1996) Re–Os agesof group IIA, IIIA, IVA, and IVB iron meteorites. Science 271,

1099–1102.

Stein H. J. (2006) Low-rhenium molybdenite by metamorphism innorthern Sweden: recognition, genesis, and global implications.Lithos 87, 300–327.

Stein H. J., Markey R. J., Morgan J. W., Du A. and Sun Y. (1997)Highly precise and accurate Re–Os ages for molybdenite fromthe East Qinling molybdenum belt, Shaanxi Province, China.Econ. Geol. 92, 827–835.

Stein H. J., Sundblad K., Morgan J. W., Markey R. J. and MotuzaG. (1998) Re–Os ages for Archean molybdenite and pyrite,

Kuittila, Finland and Proterozoic, Kabeliai, Luthuania: ametamorphic and metasomatic test for the chronometer.Mineral. Dep. 33, 329–345.

Stein H. J., Markey R. J., Morgan J. W., Hannah J. L. andScherste’n A. (2001) The remarkable Re–Os chronometer inmolybdenite: how and why it works. Terra Nova 13, 479–486.

Stein H., Scherst’en A., Hannah J. and Markey R. (2003) Subgrain-scale decoupling of Re and 187Os and assessment of laserablation ICP-MS spot dating in molybdenite. Geochim. Cos-

mochim. Acta 67, 3673–3686.

Stingl T., Mueller B. and Lutz H. D. (1992) Crystal structurerefinement of osmium(II) disulfide. Zeitschrift fuer Kristallog-

raphie 202, 161–162.

Stern E. A. (1993) Number of relevant independent points in X-rayabsorption fine-structure spectra. Phys. Rev. B48, 9825–9827.

Suzuki K., Qilu Shimizu H. and Masuda A. (1992) Determination ofosmium abundance in molybdenite mineral by isotope-dilutionmass-spectrometry with microwave digestion using potassiumdichromate as oxidizing-agent. Analyst 117, 1151–1156.

Suzuki K., Shimizu H. and Masuda A. (1996) Re–Os dating ofmolybdenites from ore deposits in Japan: implications for theclosure temperature of the Re–Os system for molybdenite andthe cooling history of molybdenum ore deposits. Geochim.

Cosmochim. Acta 60, 3151–3159.

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

Takahashi Y., Minamikawa R., Hattori H. K., Kurishima K.,Kihou N. and Yuita K. (2004) Arsenic behavior in paddy fieldsduring the cycle of flooded and non-flooded periods. Environ.

Sci. Technol. 38, 1038–1044.

Takahashi Y., Kolonin G. R., Shironosova G. P., Kupriyanova I.I., Uruga T. and Shimizu H. (2005) Determination of theEu(II)/Eu(III) ratios in minerals by X-ray absorption near-edgestructure (XANES) and its application to hydrothermaldeposits. Min. Mag. 69, 177–188.

Takahashi Y., Uruga T., Tanida H., Terada Y., Nakai S. andShimizu H. (2006) Application of X-ray absorption near-edge structure (XANES) using bent crystal analyzer to thespeciation of trace Os in iron meteorites. Anal. Chim. Acta

558, 332–336.

Terada K., Osaki S., Ishihara S. and Kiba T. (1971) Distribution ofrhenium in molybdenites. Geochem. J. 4, 123–141.

Terada Y., Goto S., Takimoto N., Takeshita K., Yamazaki H.,Shimizu Y., Takahashi S., Ohashi H., Furukawa Y., Matsush-ita T., Ohata T., Ishizawa Y., Uruga T., Kitamura H., IshikawaT. and Hayakawa S. (2004) Construction and Commissioningof BL37XU at SPring-8. AIP Conf. Proc. 705, 376–379.

Van Orman J. A., Grove T. L. and Shimizu N. (2001) Rare earthelement diffusion in diopside: influence of temperature,pressure, and ionic radius, and an elastic model for diffusionin silicates. Contrib. Mineral. Petrol. 141, 687–703.

Watanabe Y. and Stein H. J. (2000) Re–Os ages for the Erdenetand Tsagaan Suvarga porphyry Cu–Mo deposits, Mongolia,and tectonic implications. Econ. Geol. 95, 1537–1542.

Watson E. B. and Baxter E. F. (2007) Diffusion in solid-Earthsystems. Earth Planet. Sci. Lett. 253, 307–327.

Zavinsky S. I., Rehr J. J., Ankudinov A., Albers R. C. and Eller M.J. (1995) Multiple scattering calculations of X-ray absorptionspectra. Phys. Rev. B52, 2995–3009.

Zhong Z., Chapman D., Bunker B., Bunker G., Fischetti R. andSegre C. (1999) A bent Laue analyzer for fluorescence EXAFSdetection. J. Synchrotron Rad. 6, 212–214.

Associate editor: Edward M. Ripley


groer test - dynamic-eye

Documents