Chemistry of metal sulÐdes in anoxic sediments

G. Billon,a B. Ouddane,a J. Laureynsb and A. Boughriet*c

des Sciences et T echnologies de L ille 1, 59655 V illeneuve dÏAscq Cedex, Francea Universite�b L aboratoire de Spectrochimie Infrarouge et Raman CNRS: Centre dÏEtudes et de Recherches

L asers et Applications, de L ille 1, 59655 V illeneuve dÏAscq Cedex, FranceUniversite�dÏArtois, I.U.T . de de Chimie, Rue de B.P. 819,c Universite� Be� thune De� partement lÏUniversite� ,

62408 Cedex, France. E-mail :Be� thune abdelatif.boughriet=univ-artois.fr

Received 14th March 2001, Accepted 27th June 2001First published as an Advance Article on the web 31st July 2001

Using sequential extraction of solid sulÐdes, the determination of acid volatile sulÐdes (AVS) and chromiumreducible sulfurs (CRS) in anoxic sediments from the Authie Bay (in northern France) has been undertakenbecause of the importance of the sediments as sinks for iron, sulfur and trace metals and as possible sources ofpollution when reduced sediments are mixed with oxic waters (as a result of a sediment remobilization inducedby physical disturbances such as tidal currents and dredgings), and subsequently oxidized. Chemical analysis ofsolutions recovered after sequential leaching of sediments with 1 M HCl, 1 M HF and concentrated hasHNO3enabled us to obtain proÐles, vs. sediment depth, of trace metals associated with pyrite. Porewaterconcentration proÐles vs. depth have been determined for several cations (Ca2`, Cd2`, Cu2`, Fe2`, Mg2`,Mn2`, Na`, Pb2`, Sr2` and Zn2`) and anions and S2~). Using the chemical(CO32~, PO43~, SO42~equilibrium modeling program MINEQL` with these analytical data, thermodynamic calculations have giveninformation about the possibility of precipitation of discrete metal sulÐde phases (FeS as greigite andamorphous FeS ; ZnS, PbS, CuS and CdS), and coprecipitation with adsorption on solid FeS to produce solidsolutions with iron sulÐdes. The degree of trace metal pyritization, DTMP, has been determined for thesemetals and compared to the degree of pyritization, DOP. The Ðndings suggest that in Authie-bay sedimentsMn is well pyritized ; whereas Zn, Cu, Ni and above all Cd are weakly pyritized (MnAZn^Cu[NiA Cd).These observations seem to be intimately related to the existence of the discrete/separate solid phases CuS,CdS and ZnS, as predicted by thermodynamic calculations. Finally, analysis of crude sediments, heavyminerals and pyrite extracted by a heavy liquid density separation method, has been performed with a Ramanmicroprobe to gain information about the geochemical and mineralogical characteristics of these sediments.The efficiency of sequential leachings of sediments (which were used for sedimentary pyrite recovery/attack andanalysis of pyritic Fe and trace metal) has also been evaluated by these techniques.

IntroductionIn buried sediments, bacteria use oxidized forms of metals aselectron acceptors to produce soluble metallic ions mainly :Fe2` and Mn2` (as major elements) ; and Zn2`, Pb2`, Cu2`and Cd2` (as trace elements).1,2 Moreover, dissolved andH2Sother sulfur compounds (e.g., polysulÐdes) are produced inporewaters by bacterial reduction of sulfate ions.2 As a result,formation of metal sulÐdes can occur if the sedimentarymedium becomes (super)saturated with respect to the sulÐdesof reduced trace metals.3h5 Note that iron is usually the mostpredominant sulÐde-generating metal. Indeed, Fe is releasedfrom reducible minerals such as oxides/oxihydroxides and sili-cates within the sedimentary layers in which sulfate reductionoccurs6 to give ferrous iron. This subsequently reacts with dis-solved to produce amorphous FeS and/or crystallizedH2SFeS (such as mackinawite and greigite), which are consideredas precursors to pyrite generation.7h12 Numerous papers havebeen written about the possible combination of trace metalswith iron sulÐde minerals (for instance, see ref. 13È20). Theauthors have suggested that during the sulÐdization processsome trace metals, that are liberated from sedimentary com-ponents such as organic matter and metal oxides, can adsorbor coprecipitate with FeS minerals or can precipitate directlyas discrete/separate solid phases. Overall, these sulÐde min-erals constitute secondary sources of contamination when““pollutedÏÏ anoxic sediments are remobilized in the water

column under the e†ects of resuspension events, and hence,are oxidized by dissolved oxygen.

In this study, we Ðrst report analytical data on AVS (i.e.,amorphous FeS, greigite, mackinawite . . .), CRS (i.e., pyriticcompounds), and trace metals bound to pyrite in anoxic sedi-ments taken at di†erent depths from the Authie Bay (in north-ern France).

Secondly we undertake thermodynamic calculations fromexperimental results on porewaters in order to predict thepossible generation of metal sulÐdes (through precipitation,coprecipitation and/or adsorption) as discrete solid phases,coprecipitates or solid solutions.

Finally, we analyze crude sediments, heavy minerals andmore particularly pyrite (that were previously extracted byÑoatÈsink methods with dense liquids ; e.g., see ref. 21È24), andresidual solids (that were recovered from sequential leachingsof sediments) to obtain complementary information about thegeochemical and mineralogical characteristics of Authie-baysediments and to validate pyrite extraction procedures/analyses.

Experimental

Location and sampling

The sampling location is shown in Fig. 1. The Authie bay waschosen because it is considered to be a site without signiÐcant

3586 Phys. Chem. Chem. Phys., 2001, 3, 3586È3592 DOI: 10.1039/b102404n

This journal is The Owner Societies 2001(

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online / Journal Homepage / Table of Contents for this issue

Fig. 1 Location of the Authie bay (in northern France) where sedi-ment samplings were taken.

metallic pollution. Sediment sampling was carried out at lowtide during November 1998. Several cores were taken in theÐeld with a 30 cm long Perspex tube. The cores were imme-diately isolated from the atmosphere (to prevent any oxida-tion reactions) inside a plastic bag which had been previouslypurged with nitrogen. They were later sliced into 2 cm layersunder a atmosphere inside a glove box, packed individ-N2ually in plastic containers previously purged with nitrogengas, and stored in an ice box. In the laboratory, the slicedsediment samples were centrifuged at room temperature usingan X340 Prolabo centrifuge (with a rotation speed of 4000rpm). Anoxic porewaters were afterwards recovered under anitrogen atmosphere in a glove box, and Ðltered with 0.45 lmAlltech Ðlters (cellulose acetate membrane), and immediatelyacidiÐed (except for reduced sulfur analyses) with ultra-purenitric acid [100 ll in 10 ml of porewater] before elementalanalysis.

Chemicals

The water used for all the experiments was provided by aMilli-Q Plus Ðlter apparatus (Millipore). All the chemicalsused for electrochemical analysis and extraction procedureswere of analytical reagent grade : sodium chloride, sodiumhydroxide and sodium acetate (Merck, suprapur) ; sodiumhydrogenocarbonate (Prolabo, normapur) ; toluene, hydro-Ñuoric acid 40% and 1,1@,2,2@-tetrabromoethane (Merck,[98.5%) ; disodium sulÐde nonahydrate (Sigma) ; methanol(Scharlau, extra-pure), hydrochloric acid and nitric acid(Merck, suprapur).

A standard solution of elemental sulfur was prepared bydissolving 100 mg of S in 50 cm3 of toluene ; prior to use thesulfurÈtoluene mixture was diluted with extra-pure methanol.

The standard solution of sulÐde ions was prepared by dis-solving 1 g of in a 1 mol dm~3 NaOH solutionNa2S É 9H2Owhich had previously been deoxygenated. Immediately priorto use, the sodium sulÐde solution was titrated with a 10~3mol dm~3 cadmium(II) solution, because of the relative insta-bility of sulÐde ions in water (in our experimental conditions,we found that the standard solution was sufficientlyNa2Sstable for at least 10 h).

Analytical procedures

Porewater analyses. The pH was determined at variousdepths (in cm) in the Ðeld with a combined glass electrode(Ingold) which is introduced directly in the sediment core justafter the sampling.

The concentrations of the elements (Ca2`, Cd2`, Cu2`,Fe2`, Mg2`, Mn2`, Na`, Pb2`, Sr2` and Zn2`) inPO43~,porewaters (and sediment solid) were determined using induc-tively coupled plasma atomic emission spectroscopy (ICP-AES; Varian, Liberty Serie II, axial view) and graphite-furnace atomic absorption spectroscopy with Zeeman correc-tions (GFAAS-ZC; Varian ; model SpectrAA-300).

Sulfate in interstitial waters was determined by a neph-elometric method (spectrophotometer : Kontron Instruments ;model UVIKON860) at 650 nm after its precipitation andresuspension as barium sulfate in the presence of polyoxyethy-lenesorbitan monolaurate.25

Inorganic carbon (i.e., and in theH2CO3 , HCO3~ CO32~)porewaters was analyzed with a 5] 10~3 mol dm~3 HClsolution using an automatic pH titrator (Metrohm; modelTitrino 736 GP).

The analysis of sulÐde made use of the mercury drop elec-trode that interacts favorably with a number of inorganicsulfur compounds present in natural waters.26h31 To analyzereduced and dissolved sulfur species (reduced sulfurs and ele-mental sulfur), voltammograms of porewater samples wererecorded using an Autolab microprocessor which wasequipped with an IME 663 Metrohm module element, ahanging drop mercury electrode as working electrode, a glassycarbon electrode as an auxiliary electrode, and an Ag/AgCl (3mol dm~3 KCl) electrode as reference electrode.

The dissolved contents of sulÐdes, polysulÐdes, elementalsulfur and possibly colloidal sulfur (all these S species are sym-bolized in the text as reduced sulfurs) were determined inporewaters by linear-sweep voltammetry (LSV). For thatpurpose, porewater samples were previously diluted from 5 to10 times (depending upon the content of reduced S in theanalyzed sample) in a pH 10 bu†er mixture of Na2CO3Note that the instrumental parameters used for] NaHCO3 .recording linear sweep cathodic stripping voltammogramswere optimized to analyze interstitial waters adequately.32

A somewhat similar procedure described by Batina andcoworkers28 has been used for the electrochemical analysis ofelemental sulfur in porewaters, but, with some modiÐcations.Instead of acidifying the porewater sample with nitric acid, weused a bu†er mixture of acetic acid] sodium acetate at pH 5and, to this acidiÐed medium, we added an equivalent volumeof extra-pure methanol. This resulting solution was analyzedusing the LSV technique and the method of standard addi-tion.

From the electrochemical data obtained for reduced S andelemental sulfur, it is possible to calculate the content ofsulÐde ions present in the studied porewaters from therelationship :

[Reduced S]total[ [Elemental S]^ [HS~]

Sedimentary solid phase analyses. Reduced sulfur species(i.e., AVS and CRS) were determined after their conversioninto gas by following sequential extraction proceduresH2Spreviously described by CanÐeld et al.,33 Cornwell andMorse34 and, more recently, by us.32 BrieÑy, a cold extractionprocedure (6 M HCl for 1 h) is presumed to extract quantitat-ively all AVS but not pyritic sulÐdes. Consecutively, a hotdigestion for 6 h of the recovered sediment sample with Cr(II)reduction reagent leads to a signiÐcant recovery of sedimen-tary pyrite.

Reactive iron present in the sediment (which representsmainly the iron bound to the AVS and carbonates, as well aslabile iron oxides) was extracted at room temperature for 24 husing a 1 M HCl solution.13

The values of AVS, CRS and reactive iron permit us to cal-culate two ratios : the degree of sulÐdization (DOS) and thedegree of pyritization (DOP). DOS is deÐned as the molarratio of sedimentary iron bound to sulfur (i.e., Fe in

Phys. Chem. Chem. Phys., 2001, 3, 3586È3592 3587

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online

AVS] CRS) to the total reducible and reactive iron in thesediment :

DOS \MAVS-FeN] MCRS-FeNMHCl-FeN] MCRS-FeN

DOP is deÐned as the molar ratio of pyrite iron to the totalreducible and reactive iron in the sediment :

DOP\MCRS-FeN

MCRS-FeN] MHCl-FeN

Micro-Raman spectra of anoxic sediment samples(previously dried in a vacuum over at room tem-P4O10perature for 3 days) were recorded on a DILOR-JOBYN-YVON LABRAM confocal micro-spectrometer using aHeÈNe laser (632.8 nm) at a typical power of 1.5 mW in theimmediate proximity of the sample. Both the incident lightand Raman scattering were focussed and collected through anOlympus BX 40 microscope with a ULWD (ultra longworking distance) ]80/0.75 objective. Detection of theRaman scattering was achieved with a 2D, air-cooled chargecoupled device camera (CCD) from Wright Instruments. TheLABRAM spectrometer was used in the confocal laser point-by-point mode combined with a scanning sample stage whenRaman mapping is performed. Then, each pixel of the imagecontains a characteristic Raman spectrum and represents anarea of 4 lm ] 4 lm or 1 lm ] 1 lm of the sample ; for thetwo kinds of samples under study, a signiÐcant array of 160lm ] 120 lm and 64 lm ] 80 lm was predeÐned and so,40 ] 30 spectra and 64] 80 spectra were collected to elabo-rate two Raman images. The acquisition time was respectively15 s and 3 s per spectrum for a total recording time of approx-imately 6 h including automatic saving to disk. The entranceslit was set to 150 lm and the confocal hole adjusted to 500lm. When several di†erent and independent molecular speciesare present in the sample under study an interesting way toelaborate Raman mapping is by decomposition of each spec-trum of the raw image into the combination of referencespectra chosen in the set of the recorded spectra.35,36 Whenno Raman image is required, the LABRAM spectrometer wasused in the classical confocal point mode ; in this case, the]100/0.9 objective allows a lateral resolution of 2 lm andattenuates the strong Ñuorescence signals which are oftenencountered on the sediment samples.

Results

Chemical speciation of sedimentary sulfur

The oxidation of sedimentary sulÐde phases represent animportant source/release of ““ toxic ÏÏ metals to oxic overlyingwaters during resuspension events such as strong tidal cur-rents and dredgings. Therefore, to assess both the quality cri-

teria for Authie-bay sediments and the global iron and sulfurÑuxes in this sedimentary environment, it is necessary todetermine the contents of these sulfur compounds vs. sedimentdepth. For that purpose, to attack selectively AVS (amor-phous FeS, greigite, mackinawite . . .) and CRS (pyriticcompounds), we carried out two consecutive chemical treat-ments on the sliced sediments. The AVS and CRS values vs.depth are reported in Table 1. They display an abrupt increasebelow 1 cm depth. The AVS proÐle reaches a Ðrst maximumvalue at a depth of 9È11 cm (ca. 1.09È1.21 g of S per kg ofcrude sediment), and below this depth the AVS content Ñuctu-ates between 0.80 and 1.73 g of S per kg of crude sediment.The CRS proÐle attains a maximum value more rapidly at adepth of 5 cm (ca. 1.8 g of S per kg of crude sediment), andbelow this depth it remains nearly constant at a sulfur contentof ca. 1.3 g of S per kg of crude sediment.

The values of the DOS and the DOP vs. depth are alsoreported in Table 1. The DOS proÐle is characterized by aslight increase from 0.4 up to a maximum value of ca. 1.0which is reached at a depth of 9È10 cm, below this depth theDOS Ñuctuates weakly between 0.9 and 1.2. Note that someDOS values in Table 1 are greater than 1.0 because of experi-mental error (D10%) and also the possible existence of otherforms of sulÐde and iron compounds at depths P9 cm whichmight be extracted by the chemical treatment used for thedetermination of AVS. The DOP proÐle remains nearly con-stant along the entire length of this sediment core. The ratio ofsulfur contents in AVS and CRS, (AVS)-S/(CRS)-S, increasessigniÐcantly with depth to reach a maximum value, 2.5, at adepth of 17È18 cm.

Overall, these experimental data indicate that the conver-sion of AVS (FeS) into CRS is difficult in this sedimen-(FeS2)tary medium when compared to AVS ] CRS transformationrates measured in other coastal marine sediments.16

The low contents of AVS and CRS found in Authie-baysediments, the rapid oxidation of AVS and the difficultiesinvolved in their separation from the bulk sediment have pre-vented us carrying out direct analysis of reduced sulfur solids.Thermodynamic calculations have thus been used as an indi-rect means of gaining information about the exact chemical/crystalline forms of the reduced-metal sulÐdes involved inthese sediments.

Thermodynamic predictions

To appraise the extent of the interactions between reducedmetals and dissolved sulfur species in anoxic sediments fromthe Authie bay, porewater concentration proÐles were Ðrstdetermined for di†erent ionic species present in interstitialwaters [cations : Ca2`, Cd2`, Cu2`, Fe2`, H`, Mg2`, Mn2`,Na`, Pb2`, Sr2` and Zn2` ; anions : CO32~, PO43~, SO42~and S2~]. These analytical data were afterwards studied withthe chemical equilibium modeling program, MINEQL`elaborated by Schecher and McAvoy,37 in order to determine

Table 1 Values vs. depth of AVS, CRS, DOS, DOP and the ratio AVS/CRS, measured in the sediment core collected in the Authie bay

Depth/cm AVS/(mg of S) kg~1 AVS/(mg of S) kg~1 AVS/CRS DOSa DOPa

1 68 691 0.20 0.44 0.363 503 1037 0.97 0.75 0.385 701 1763 0.80 0.78 0.447 765 1324 1.16 0.86 0.409 1093 1184 1.85 1.18 0.41

11 1211 1367 1.77 0.97 0.3513 804 1197 1.34 0.91 0.3915 1168 1156 2.02 1.00 0.3317 1727 1268 2.73 1.24 0.3319 1332 1113 2.39 1.15 0.34

a Experimental error : D10%.

3588 Phys. Chem. Chem. Phys., 2001, 3, 3586È3592

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online

Table 2 IAP, calculated vs. depth for the following cations (in interaction with the anion S2~) : Cd2`, Cu2`, Fe2`, Mn2`, Pb2` and Zn2`

Depth/cm [log[(Fe2`)(S2~)] [log[(Mn2`)(S2~)]

1 20.0 19.43 20.4 19.75 20.2 19.67 20.0 19.59 18.8 18.9

11 18.6 18.613 18.4 18.315 18.6 18.319 19.2 19.0

pKs : 17.5 (mackinawite) ; 18.3 (greigite)a 10.5 (pink) ; 13.5 (green)b

[log[(Cu2`)(S2~)] [log[(Cd2`)(S2~)] [log[(Pb2`)(S2~)] [log[(Zn2`)(S2~)]

1 24.1 24.8 23.1 21.43 25.1 È 23.2 21.25 26.1 25.2 22.9 21.27 25.4 25.4 23.0 21.29 È È È 21.6

11 28.5 È 22.6 22.113 È 25.5 23.0 22.315 28.9 26.1 23.2 22.419 27.8 25.1 22.5 21.9

pKs : 36.1b 27.0c 27.5b 24.7 (a) ; 22.5 (b)c

a Ref. 46. b Ref. 38. c Ref. 39.

the concentration of free Me2` (dissolved reduced metal) andS2~ ions and thereby to obtain the ionic activity product :IAP\ (Me2`)(S2~) [where (Me2`) and (S2~) represent theactivities of Me2` and S2~, respectively].

Fe and Mn sulÐdes. Values of IAP vs. depth for iron andmanganese are reported in Table 2, where they are also com-pared with the solubility products of FeS and MnS. The(pKs)IAP values obtained vs. depth for porewater iron sulfur, ingeneral, agree best with the solubility products of amorphousFeS and greigite although the existence of mackinawite insome sedimentary layers cannot be excluded. In the case ofmanganese, the porewaters are undersaturated with respect tothe solubility product of pure MnS in the crystalline form:suggesting that removal of Mn2` ions from anoxic sedimentsoccurs by coprecipitation with or adsorption on solid FeS to

Fig. 2 DTMP vs. DOP determined for Ni, Zn, Mn, Cd, Cu and Pbin some Authie-bay samples.

produce solid solutions with iron sulÐdes. This phenomenonhas also been observed, for instance, by Huerta-Diaz andcoworkers19 for freshwater sediments collected in two Cana-dian Shield lakes (Chevreuil and Clearwater).

Trace metal sulÐdes. The IAP proÐles vs. depth determinedfor Zn, Pb, Cu and Cd indicate that the porewaters are over-saturated with respect to the solubility product of ZnS, PbS,CuS and CdS, respectively. This implies that the reduced tracemetals (Zn2`, Pb2`, Cu2` and Cd2`) react in anoxic porewa-ters (extracted from the Authie-bay sediments at di†erentdepths) with sulÐde ions to give discrete/separate solid phasesof ZnS, PbS, CuS and CdS, in agreement with recentlypublished works on anoxic sediments derived from two Cana-dian Shield lakes.19

Trace-metal pyritization. Pyrite is an important sink in sedi-ment for certain trace elements.13,19,20 Detailed investigationson the mechanism of incorporation of some trace metals suchas Hg, Cu, Ni, Co, Mn, Pb, Zn and Cd into pyrite have beenmade recently by Huerta-Diaz and coworkers19 and Morseand Luther III.20 Note further that trace metal-pyrite inter-actions can be important factors controlling the bio-availability of toxic metals in sediments.40h42 In this context,the sinking/scavenging characteristics of pyrite for some tracemetals (Zn, Pb, Cu and Cd) in anoxic sediments from theAuthie bay is examined. Three chemical treatments werecarried out on heavy minerals (after their extraction by ÑoatÈsink methods with dense liquids) : (i) 1 M HCl at room tem-perature over night to dissolve reactive-iron fractions,namely : carbonates, AVS and most of the Fe and Mn oxides(except goethite) ; (ii) 1 M HF digestion at room temperaturefor 24 h to remove silicate-iron phases ; and (iii) concentratednitric acid to attack recovered pyrite. The analysis of theresulting solutions permits the determination of pyritic tracemetal pyritization, DTMP, deÐned as :13

DTMP\MCRS-MeN

MCRS-MeN] MHCl-MeN100

where MCRS-MeN and MHCl-MeN represent the contents of thetrace metal, Me, present in pyrite and in the HCl fraction (asdescribed above), respectively.

This parameter is often used to appraise the extent of pyriteas a trace-metal reservoir in sediments [e.g., ref. 19 and 20].

Phys. Chem. Chem. Phys., 2001, 3, 3586È3592 3589

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online

Fig. 3 2D and 3D views of Raman images of molecular species and corresponding typical spectra. A : (anatase) ; B : (rutile) ; C :TiO2 TiO2kaolin ; D: zircon ; E : pyrite ; F : Ñuorescent unknown particle ; G: paper Ðlter.

DTMP is plotted against DOP for di†erent sediment samplescollected from the Authie Bay (see Fig. 2) to Ðnd the corre-lation between the trace-metal pyritization process andprimary pyrite generation in the bulk sediment. These plotsshow that in Authie-bay sediments Mn is most extensivelypyritized, whereas Zn, Cu, Ni and especially Cd are much lesspyritized. In the case of lead, this metal is associated with sul-Ðdes in the sediments to give both a solid solution with pyriteand a discrete/separate solid phase (as PbS). Overall, DTMPmeasured at di†erent sediment depths decreases in the follow-ing order : Note this generalMn A Zn^ Cu [NiACd.pattern does not di†er signiÐcantly from those described byMorse and Luther III20 and Huerta-Diaz and coworkers14 inother anoxic sediments, except for copper, probably becauseof the possible co-existence of several Cu-sulÐde minerals insedimentary environments.20 Accordingly, the large discrep-ancies of biogeochemical properties of anoxic sediments foundin estuaries should explain the wide and variable scatter oftenobserved for DTMP-copper, and hence, any comparison ofDTMP-Cu values with those reported in the literature mustbe treated with caution. As pointed out by Morse and LutherIII,20 to better understand the general pattern found forDTMP-trace metal, the thermodynamic and kinetic character-

istics of these metals should be addressed. In the case of man-ganese, its incorporation into pyrite occurs more easily fortwo principal reasons : (i) MnS is not generated as a discrete/pure phase in the bulk sediment but as a solid solution in ironsulÐdes (as predicted above by thermodynamic calculations) ;and (ii) the presence of solid solutions, inMn

xCa1~x

CO3 ,sediments is generally considered by geochemists as a relevantsource/reservoir of dissolved manganese(II) for interstitialwaters that can afterwards adsorb onto FeS phases and bepyritized with time.43,44

The potential existence of solid solutions inMnxCa1~x

CO3Authie-bay sediments is discussed below. In contrast, zinc,copper and above all cadmium have lower DTMP values vs.DOP, indicating a poor association of these trace metals withsedimentary pyrite. A possible explanation of this behavior isthat Zn, Cu and Cd form discrete/pure MeS phases in AVS (aspredicted above) that dissolve in cold HCl, prior to FeS for-mation and subsequent pyrite formation. Indeed, the gener-ation of pure zinc, copper and cadmium sulÐdes in theseanoxic sediments should slow down the process of Zn, Cd andCu incorporation into pyrite if there are kinetic constraints onthe dissolution of these generated sulÐde minerals in intersti-tial waters.

3590 Phys. Chem. Chem. Phys., 2001, 3, 3586È3592

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online

Fig. 4 2D and 3D views of Raman images of molecular species and corresponding typical spectra. A : (anatase) ; B : (rutile) ; C :TiO2 TiO2kaolin ; D: zircon ; E : Ñuorescent unknown particle ; F : paper Ðlter.

To test the efficiency of the chemical extraction proceduresused in the present work for sedimentary pyrite recovery(which is hardly demonstrated by conventional analyticaltechniques because of the low contents of pyrite in our sedi-ment samples), micro-Raman analyses have been undertakenon solids recovered after the two acidic treatments : (i) 1 MHF digestion at room temperature for 1 day ; and (ii) concen-trated nitric acid at 100 ¡C for 12 h.

Micro-Raman spectroscopic studies

Micro-Raman spectroscopy has proved to be a valuable tech-nique for the identiÐcation of sedimentary components andtheir morphological aspects in sedimentary materials.45 In thiswork, the combined use of an optical microscope with aRaman spectrometer has allowed us to visually select analyti-cal areas where interesting mineral specimens are present, andsubsequently to perform both morphological and chemicalanalyses on randomly chosen specimens. This technique wasapplied to particles recovered after the chemical extractionstep employing HF, and subsequently after the chemicalextraction with concentrated Note that these particlesHNO3 .are deposited on cellulose-acetate Ðlters (Alltech) and analyzeddirectly on these Ðlters after drying on a horizontal laminar-Ñow hood (Class 100 clean air). Micro-Raman spectra of the

di†erent minerals detected on these Ðlters (i.e., zircon, rutile,anatase, kaolin, pyrite . . .) were recorded satisfactorily afteroptimization of the principal Raman instrumental parameters(see the Experimental). Micro-Raman spectroscopy alsopermits the elaboration of chemical mapping which is, in thiscase, particularly useful in identifying/visualizing mixed-phaseaggregates. It should be noted that such a microanalysis hasnevertheless a severe disadvantage compared to conventionalmicroanalysis because the chemical nature of these sedimen-tary solids is very heterogeneous and their chemical composi-tion then varies strongly from one analytical area to another,even after chemical extraction procedures. Hence, only limitedqualitative identiÐcation of the sedimentary phases can beobtained by Raman microprobe.

Micro-Raman point-by-point analysis of the randomlychosen specimens on the surface of sedimentary solids recov-ered after HF treatment (and after concentrated HNO3treatment) was performed. The surface areas analyzed are19 200 lm2 (Raman image number 1) and 5120 lm2 (Ramanimage number 2).

The decomposition of each experimental spectrum of theprobed zone into the combination of model spectra selected inthe set of the recorded spectra has enabled us to constructRaman mapping. The Raman spectra exhibit several charac-teristic Raman regions attributed to the minerals : zircon,

Phys. Chem. Chem. Phys., 2001, 3, 3586È3592 3591

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online

anatase, rutile, kaolin and pyrite in addition to that corre-sponding to the Ðlter (Fig. 3). This micro-Raman study con-Ðrms the selective elimination of pyrite during the lastsequential extraction step [using concentrated (Fig.HNO3]4). The solution resulting from attack has been usefulHNO3in the present work, Ðrst for the determination of pyritic tracemetal and Fe contents in anoxic sediments from the Authiebay and secondly for evaluating DTMP.

Conclusion

Quantitative determination of AVS and CRS and trace metalsin sedimentary pyrite has helped improve our knowledge onscavenging , mobility and diagenetic pathways of Fe, Mn, Ni,Zn, Pb, Cu and Cd in anoxic sediments from the Authie bay.In addition, from analytical data on porewaters, thermodyna-mic calculations have proved the occurrence of precipitationof iron sulÐdes, mainly in the forms of greigite and amorphousFeS, although mackinawite might exist in the bulk sediment.We have also found substantial evidence of association of Cd,Cu, and Zn with sulfur to give metal sulÐdes as discretephases, and of coprecipitation with and/or adsorption of Mnon iron sulÐdes. Direct analysis of sedimentary pyrite whichwas previously extracted by acidic treatments, has yieldedweak Cd, Ni, Cu and Zn concentrations (i.e., low DTMP) andsigniÐcant Mn concentrations (i.e., high DTMP) as a functionof sediment depth. The plots : DTMP vs. DOP obtained forthese metals indicate strong dependence of the Mn contentson the pyrite content, suggesting that manganese is extensivelypyritized. In contrast, because zinc, copper, nickel andcadmium hardly vary with the amounts, these four ele-FeS2ments are weakly pyritized and consequently are mainlypresent in the form of ZnS, CuS, NiS and CdS. On the otherhand, the micro-Raman spectroscopic technique has beenapplied successfully to the identiÐcation of sedimentary min-erals in heavy minerals extracted by density separation withheavy liquids, and in solids recovered after each sequentialleaching of sediments (with HCl, HF and This tech-HNO3).nique has proved the relative efficiency of these chemical treat-ments, Ðrstly, for an enrichment of sedimentary pyrite inrecovered solids, and secondly, for the total dissolution ofpyrite with concentrated Note that, in all cases, recov-HNO3 .ered pyrite is accompanied by numerous residual compoundssuch as rutile, zircon, . . . that are hardly decomposed by suchacidic treatments.

References1 R. Chester, Marine Geochemistry, Unwin Hyman, London, 1990,

p. 698.2 R. W. Howarth, J. W. B. Stewart and M. V. Ivanov, Sulfur

Cycling on the Continents : W etlands, T errestrial Ecosystems andAssociated W ater Bodies, Wiley, Chichester, 1992, vol. 48, p. 350.

3 W. Davison, Geochim. Cosmochim. Acta, 1980, 44, 803.4 R. B. Cook, Can. J. Fish. Aquat. Sci., 1984, 41, 286.5 D. R. Lovley, Geomicrobiol. J., 1987, 5, 375.6 D. E. CanÐeld, Geochim. Cosmochim. Acta, 1989, 53, 619.7 R. A. Berner, Am. J. Sci., 1970, 268, 1.8 D. Rickard, Am. J. Sci., 1975, 275, 636.9 G. W. Luther-III, Geochim. Cosmochim. Acta, 1991, 55, 2839.

10 M. A. A. Schoonen and H. L. Barnes, Geochim. Cosmochim. Acta,1991, 55, 1495.

11 M. A. A. Schoonen and H. L. Barnes, Geochim. Cosmochim. Acta,1991, 55, 1505.

12 M. A. A. Schoonen and H. L. Barnes, Geochim. Cosmochim. Acta,1991, 55, 1515.

13 M. A. Huerta-Diaz and J. W. Morse, Mar. Chem., 1990, 29, 119.14 M. A. Huerta-Diaz, R. Carignan and A. Tessier, Environ. Sci.

T echnol., 1993, 27, 2367.15 J. W. Morse and T. Arakaki, Geochim. Cosmochim. Acta, 1993,

57, 3635.16 C. Gagnon, A. Mucci and E. Pelletier, Geochim. Cosmochim. Acta,

1995, 59, 2663.17 W. G. Brumbaugh and J. W. Arms, Environ. T oxicol. Chem.,

1996, 15, 282.18 S. L. Simpson, S. C. Apte and G. E. Batley, Environ. Sci. T echnol.,

1998, 32, 620.19 M. A. Huerta-Diaz, A. Tessier and R. Carignan, Appl. Geochem.,

1998, 13, 213.20 J. W. Morse and G. W. Luther-III, Geochim. Cosmochim. Acta,

1999, 63, 3373.21 I. I. Volkov and L. S. Fomina, in T he Black Sea, Geology, Geo-

chemistry and Biology, ed. E. T. Degens, D. A. Ross, AmericanAssociation Petroleum Geologists Memoir, Tulsa, OK, 1974, vol.20, pp. 45È477.

22 R. Raiswell and J. Plant, Econ. Geol., 1980, 75, 684.23 F. I. Roberts, J. Geochem. Explor., 1982, 17, 49.24 G. E. Batley, T race Element Speciation : Analytical Methods and

Problems, CRC Press, Boca Raton, FL, 1990, p. 350.25 A. N. Afnor, in Eaux dÏEssais, ed. A. N. Afnor, Paris,Me� thodes

1990, p. 191.26 G. W. Luther-III, L imnol. Oceanogr., 1985, 30, 727.27 W. Davison, J. Buffle and R. Devitre, Pure Appl. Chem., 1988,

60(10), 1535.28 N. Batina, I. Ciglenecki and B. Cosovic, Anal. Chim. Acta, 1992,

267, 157.29 I. Ciglenecki and B. Cosovic, Electroanalysis, 1997, 9(10), 1.30 A. Bertolin, G. A. Mazzocchin, D. Rudello and P. Ugo, Mar.

Chem., 1997, 59, 127.31 I. Ciglenecki and B. Cosovic, Mar. Chem., 1996, 52, 87.32 G. Billon, B. Ouddane and A. Bougrhiet, T alanta, 2001, 53, 971.33 D. E. CanÐeld, R. Raiswell, J. T. Westrich, C. M. Reaves and R.

A. Berner, Chem. Geol., 1986, 54, 149.34 J. C. Cornwell and J. W. Morse, Mar. Chem., 1987, 22, 193.35 S. Sharonov, I. Nabiev, I. Chourpa, A. Feofanov, P. Valisa and

M. Manfait, J. Raman Spectrosc., 1994, 25, 699.36 S. Sharonov, I. Chourpa, P. Valisa, F. Fleury, A. Feovanov and

M. Manfait, Eur. Microsc. Anal., 1994, 23.37 W. D. Schecher and D. C. McAvoy, MINEQL ] : A Chemical

Equilibrium Modeling System, version 4 for Windows, Environ-mental Research Software, 1998.

38 R. M. Smith and A. E. Martell, Critical Stability Constants,Plenum Press, New York, 1981, vol. XIII, p. 257.

39 S. Kotrly and L. Sucha, in Handbook of Chemical Equilibria inAnalytical Chemistry, ed. R. A. Kottler and M. Masson, Wiley,Chichester, 1985, p. 118.

40 D. M. Di-Toro, J. D. Mahony, D. J. Hansen, K. J. Scott, M. B.Hicks, S. M. Mayr and M. S. Redmond, Environ. Sci. T echnol.,1990, 9, 1487.

41 D. M. Di-Toro, J. D. Mahony, D. J. Hansen, K. J. Scott, A. R.Carlson and G. T. Ankley, Environ. Sci. T echnol., 1992, 26, 96.

42 J. W. Morse, Mar. Chem., 1994, 46, 1.43 W. A. Kornicker and J. W. Morse, Geochim. Cosmochim. Acta,

1991, 55, 2159.44 T. Arakaki and J. W. Morse, Geochim. Cosmochim. Acta, 1993,

57, 9.45 A. Boughriet, B. Ouddane, C. Cordier and J. Laureyns, J. Chem.

Soc., Faraday T rans., 1998, 94, 3677.46 W. Stumm, Chemistry of the Solid-W ater Interface : Processes at

the Mineral-W ater and Particle-W ater Interface in NaturalSystems, Wiley-Interscience, New York, 1992, p. 428.

3592 Phys. Chem. Chem. Phys., 2001, 3, 3586È3592

Publ

ishe

d on

31

July

200

1. D

ownl

oade

d by

Uni

vers

ity o

f W

est F

lori

da o

n 27

/08/

2013

21:

59:2

2.

View Article Online