ebook img

Experimental studies of equilibrium iron isotope fractionation in ferric aquo–chloro complexes PDF

16 Pages·2009·0.64 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Experimental studies of equilibrium iron isotope fractionation in ferric aquo–chloro complexes

Available online at www.sciencedirect.com GeochimicaetCosmochimicaActa73(2009)2366–2381 www.elsevier.com/locate/gca Experimental studies of equilibrium iron isotope fractionation in ferric aquo–chloro complexes Pamela S. Hilla,*, Edwin A. Schaublea, Anat Shaharb, Eric Tonuia, Edward D. Younga,c aDepartmentofEarthandSpaceSciences,UniversityofCalifornia,LosAngeles,CA90095,USA bGeophyscialLaboratory,CarnegieInstitutionofWashington,WashingtonDC20015,USA cInstituteofGeophysicsandPlanetaryPhysics,UniversityofCalifornia,LosAngeles,CA90095,USA Received7July2008;acceptedinrevisedform22January2009;availableonline4February2009 Abstract Herewecomparenewexperimentalstudieswiththeoreticalpredictionsofequilibriumironisotopicfractionationamong aqueousferricchloridecomplexes(Fe(H O) 3+,FeCl(H O) 2+,FeCl (H O) +,FeCl (H O) ,andFeCl –),usingtheFe–Cl– 2 6 2 5 2 2 4 3 2 3 4 H Osystemasasimple,easily-modeledexampleofthelargervarietyofiron–ligandcompounds,suchaschlorides,sulfides, 2 simple organic acids, and siderophores. Isotopicfractionation (56Fe/54Fe) among naturally occuring iron-bearing species at Earthsurfacetemperatures(upto 3&)isusuallyattributedtoredoxeffectsintheenvironment.However,theoreticalmod- ! eling of reduced isotopic partition functions among iron-bearing species in solution also predicts fractionations of similar magnitudeduetonon-redoxchangesinspeciation(i.e.,ligandbondstrengthandcoordinationnumber).Inthepresentstudy, fractionationsaremeasuredinaseriesoflowpH([H+]=5M)solutionsofferricchloride(totalFe=0.0749mol/L)atchlo- rinities ranging from 0.5 to 5.0mol/L. Advantage is taken of the unique solubility of FeCl – in immiscible diethyl ether to 4 createaseparatespectatorphase,usedtomonitorchangingfractionationintheaqueoussolution.D56Fe =d56Fe(total aq-eth Feremaining inaqueous phase) d56Fe(FeCl – inetherphase) isdetermined foreach solutionvia MC-ICPMSanalysis. Bothexperimentsandtheoret"icalcalculatio4nsofD56Fe showadownwardtrendwithincreasingchlorinity:D56Fe aq-eth aq-eth isgreatestatlowchlorinity,whereFeCl (H O) +isthedominantspecies,andsmallestathighchlorinitywhereFeCl (H O) is 2 2 4 3 2 3 dominant.TheexperimentalD56Feaq-ethrangesfrom0.8&at[Cl–]=0.5Mto0.0&at[Cl–]=5.0M,adecreaseinaqueous– etherfractionationof0.8&.ThisisveryclosetothetheoreticallypredicteddecreasesinD56Feaq-eth,whichrangefrom1.0to 0.7&, dependingonthe abinitio model. Therateofisotopicexchangeandattainmentofequilibriumareshownusingspikedreversalexperimentsinconjunction withthe two-phase aqueous–ethersystem. Equilibrium underthe experimental conditionsis established within30min. Thegeneralagreementbetweentheoreticalpredictionsandexperimentalresultspointstosubstantialequilibriumisotopic fractionationamongaqueousferricchloridecomplexesandadecreasein56Fe/54FeastheCl–/Fe3+ionratioincreases.The effects on isotopic fractionation shown by the modeling of this simple iron–ligand system imply that ligands present in an aqueous environment are potentially important drivers of fractionation, are indicative of possible fractionation effects due tootherspeciationeffects(suchasiron–sulfidesystemsorironbondingwithorganicligands),andmustbeconsideredwhen interpreting ironisotopefractionation in the geologicalrecord. !2009Elsevier Ltd. Allrights reserved. 1. INTRODUCTION Isotopic fractionation among iron-bearing species is * Correspondingauthor. found in both geological and biological systems at Earth- E-mailaddress:[email protected](P.S.Hill). 0016-7037/$-seefrontmatter!2009ElsevierLtd.Allrightsreserved. doi:10.1016/j.gca.2009.01.016 FeisotopefractionationexperimentswithFe–Clcomplexes 2367 surface temperatures. Large fractionations have been to study the effects of solution chemistry and iron–ligand attributed to redox reactions between ferrous and ferric speciation on equilibrium mass-dependent iron isotopic species. Both theoretical studies (Polyakov and Mineev, fractionation. We measured the extent of Fe-isotope frac- 2000; Schauble et al., 2001; Anbar et al., 2005; Polyakov tionation among aqueous Fe–Cl complexes relative to a et al., 2007; Tang and Liu, 2007; Hill and Schauble, 2008) monospecific etherspectator phase,under arangeofchlo- andexperimentalstudies(e.g.,Anbaretal.,2000;Johnson ride concentrations (0.5–5.0mol/L) in low pH solutions, etal.,2002)havedemonstratedFe3+/Fe2+fractionationof andthencomparedourresultsagainstthetheoreticalequi- atleast 2–3&atroom temperature. librium mass-dependent isotopic fractionations predicted Theoretical studies also indicate that changes in bond byseveralsetsofabinitiomodelsdevelopedforthesecom- partners (i.e., different ligands or coordination numbers) plexes byHillandSchauble (2008). shouldresultinsignificantisotopicfractionation(Polyakov These studies demonstratea potentially significantcon- and Mineev, 2000; Schauble et al., 2001; Polyakov et al., tributiontoironisotopesignaturesinnaturalsystemsfrom 2007; Hill and Schauble, 2008). The omnipresence of dis- thenon-redoxeffectsofiron–ligandspeciation.Further,the solved aqueous species (e.g., chlorides, sulfides, organic li- techniques used in the present study, viz. separation of a gands, and siderophores) and their potential influence on spectator phase using diethyl ether and the use of spiked thesolubility,mobility,andavailabilityofiron-bearingspe- two-phase equilibrium experiments, can be used to make cies make the isotope geochemistry of aqueous iron com- reversible measurements of both Fe3+/Fe2+ fractionation plexesanimportant areaof study. andfractionationsassociatedwithsmall,labileinorganicli- Chloride,aubiquitousaqueoussolute,isofinterestasa gands analogousto biological molecules. These techniques bondingpartnerwithironbecauseofitssimilaritytoother shouldalsobeapplicabletomeasurementsofisotopicfrac- likelyiron–ligandpartners,anditsfacilityinaqueous-phase tionation of otherheavy elements in solution. experiments.Forexample,S2–,acommonpartnerofironin bothnaturalandbiologicalsystems,hasanelectronicstruc- 1.1.Speciationandstructuresoftheferricchloridecomplexes ture similar to that of Cl– and can form similar structures withsimilar56Fe/54Fetrends(Schaubleetal.,2001;Polya- According to several spectroscopic studies (e.g., Lind, kovetal.,2007).Ferricchloridesandsulfides/sulfatesoccur 1967; Cotton and Gibson, 1971; Sharma, 1974; Best naturallyinsomehighlyacidicextremeterrestrialenviron- et al., 1984; Murata and Irish, 1988; Murata et al., 1989), mentsiftheelementalcomponentsareavailable(e.g.,acidic themajorcomplexesinanacidicaqueoussolutionofferric sulfate–chloridelakesorhotspringsassociatedwithsubaer- chloride are Fe(H O) 3+, Fe(H O) Cl2+, Fe(H O) Cl +, 2 6 2 5 2 4 2 ial volcanic activity (e.g., Delmelle and Bernard, 1994), Fe(H O) Cl 0,andFeCl –(Table1).Followingtheconven- 2 3 3 4 acidic evaporative basins or lakes (e.g., Bowen et al., tioninHillandSchauble(2008),wehaveabbreviatedthese 2007), and highly acidic mine tailings such as Spain’s Rio complexes according to the number of Fe–Cl bonds (i.e., Tinto(e.g.,Amilsetal.,2007)ortheRichmondmine,Iron C0=Fe(H O) 3+, C1=(Fe(H O) Cl)2+, etc.). The com- 2 6 2 5 Mountain, California (e.g., Jamieson etal., 2005)). plexes C0, C1 and C2 have octahedral geometries, while The formation of Fe3+–Cl––H O complexes is expected C4isclearlytetrahedral(MaginiandRadnai,1979;Apted 2 to generate inter-species fractionations nearly as large as etal.,1985).ThestructureofferrictrichlorideC3islesscer- the 3& fractionation between Fe3+(aq) and Fe2+(aq) tain: it may be either octahedral or trigonal bipyramidal ! (Schauble et al., 2001; Hill and Schauble, 2008). However, (Lind, 1967; Bjerrum and Lukes, 1986) and was modeled the experimental record is mixed. In low-chlorinity experi- both ways by Hill and Schauble (2008). The octahedral ments([Cl–]60.1mol/L),noisotopiceffecthasbeenfound structuresofC2andC3weremodeledrespectivelyastrans (Welch et al., 2003); at higher chloride concentrations (i.e.,withCl–onoppositecornersoftheoctahedron)andas ( 2MHCl)inresin/solutionpartitioningexperimentsAn- fac(i.e.,withCl–ionsonthethreecornersofonefaceofthe ! bar et al. (2000) found a modest ( 0.1&) fractionation, octahedron).WeusedvaluesforC2astransandC3asfac ! while Fujii et al. (2006) found fractionations from 1.6 to octahedral whencalculating ourtheoretical predications. 0.6&betweenthedominantFe3+–Cl–speciesinanaqueous There is also evidence for the existence of further com- HCl solution and FeCl – in crown ether at [HCl]=1.6– plexes such as FeCl (H O)2– and FeCl 3– (e.g., Shamir, 4 5 2 6 3.5M. 1991; Inada and Funahashi, 1999); however, the relative Thisstudy integrates experimental studies withtheoret- abundances of these complexes in solution are too low to ical predictions in the aqueous ferric aquo–chloro system be considered in our experiments. Ferric hydroxide com- Table1 Aqueousferric–chloridecomplexes. FerricComplex Abbreviation(#Clatoms) #Fe–Clbonds Charge Structure Fe(HO) 3+ C0 0 +3 Octahedral 2 6 FeCl(HO) 2+ C1 1 +2 Octahedral 2 5 FeCl(HO)+ C2 2 +1 Octahedral(trans) 2 2 4 FeCl(HO)0 C3 3 0 Octahedral(fac) 3 2 3 FeCl– C4 4 1 Tetrahedral Fe(H4O) 3+NO– 0 +"2 Nitrateinouterhydrationsphere 2 6 # 3 2368 P.S.Hilletal./GeochimicaetCosmochimicaActa73(2009)2366–2381 plexes are also considered negligible due to the low pH of more C4 from the aqueous phase at low chlorinities than the solutions ( 0); a preliminary report by Liu and Tang does isopropyl ether (Dodson et al., 1936). Nachtrieb and ! (2006) suggests that fractionation of the ferric hydroxides Conway (1948)noted thatthe timerequired formaximum is similar to that of C0. Nitric acid (HNO ) is used in the extractionofironintotheetherdependsappreciablyupon 3 experiments to control the pH and ionic strength of the theamountofmixingoftheetherandtheaqueoussolution; solution while allowing for a range of chloride activities. morethan98%ofthetotalirondissolvedintheetherwas Nitrate is believed to interact weakly with the ferric ionin extracted within 20min after gentle inversions of the solu- the outer hydration sphere as Fe(H O) 3+.NO – (Horne tion container. We gave each of our ether–aqueous mix- 2 6 3 etal.,1964;MorrisandSturgess,1969);preliminaryabini- tures a vigorous shake and then let them equilibrate for tiomodelingbyHillandSchauble(2008)indicatesthatthe at least 40min before proceeding, so that the ether and reduced partition function ratio of Fe(H O) 3+.NO – aqueous phases would have a chance to separate after the 2 6 3 should besimilar to, orslightlysmaller than,that of C0. initial emulsification. Speciationmodelingoftheaqueousferricchloridecom- Thereisaslightdissolutionofhydrochloricacid(i.e.,the plexes (this work and others such as Rabinowitch and aqueous phase) into the ether phase upon addition of the Stockmayer, 1942; Bjerrum and Lukes, 1986; Bethke, ether.Dodsonetal.(1936)reportedadecreaseintheaque- 1996; Lee et al., 2003) indicates that the initial concentra- ous volume 62% with the addition of an equal amount of tion of chloride, the acidity, and the ionic strength of the ether. For a concentration of 0.0794M Fe aq, the 2% solution determine the relative concentration of each Fe– reduction in volume results in an increased Fe concentra- Cl complex. As the Cl–/ Fe3+ ratio increases, the relative tionequalto0.0810MFe,notenoughtoinfluencethelevel abundances of species with more Fe–Cl bonds increases. of accuracy of our experiments. We ran a control experi- Thedesignofourexperimentsmakesuseofthesechanging ment with [Cl–]=0.0M to ensure that no measurable relative abundances. amountofaqueousironwasextractedintotheetherphase without the presence ofHCl (seeSection 3). 1.2.Ether withdissolved FeCl –as spectatorphase 4 1.3. Abinitio models ofequilibrium mass-dependent Fe We took advantage of the selective solubility of isotope fractionation in Fe–Cl species C4 (FeCl –) in organic solvents at low pH to create a 4 separable spectator phase. The use of ether to extract Four different ab initio models of equilibrium mass- FeCl –fromasolutionofferricchloridedissolvedinhydro- dependent Fe isotopic fractionation among the ferric 4 chloricacidisanoldtechnique,developedbyJ.W.Rothein chloride complexes, (from Hill and Schauble, 2008), are thelate1800s(NachtriebandConway,1948).Becauseether used to generate theoretical predictions for the aque- andwaterareimmiscible,theethersolutioncontainingthe ous–ether fractionations resulting from our experiments. dissolvedC4formsaseparatephase,whichrestsontopof The models employ Unrestricted Hartree Fock (UHF) the aqueousphase. and hybrid Density Functional Theory (DFT), B3LYP A monospecific ether phase can easily be equilibrated (Becke, 1993), methods paired with different basis sets. withanaqueousferricchloridesolution,andquantitatively The UHF model is paired with the 6-31G(d) basis set separatedfromthatsolutionwithoutdisturbingtheoverall (Rassolov et al., 1998), and the hybrid DFT models, with equilibrium of the two-phase solution. This procedure is basis sets 6-31G(d), 6-311G(d) (Frisch et al., 2004), and analogoustotheuseofCO gastomonitortheoxygen-iso- Ahlrich’s VTZ (Scha¨fer et al., 1992). 6-31G(d) is a double 2 topefractionationbehaviorofaqueoussolutionsofvarying split-valence basis set; 6-311G(d), a triple split-valence ba- chemistries(e.g.,O’NeilandTruesdell,1991).Theexistence sis set; and Ahlrich’s VTZ, a triple zeta basis set. The ofatwo-phasesystemalsofacilitatesreversalexperiments, four models are abbreviated: (1) UHF/6-31G(d); (2) as described below. B3LYP/6-31G(d); (3) B3LYP/6-311G(d); and (4) Ether extraction of a spectator phase has advantages B3LYP/VTZ. The latter model uses Ahlrich’s VTZ for over precipitation methods (e.g., Johnson et al., 2002) or Fe and Cl, and basis set 6-31G(d) for O and H. The columnelutionmethods(e.g.,Anbaretal.,2000;Mathews models are in vacuo, i.e., modeled with only an inner et al., 2001). Once the system comes to equilibrium, one hydration sphere as gas phase molecules. neednotworryaboutkineticeffectsorfurtherfractionation The reduced partition function ratios (b) of the four duringthesamplingprocess.Conversely,bothprecipitation models for each complex are listed in Table 2 as 1000 lnb methods and column elution methods involve potential b =56Fe/54FereducedpartitionfunctionratioofAXrel- AX fractionation in order to separate the phases being com- ativetoadissociatedFeatomX,i.e.,ironvapor.1000lnb pared, the effects of which must be subtracted from the valuesare in parts permil (&). measured fractionation. The isotopic fractionations of each complex Ci relative Earlystudiesshowedthatthe amountof ironextracted to C4 (denoted as a ) are shown in Fig. 1 where (Ci C4) by the ether phase (most likely as FeCl .H O.nH O) in- 1000 lna =1000" lnb 1000 lnb (i.e., creasesastheconcentrationofironinthe4aci3dsolut2ionin- a b(Ci"/Cb4) ). The frac(tCioi)"nation facto(rC4)between (Ci C4)= (Ci) (C4) creasesandastheamountofHClinthesolutionincreases each"complex and C4 decreases as the number of Fe–Cl (Dodsonetal.,1936;Sandell,1944;NachtriebandFryxell, bonds increases (Table 3). The ranges of fractionation be- 1948). We chose diethyl ether overthe less volatile isopro- tween C0 and C4 predicted by the models are 1.5– ! pyletherasasolvent forC4becausediethyl etherextracts 2.5&. FeisotopefractionationexperimentswithFe–Clcomplexes 2369 Table2 Equilibrium Fe isotopic fractionation (1000 lnb) for the Fe–Cl complexesat293.15K(20"C)fromfouroftheabinitiomodelsof HillandSchauble(2008). Complex UHF/6- B3LYP/ B3LYP/ B3LYP/ 31G(d) 631G(d) 6311G(d) VTZ C0 9.73 8.93 9.41 9.59 C1 9.12 8.04 8.54 8.57 C2 8.73 7.61 8.1 8.13 C3a 7.74 7.05 7.13 7.13 C4 8.25 7.46 7.17 7.14 a C3modeledasFeCl(HO) (facoctahedralstructure). 3 2 3 Fig.2. Separationofetherandaqueousphasesinaferricchloride solution.Theetherphaseisonthetop. 2.1.4.). In each series the chlorinity was varied from 0.5 to 5.0M by the addition of HCl. Total acidity was held constant ([H+]=5mol/L), as well as the initial ionic strength ( 5 – see Section 3.2 on speciation), by adding ! appropriate amounts of HNO . All measurements were 3 madewithanEppendorfTMauto-pipette.Thesolutionswere gentlyshakenandthenallowedtorestfor10mininasep- aration funnel. 2.1.2. Etherphase Anequalvolumeofimmisciblediethyletherwasadded to the aqueous solution. The ether–aqueous mixture was briskly shaken, and then allowed to rest for at least 40min. The ether almost immediately formed a separate, lessdense,liquidphaseontopoftheaqueousphase.Dur- ingthistime,aportionofFeCl –passedfromtheaqueous Fig. 1. Predicted Fe isotopic fractionation of ferric chloride 4 complexes (Ci) relative to C4 expressed as 1000 lna where 1000 phaseintotheetherphase,turningtheetherphaseapale– lna (Ci C4)=1000 lnb (Ci)–1000 lnb (C4) (Hill and Schauble, yellow color. A sharp boundary between the aqueous and " 2008).(ComparewithTables2and3.) ether phases was visible almost immediately (Fig. 2). No second ether phase was seen; the formation of a second phase, mentioned by Dodson et al. (1936), appears to re- quirechlorinityP7M.SomeFeCl –isalsoexpectedtore- 4 Table3 mainin the aqueousphase. PredictedFractionationofC0RelativetoC4expressedas1000lna When the ether–aqueous mixture is shaken, there is a wSchhearueb1le0,02000ln8)a.(C(S0"eeC4F)i=g.110.0)0 ln b(C0)" 1000 lnb(C4) (Hill and tferonmdenthceytfoopraofvtehryessempaalrlaatimonoufnutnnoeflt,hdeumetixotuthreetinocersecaaspede UHF/ B3LYP/ B3LYP/ B3LYP/ vaporpressureoftheether.Lossofsolutionwasavoidedby 6-31G(d) 6-31G(d) 6-311G(d) VTZ gentlereleaseandthenreinsertionofthefunnelcap.Volu- metricethermeasurementsweremadewitha25mlgradu- 1000lna(C0 C4) 1.48 1.47 2.24 2.45 " atedglass cylinder. 2.1.3. Separationofphases 2.METHODS After the designated equilibrium time had elapsed in each experiment, separate aliquots of the ether (1ml) and 2.1.Fractionation experiment procedures aqueous(0.25ml)phaseswereremovedimmediatelybydis- posablepipettefromthetopofthefunnelandthroughthe 2.1.1.Aqueous phase stopcockatthebottomofthefunnel,respectively,toSavil- Series of aqueous ferric chloride solutions (20ml each) lexTM beakers for isotopic analysis. One milliliter of 1M were created from stock solutions of crystalline ferric ni- HCl was added to each ether sample to ensure dissolution trate Fe(NO ) .9(H O), ([Fe]=0.0794mol/L). The ferric ofallironastheetherevaporated.Thesamplesweredehy- 33 2 nitrate crystals were ultra pure (>99.9%) in order to avoid drated on a hot plate (100"C for aqueous samples; 50"C contamination by Cr in the MC-ICPMS (See Section forethersamples)toremoveanyremainingHCl,andthen 2370 P.S.Hilletal./GeochimicaetCosmochimicaActa73(2009)2366–2381 rehydrated with dilute nitric acid. Further dilutions were intensity of the color depends on the concentration of madewith2%nitricacidasnecessarytomatchtheconcen- Fe2+. The reducing agent hydroxylamine hydrochloride tration of the Fe isotope standard used for mass was added to the samples to reduce all ferric species. Ten spectrometry. timestherequiredamountsofreducingagentandofferro- Afewmillilitersofaqueoussolutionwereallowedtorun zinewereaddedtothesampletoensurethatallferriciron through the stopcock before samples were gathered for reacted. analysis. Theaqueoussample wasthengathered in a glass Maximum absorbance for iron occurs between pH val- beaker, from which measurements were made with an ues of 4 and 9 at a wavelength of 562nm (Stookey, auto-pipette. 1970).Weaddedanaceticacid–ammoniumhydroxidebuf- Thevolatilityofethercausesproblemswithanauto-pip- ferofpH 5.7tothesamplestoensurethecorrectrangeof ! ette.Therefore,measurementsoftheethersolutionwereta- pHvalues.Reducer,ferrozine,andbufferwerepremixedin ken with a disposable pipette with gradations at 0.25ml. a1:2:4ratioandthenaddedtothesamplestofacilitatethe We found that the ether in the disposable pipette has less measurement process. tendency to shoot out prematurely if the pipette is held Sincetheabsorbencyofsamplesdriftsslightlyovertime, vertically. each sample with added ferrozine, reducer, and buffer was allowedtorestauniformtenminutesbeforemeasurement 2.1.4.Isotopic measurements in the spectrophotometer. The Fe isotope ratios 56Fe/54Fe and 57Fe/54Fe of both Themolarabsorptivityconstantforourspectrophotom- theaqueousphaseandtheetherphaseforeachsamplewere eter,determinedempiricallybymeasurementofaliquotsof measured on the UCLA Thermo Finnegan Neptune MC- knownstandards,is26107±1441,inreasonableagreement ICPMS (Multiple Collector Inductively Coupled Plasma with measurements by other workers (e.g., 27,000 molar Mass Spectrometer). The instrument has a fixed array of absorbtivity, Stookey, 1970). Samples were diluted to nine Faraday collectors with 1011X amplifiers. The mass [Fe]= 40lM(absorbancy= 1)via2to3separatedilu- ! ! spectrometer was operated at a mass resolving power of tionsteps,usinganEppendorfTMauto-pipette.Thefirstdilu- 12,000 (M/dm) to eliminate ArO+, ArOH+, and ArN+ tion(s)werewithdeionizedwater;thelastwiththereducer– m!assinterferencesatcardinalmasses56,57and54,respec- ferrozine–buffermixture.Allaqueoussampleswerediluted tively(e.g.,WeyerandSchwieters,2003).Solutionswerein- and measured independentlyatleast threetimes. spected for 52Cr+ to monitor possible interference from Measuring the iron concentration in the ether phase 54Cr+.Inallcases,Crwasfoundtobebelowdetectionlev- posedpotentialproblemsbecausetheetherevaporatesrap- els. Potential instrumental mass bias due to inter-element idly, leaving behind an unknown volume of solution. To matrix effects was not an issue since samples consisted of avoidthisproblem,weremoved3mlofethersolutionfrom pureFe.Sampleswererunindryplasmausingadesolvat- the separation funnel to a 10ml graduated glass cylinder ingnebulizer(CetacAridus)at 1ppmFein 2%HNO . with a disposable pipette. We immediately added 1ml of Ironisotoperatioswerecorr!ectedforinstru!mentalfrac3- 1M HCl to the cylinder and then set the cylinder aside tionation using standard-sample-standard bracketing. Iso- for a few days, lightly covered with ParafilmTM, until all tope ratios are reported as d56Fe and d57Fe values relative the ether had evaporated. Deionized water was added to to astandard: the remaining liquid to return the total volume to 3ml. The resultant iron solution concentration, with ether re- iFe=54Fe moved and with a known dilution factor, was then mea- diFe¼103 ð iFe=54FÞeSample"1! ð1Þ sured with the spectrophotometer as described above. ð ÞStd Eachsamplewasdilutedandmeasuredthreetimesindepen- dently.However,measurementsofetherealironhavelarger where i represents either 56 or 57, Std represents the in- errors than aqueous iron measurements due to the addi- house standard, and Sample signifies the sample. In this tionalmeasurementerrorfromtheuseofbothadisposable study all Fe isotope ratios are reported relative to pipette and a graduated cylinder in the first steps of the the working standard Spex-1; the isotopic composition measurement process. All three sets of separate dilution of this standard is related to the IRMM14 measurements for the ethereal ironsample will include the iron standard by the experimentally determined d56Fe(IRMM14)=0.99979 (d56Fe(Spex-1)) 0.20845. same error resulting from the graduated cylinder and dis- " posablepipette.Therefore,allthreeetherealmeasurements IronisotopefractionationbetweenaqueousFeandethe- real FeCl – is expressed as d56Fe =d56Fe (total Fe willhavethesamecomponentofthisbiasintheirresultant inaqueou4sphase) d56Fe (FeC(laq–-eitnh)etherphaasqe).Values measurement errors. forD56Fe we"redeteertmhinedfo4raqueous–etherpairsof Two other measuring techniques for the ether phase eachmola(arqit-eytho)f Cl–. were also tried: (1) A known amount of ether phase was added immediately to a premeasured amount of reducer– buffer–ferrozine mix and covered with parafilm for 10min 2.1.5.Fe concentration measurements before going to the spectrophotometer. (2) A known The concentration of total iron in both the ether and amount of ether phase with added 1M HCl was put into aqueous phases for each sample was measured with a GENESYS20spectrophotometerusingtheferrozinemeth- od(Stookey,1970;Vollieretal.,2000).Ferrozineinteracts with ferrous iron to create a purple–hued complex; the 1 Standarderroronesigma;arelativestandarderrorof 0.5%. ! FeisotopefractionationexperimentswithFe–Clcomplexes 2371 aPetridishandpartiallycoveredwiththePetridishlidto sive 54FeCl –) was added to the ether of the normal - iso- 4 hasten evaporation of the ether (due to a larger surface tope forward experiment, creating the reversal area). Then an additional measured amount of 1M HCl experiment. Thus, the reversal experiment began at time was added to the dish and gently swirled to pick up any t=0 with 20ml aqueous phase and 30ml ether phase. remnants of Fe. However, measurements from these two We measured the rate and progress of the reversal experi- methodswerenotasreproducibleasthegraduatedcylinder ment toward equilibrium by removing small aliquots from method describedabove. both the ether and aqueous phases after 20, 30 and 40min (without disturbing the mixtures) and measuring 2.2.Reversal experiments the aqueous–etherisotopicfractionationateachtimestep. In order to quantify progress towards equilibrium, we Reversal experiments demonstrate equilibrium by run- usedthethree-isotopeexchangemethodtobracketequilib- ning the same reaction in opposite directions (forward rium(Matsuhisaetal.,1978;Mathewsetal.,1983;Shahar andreverse),therebybracketingtheequilibriumpointfrom etal.,2008).Thismethodisbasedontheexponentialrela- two different paths. In the forward experiment, the direc- tionshipbetweentheequilibriummass-dependentfraction- tion of iron transport is from the aqueous phase into the ation factors ab=a and ac=a between phases i and j for three i j i j ether phase as some FeCl –(aq) dissolves into the ether differentisotope"sa,b,an"dcofagivenelement(Matsuhisa 4 phase et al.,1978; Young etal.,2002).Inthiscase, 3 a56=54 a57=54 c 6 n 0FeClnðH2OÞ6"n3"nðaqÞþFeCl4"ðaqÞ c¼ð1¼=mð54"1Þ=m56Þ=ð1=m54"1=m57Þ ðð7ÞÞ X¼ 3 and 103ln a56=54 c103ln a57=54 8 )n 0FeClnðH2OÞ6"n3"nðaqÞþFeCl4"ðaqÞþFeCl4"ðethÞ where atomic ðmasseÞs¼mi areð: m54Þ=53.93961, m56ð=Þ X¼ 2 55.93494, and m57 =56.93540amu (Baum et al., 2002). ð Þ Using these values the exponent c=0.678. In d notation Thiscreatesanisotopicfractionationbetweentheaque- the above equilibrium relationship between the three iso- ousandetherphases,themagnitudeofwhichisdependent topes becomes uponthe chlorinity of thesolution. In the reversal experiment, the direction of iron trans- d56Fe ¼ ð1000 þ d56FerefÞ½ð1000 þ d57FeÞ portisfromtheetherphaseintotheaqueousphase.Ether = 1000 d57Fe c 1000 9 containingspikedC4(i.e.,54FeCl –)isaddedtotheetherof ð þ refÞ) " ð Þ 4 a forward, isotopically normal experiment, which has al- where ref refers to any point on the above fractionation ready achieved equilibrium. This forces isotopic transport line.Inthree-Fe-isotopespace(d56Fevs.d57Fe)Eq.(9)de- ofspikedC4fromtheetherphaseintotheaqueousphase, fines a slightly curved, primary (terrestrial) mass fraction- where speciation and isotopic exchange with the aqueous ation line, along which naturally occurring, iron-bearing iron occur. Further exchange will take place between the terrestrial samples lie (Matsuhisa et al., 1978; Young etherandtheaqueousphasesuntilanisotopicequilibrium et al.,2002). isreached. Atthestartofthereversalexperiment,whenthespiked etherisaddedtotheetheroftheunspikedforwardexperi- 3 ment,theresultingethermixtureliesofftheterrestrialmass 54FeCl"4ðethÞ) 54FeCl4"ðaqÞ) n 054FeClnðH2OÞ6"n3"nðaqÞ fractionation line (TFL)due to excess54Fe (Fig. 3).Isoto- þ54FeCl4"ðaqÞ X¼ ð3Þ pkiecdexacqhuaenoguesbepthwaeseentwhiellspeiskteadbleisthherepqhuailsiberaiunmdthaelounngspia- 3 54FeCl4"ðaqÞþn 056FeClnðH2OÞ6"n3"nðaqÞ()56FeCl4"ðaqÞ stheceonfidrsatry(Ffirga.ct3io).naTthioenploinseiti(oSnFLin),tpharerae-lliesolttoop,ebusptabceeloowf 3 X¼ the new SFL will pass through the bulk composition of þn 054FeClnðH2OÞ6"n3"nðaqÞ ð4Þ tehteal.m,1ix9t8u3r)e.,Thsiegnloificaedtionasofdt5h6Feeb0ulakncdomdp57oFsiet0ion(Minat3h-eiswos- X¼ 54FeCl"4ðethÞ þ56FeCl"4ðaqÞ()56FeCl"4ðethÞ tope spaceisdefined byisotopic mass balancesuch that þ54FeCl"4ðaqÞ ð5Þ ð56Fe=54FeÞ0¼ð56Fe=54Þaq*ðfÞ* ½Feaq)=½Fe0) Ifequilibriumisattainedintheseexperiments,theresul- þð56Fe=54FeÞeth*ð1"fÞ!*ð½Feeth)=½F"e0)Þ ð10Þ tant fractionation between the aqueous and ether phases and Fe f Fe 1 f Fe 11 shouldbethesameastheinterphasefractionationofafor- ½ 0)¼ *½ aq)þð " Þ*½ eth) ð Þ wardexperimentofthesamechlorinity,acidityandnitrate Note 56Fe=54Fe ; 56Fe=54 ,and 56Fe=54Fe arethe ð Þ0 ð Þaq ð Þeth concentration. isotopicratiosofthebulkcompositiontheaqueousphase, TheetherwithspikedC4forthereversalexperimentwas and the ether phase respectively; f is the fractional volume prepared via a forward experiment prepared with aqueous oftheaqueousphasetothetotalvolume;and[Feaq],[Feeth], 54Fe spike. Both the spiked and normal solutions had and[Fe0]aretheconcentrationsofFeintheaqueousphase, [Cl–]=1.0M. Once each mixture equilibrated, half the the ether phase, and the entire solution in moles/liter ether(10ml)fromthespikedexperiment(containingexces- (Faure, 2005). 2372 P.S.Hilletal./GeochimicaetCosmochimicaActa73(2009)2366–2381 of d56Fe) on the SFL should ideally be equal to the D56Fe of theforward experiment. aq eth Equi"librium is thus demonstrated in two ways: (1) Fe isotopic composition in both the ether and the aqueous phases of the reversal experiment moves to the same mass fractionation line, and 2) the D56Fe values of both (aq eth) theforwardandthereversalexperiment"sarethesamewith- in analytical uncertainties. Fe concentrations of each phase of the reversal experi- mentweremeasuredat40min(inordertominimizedisrup- tionofthereversalexperiment)andareusedincalculating the bulk isotopic composition ofthe mixture. 3. RESULTS 3.1. Dissolution ofFeCl –intothe ether phase 4 Fig.3. Schematicdiagramofprogressionofareversalexperiment toward equilibrium in three-isotope space. At time=0, the Thepercentageoftotalironextractedfromtheaqueous aqueousphase(toprightbluehexagon)sitsontheterrestrialmass phase by dissolution into the diethyl ether phase increases fractionation line (TFL) while the ether phase, with excess as the chlorinity of the solution increases, changing the 54FeCl–, rests below the TFL (red triangle in lower left corner). 4 ether from a pale yellow to a deep amber color. In our Asthemixtureequilibratesisotopically,bothphasesmovetoward experiments,inwhichbothpHandionicstrengthareheld equilibriumonthesecondarymassfractionationline(SFL)(lower roughly constant ( 0 and 5, respectively), the ethereal line), each along its own mixing line. The ether phase moves up ! ! along the lower (red) line while the aqueous phase moves down iron climbs from 0.0% at 0.0 M Cl" to >95% at 5.0M alongtheupper(blue)lineinthedirectionofthearrows.TheSFL (Table 4and Fig. 4). mustpassthroughthebulkcompositionofthetwophases,shown Whenacidity,ionicstrength,andchlorinityincreaseto- bythegreendiamond.Equilibriumisachievedwhenbothphases gether, as a function of increasing concentrations of HCl, lieontheSFL.Thedistancebetweenthetwophasesonceonthe the amount of initial iron dissolved into diethyl ether in- SFL(intermsofd56Fe)isD56Feaq eth.(Forinterpretationofthe creasesmoreslowly(Dodsonetal.,1936).Apparentlydis- rthefeerweenbcevsetrosiocnolooufrtshiisnptahpeefir.g)urele"gend,thereaderisreferredto solutionofFeCl4"increasesasbothmoreH+andCl"are madeavailable because theironismost likelyextracted as An analogous relationship holds for 57Fe. d56Fe and stoichiometric HFeCl4" (i.e., FeCl4.H3O.nH2O see section o d57Fe are thencalculatedfrom Eq.(1). 1.2).Dodsonetal.’setherealironconcentrationsareshown o in Fig. 4 for comparison. Asisotopicexchangeprogressestowardequilibrium,the Mass balance calculations of aqueous and ether iron etherphasewillmoveupamixinglinetowardapositionon concentration measurements are on average within 8% of the SFL while the aqueous phase will travel down its own thetotalinitialiron(comparablewiththetotaluncertainty mixing line toward itsposition on the same mass fraction- obtainedbypropagationofmeasurementerror–seeTable ationline(Fig.3).Equilibriumisreachedwhenbothphases 4). This is similar to mass balance molarity measurements arrive at the new equilibrium mass fractionation line (i.e., by Nachtrieb and Conway (1948). Measurement errors of the SFL). The distance between the two phases (in terms Table4 Divisionoftotalironbetweenaqueousandetherphasesfordifferentchlorinities.Eachconcentrationwasmeasuredindependentlyatleast threetimes.Reportederrorsarestandarderror1randaretheoriginalmeasurementerrorspropagatedthrough2–4separatedilutionsteps. Unitsaremol/liter.TotalFeisthecalculatedconcentrationoftheoriginalironsolutionbeforetheadditionofether.FeaqandFeethare measuredquantities.Feaqerrorsresultfrommeasurementwithanauto-pipettewhileFeetherrorscontainmeasurementserrorsfroman auto-pipette,adisposablepipette,anda10mlgraduatedcylinder(seeSection2.1.5).(SeeFig.4.) Cl– TotalFe Feaq Feeth Feaq+Feeth Feeth/(Feaq+Feeth) Feaq/(Feaq+Feeth) 0.0 0.0794 0.079±0.0003 0.000±0.0000 0.079±0.0003 0.0000 1.0000 0.5 0.0794 0.078±0.0003 0.003±0.0001 0.080±0.0003 0.0359 0.9641 0.5 0.0794 0.073±0.0007 0.001±0.0000 0.074±0.0007 0.0129 0.9871 1 0.0794 0.085±0.0001 0.005±0.0001 0.090±0.0001 0.0546 0.9454 1 0.0794 0.068±0.0011 0.008±0.0002 0.075±0.0011 0.1032 0.8968 1.5 0.0794 0.057±0.0000 0.029±0.0008 0.086±0.0008 0.3386 0.6614 1.75 0.0794 0.042±0.0001 0.035±0.0008 0.076±0.0008 0.4524 0.5476 3 0.0794 0.018±0.0000 0.074±0.0018 0.092±0.0018 0.8017 0.1983 3 0.0794 0.019±0.0000 0.066±0.0016 0.085±0.0016 0.7742 0.2258 3 0.0794 0.019±0.0002 0.065±0.0039 0.084±0.0039 0.7748 0.2252 4 0.0794 0.012±0.0001 0.073±0.0019 0.085±0.0019 0.8616 0.1384 5 0.0794 0.004±0.0002 0.086±0.0023 0.090±0.0023 0.9513 0.0487 5 0.0794 0.004±0.0001 0.079±0.0042 0.083±0.0042 0.9555 0.0445 FeisotopefractionationexperimentswithFe–Clcomplexes 2373 etherealironconcentrationingeneralarelargerthanthose ofaqueousironduetothevolatilityoftheetherandtodif- ferences in the measurement process (see Section 2.1.5 for further error discussion). Measurements may also be slightlyaffectedbyminutechangeinvolumeduetothedis- solution of some HCl (i.e., the aqueous phase) into the etherphase (see Section1.2). 3.2. Speciationof FeClcomplexes with varyingchlorinities The relative abundances of the aqueous ferric chloride complexes depend on the amount of available chloride, theacidity,andtheionicstrengthoftheaqueoussolution. Asthechlorinitytoironratioincreases,Fe–Clspecieswith Fig.4. Percentageoftotalinitialironextractedintoetherphasefor more Fe3+–Cl" bonds become more plentiful (Fig. 5 and Table 5). According to our speciation models, with pH variouschlorinities.Theamountofetherealironfromthisworkis 0andionicstrength 5,C2isthemostabundantcomplex comparedtoexperimentsfromDodsonetal.(1936)inwhichboth ! ! [H+] and [Cl–] were increased together (as [HCl]); in our experi- between0.5and3Mchlorinity,whileC3ismostabundant ments[H+]=5Mforall[Cl–].ThegreaterdissolutionofFeinto from 3.2 to 5M chlorinities. C4 abundances are initially ! theetherphaseatlowerchlorinitiesinourexperimentsshowsthat low but increase continuously with rising chlorinity. C0 [H+]isalimitingfactorsincetheetherealironisprobablydissolved and Fe(H2O)63+.NO3" and, to a lesser extent, C1, are asstoichiometricHFeCl4(seeSection1.2).(SeeTable4.) important at [Cl"]=.5–1M,but decease rapidly at higher [Cl"]. Our models for the ferric chloride complexes at each chlorideconcentrationwerecalculatedusingtheFeClasso- ciation equilibrium constants for C1, C2, C3 and C4 (30, 4.5,0.15and0.0078respectively)fromBjerrumandLukes (1986), the FeNO 2+ association equilibrium constant (10) 3 from Delany and Lundeen (1990), and the activity models forHClandHNO fromBromley(1973).Calculationswere 3 iterated to determine final ionic strength and species con- centrations.Ateachchlorinity,theamountofaqueousiron (i.e.,theamountofironleftintheaqueousphaseafterdis- solution of some iron into the ether phase – see Table 4) was used as the total aqueous iron concentration. Since FeCl4" is dissolved into the ether phase as stoichiometric HFeCl (FeCl .H O+.nH O see Section 1.2), the initial 4 4 3 2 aqueous chloride concentration was decreased by four times the amount of ethereal iron while the initial aqueous H+ concentration was decreased by the amount of ethereal iron. (For example, at [Cl"]=3M and [Fe ]=0.0794M, [Fe ]=0.017M so the amounts of total aq Fe3+, Cl" and H+ sequestered from the aqueous to the aFvqaigruy.eion5ug.s[MCploh–l]ae,sreefrlaaaftctitevireontdosistsohofelutttohioteanlinaomdfiovsiuodnmutaeolfFFiereCo–nlC–rleimcnotamoinpitnlhegexeines,ththaeert [eHtheethr+]p=ha0se.06aMre.[FTeheuths],=[C0.l0aq6"M]=, 2[C.9l7et6h"M]=a0n.d024[HMaq+a]n=d 4 4.94M). These small changes in aqueous concentrations phase.(SeeTable5.) Table5 CalculatedmolefractionsofeachFe–Clcomplex,forvaryingchlorinities,relativetothetotalmolesofaqueousiron,afterequilibrationwith the ether phase. Ave. Fe aq is calculated from the average percentage of Fe aq (Table 4, column 7) based on an initial total calculated Fe=.0794M.(SeeFig.5.) AveFeaq(M) Cl" Fe(H2O)63+ FeCl(H2O)52+ FeCl2(H2O)4+ FeCl3(H2O)3 FeCl4– Fe(H2O)6#NO32+ 0.077 0.5 0.005 0.130 0.543 0.076 0.001 0.246 0.073 1.0 0.001 0.074 0.664 0.199 0.003 0.058 0.053 1.5 0.001 0.047 0.636 0.289 0.007 0.021 0.043 1.75 0.000 0.038 0.612 0.327 0.009 0.014 0.017 3.0 0.000 0.017 0.492 0.465 0.023 0.002 0.011 4.0 0.000 0.011 0.417 0.536 0.036 0.001 0.004 5.0 0.000 0.007 0.359 0.584 0.049 0.000 2374 P.S.Hilletal./GeochimicaetCosmochimicaActa73(2009)2366–2381 of Fe, Cl and H did not have a noticeable impact on speciation. Somespeciationmodelsofferricchloridecomplexesvs. chlorinity are shown in which chlorinity is increased by adding increasing amounts of HCl (e.g., Roe et al., 2003; Fujiietal.,2006).Inthesemodels,incontrasttoourmodel, pH decreases and the ionic strength increases in the aque- ous solution as chlorinity increases, resulting in a slightly differentdistributionofferricchloridecomplexabundances (discussedfurther in Section4.5). The presence of nitrate ions in the aqueous solution at lowchlorinitiesinfluencesthespeciationofironbycausing formationofFe(H2O)63+#NO3"inadditiontoFe(H2O)63+. In our experiments the NO3" ion is most abundant at the verylowestchlorinities,sincetheamountofnitricacidnec- Fig. 6. Experimental data for D56Feaq eth=d56Fe (total Fe essary to keep the pH constant decreases as more HCl is remaining in aqueous phase)"d56Fe (FeCl"4– in ether phase) (&). (pH 0,T=20"C).(SeeTable6.) added. ! Complexation of the nitrate ion with some of the iron sequesterspart of the ironthat mightotherwise have gone intoironchloridespecies,butthisapparentlyhasnosignif- icant effect on the relative abundances of the non-nitrate complexes northe dissolution of Feinto the ether. 3.3.Experimental fractionation betweenaqueousandether phaseswith increasing chlorinity The measured fractionation between the aqueous and ether phases, defined as D56Fe =d56Fe (total Fe (aq eth) remaining in aqueous phase)"d5"6Fe (FeCl4" in ether phase), shows a clear downward trend with the increasing chlorinity (Table 6 and Fig. 6). At [Cl"]=0.5M, D56Feaq eth=0.8&; at 5.0M, 0.0&. The d56Fe and d57Fe experim"entalvaluesplotalongtheterrestrialmassfraction- Fig.7. Experimentalvaluesofd56Fe(&)andd57Fe(&)pairsfrom ationline,ademonstrationofmassdependentfractionation D56Feaq eth (&) plotted along the terrestrial mass fractionation anda lackof Crinterference (Section 2.1.4)(Fig. 7). line. Err"or bars represent 1-sigma standard error. The error bars IsotopicFemassbalancecalculationsonaqueous–ether aresmallerthantheplottedsymbolsonthegraph.(SeeTable6.) pairswerewithin±0.05&oftheinitialferricnitratestock D56Fe 1000 lnb X solution. aq"eth ¼ ð FeCliÞð iaqÞ i X 1000 lnb 12 3.4.Predicted D56Fe fromthe abinitio fractionation " FeCl"4ether ð Þ aq-eth models whereXiisthefractionofFeincomplexi(relativetototal Feintheaqueoussolution)andb isthereducedparti- PredictionsofD56Fe usingthefourabinitiomod- tion function ratio for the ith FeCFleCcolimplex. This assumes aq eth els werecalculated as " there is no fractionation between C4aq and C4eth when C4 Table6 Measuredfractionation(in&)betweenaqueousandetherphases,definedasD56Fe(aq eth)=d56Fe(totalFeremaininginaqueousphase)– d56Fe (FeCl4– in ether phase), versus chlorinity. d56Fe and d57Fe for both ether and a"queous phasesare shown, expressedrelative to the inhousestandardFeSpex-1(seeSection2.1.4).Standarderrorsarefor1sigma.Theetherresttimeisthetimeinminutesbetweentheaddition ofethertotheaqueoussolutionandtheremovalofaqueousandethersamplesforisotopicmeasurement.(SeeFig.6.) Exp’t Cl D56Fe(aq eth) Aqd56Fe Etherd56Fe Aqd57Fe Etherd57Fe Etherresttime(min) " FN30 0.5 0.817±0.011 0.611±0.013 0.206±0.007 0.882±0.020 0.323±0.023 41 " " FN23 0.5 0.793±0.011 0.509±0.011 0.284±0.016 0.777±0.017 0.402±0.023 45 " " FN31 1 0.614±0.010 0.614±0.003 0.000±0.011 0.909±0.017 0.005±0.024 45 " FN20 1 0.540±0.012 0.644±0.016 0.104±0.010 0.978±0.015 0.130±0.016 42 FM32 3 0.186±0.016 0.717±0.004 0.531±0.017 1.051±0.010 0.747±0.013 41 FN21 3 0.365±0.008 0.868±0.011 0.503±0.008 1.280±0.017 0.684±0.008 42 FN33 5 0.014±0.017 0.679±0.018 0.666±0.006 0.956±0.015 0.918±0.016 45 FN22 5 0.011±0.003 0.553±0.011 0.542±0.011 0.837±0.015 0.836±0.011 45 FeisotopefractionationexperimentswithFe–Clcomplexes 2375 the forward and the reversal experiments achieved similar values of D56Feaq-eth: 0.828±0.0415& (1 sigma std err) and 0.783±0.0249&, respectively (Table 8 and Fig. 9). Second, both the aqueous and ether phases of the reversal experimenthavearrivedatthesameequilibriummassfrac- tionationline(SFL–seesection2.2)inthree-isotopespace in time t, where 20<t630min. On a plot of d57Fe vs. d56Fe, at the start of the reversal experiment (t=0), the unspiked aqueous phase and the spiked ether phase are on different mass fractionation lines. At t=20min, both phases have moved much closer; the ether phase has reached 99.8%2 of its equilibrium d56Fe , D56Fe at eth aq eth 20minisat97.4%2ofitsequilibriumvalue,whilethe"aque- Fig.8. Comparisonofabinitiomodelpredictionsandexperimental ousphasehasachieved71%2ofitsfinald56Feat20min(see dataforD56Feaq ethvs.chlorinity.Thesolid(red)lineisalogarithmic Section4.4).At30min,bothphasesareonthesameSFL. bestfittotheex"perimentaldata;bestfitequationsforexperimental Thelocationofthebulkcompositionoftheequilibrated dataandmodelpredictionsarelistedinTable9.(SeeTables6and7.) mixturein3-isotopespace,ascalculatedfromthemeasure- (Forinterpretationofthereferencestocoloursinthefigurelegend, ments of Fe concentrations and delta values at t=40min thereaderisreferredtothewebversionofthispaper.) (Table 8), also lies on the SFL as it should (see Section 2.2). (d57Fe, d56Fe (bulk composition))=( 3.50±.02&, " Table7 3.70±.02&, std err1 sigma) (Fig. 9). PredictionsofD56Feaq eth(&)foreachchlorinity,calculatedfrom " the fractional concen"trations of each Fe–Cl complex and the 4. DISCUSSION correspondingabinitio1000lnbfactors.Thepredicteddifferences inD56Feaq eth(&)for[Cl–]=0.5and5.0Mareshowninthelast 4.1. Fractionation amongtheFe–Cl complexes row.(SeeF"ig.8.) Cl"(M) UHF/ B3LYP/ B3LYP/ B3LYP/ Themostsignificantfeatureofboththetheoreticaland 6-31G(d) 6-31G(d) 6-311G(d) VTZ theexperimentalD56Fe dataistheprominentsmooth aq eth 0.5 0.71 0.49 1.24 1.34 decreaseinaqueous–ether"isotopicfractionationwithrising 1.0 0.37 0.15 0.84 0.91 chlorinity(Fig.8).Experimentsshowadecreaseinfraction- 1.5 0.23 0.04 0.69 0.75 ationfrom0.8&at0.5MClto0.0&for5.0MCl.Thethe- 1.8 0.18 0.00 0.64 0.69 oretical models predict a decrease of 0.7 to 1.0 &, 3.0 0.02 0.10 0.47 0.51 " dependingonthedetailsoftheabinitiomethod(Table7). 4.0 0.06 0.15 0.38 0.42 " " Thecloseagreementoftheexperimentalresultswiththe- 5.0 0.12 0.18 0.32 0.36 D56Feaq eth(Cl= "0.82 "0.68 0.92 0.98 oreticalpredictionsstronglysupportsthepresenceofnota- 0.5M") D56 ble fractionation resulting from speciation and/or Fe ("Cl=5.0M) complexing in the ferric iron–chloride–water system. This aq eth " is significant for two reasons: (1) it is of similar order of magnitude to Fe3+/Fe2+ fractionation; and (2) it implies thatligandspresentinanaqueousenvironmentarepoten- dissolves into the ether. All four ab initio models and the tially important drivers of fractionation, and must also be experimentalresultsdisplayasimilardecreaseintheaque- consideredwheninterpretingironfractionationingeologi- ous–ether fractionation with increasing chlorinity (Fig. 8, cal records. Tables 6 and 7). Eq. (12) is derived from the following Thereductioninaqueous–etherphasefractionationwith relationships. increasing chlorinity is most simply explained by the D56Fe d56Fe d56Fe 13 changes in aqueoussolution chemistry.As chlorinity rises, aq"eth ¼ AX " BX ð Þ I000lna d56Fe d56Fe 14 species with higher Cl/Fe ratios become more abundant, ðAX"BXÞ! AX " BX ð Þ thus decreasing the overall isotopic fractionation of the b aðAX"BXÞ ¼ bABXX"XX ð15Þ a(sqeueeoFuisgsp.ha1searnedlat5iv).e tTohiCs4t(rFenedCl4t"o)wianrdthereedtuhcetriopnhaosef " 1000 ln aðAX"BXÞ ¼ 1000 lnbAX"X "1000 lnbBX"X ð16Þ 56Fe/54Feoftheaqueoussolutionduetosolutionchemistry should also be apparent in fractionation between the solu- Preliminarycalculationsindicatethatthereducedparti- tion function ratio for Fe(H O) 3+.NO – is similar to that tion and any associated minerals (see Hill and Schauble, ofC0;hence,Fe(H O) 3+.NO2 –6wastre3atedasC0(seeHill 2008). 2 6 3 Acomparison ofthetheoretical modelswiththe exper- andSchauble, 2008). imentaldatashowsthattheyallhavethesamequasi-loga- 3.5.Reversal experiments 2 We define the % equilibrium value of X=100 (X @20min- demTohnestartattaeidnminenttwoof weqauyisli.bFriiurmst,waitthtihnirtthyirmtyinmutinesutbesotihs DX5@6F0meianq)/(etXh.@30min"X@0min) where X is d56Feeth, d56Feaq, or "

Description:
the solutions (!0); a preliminary report by Liu and Tang tope fractionation behavior of aqueous solutions of varying chemistries (e.g., O’Neil and Truesdell, 1991).
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.