Geobiology. 2017;15:353–365. wileyonlinelibrary.com/journal/gbi | 353

Received:25June2016  |  Accepted:16December2016DOI:10.1111/gbi.12227

O R I G I N A L A R T I C L E

Sulphur cycling in a Neoarchaean microbial mat

N. R. Meyer1 | A. L. Zerkle1 | D. A. Fike2

1SchoolofEarthandEnvironmentalSciences,UniversityofStAndrews,StAndrews,UK2DepartmentofEarthandPlanetarySciences,WashingtonUniversity,St.Louis,MO,USA

CorrespondenceN.R.Meyer,DepartmentofEarthSystemScience,StanfordUniversity,Stanford,CA,USA.Email:nrmeyer@stanford.eduandA.L.Zerkle,SchoolofEarthandEnvironmentalSciences,UniversityofStAndrews,StAndrews,UK.Email:az29@st-andrews.ac.uk

Current addressN.R.Meyer,DepartmentofEarthSystemScience,StanfordUniversity,Stanford,CA94305,USA

Funding informationGeologicalSocietyofLondon’sAlan&CharlotteWelchFund;APalaeontologicalAssociationUndergraduateResearch,Grant/AwardNumber:PA-UB201504;UniversityofStAndrew’sZannahStephenMemorialTravelScholarship;NaturalEnvironmentResearchCouncilFellowship,Grant/AwardNumber:NE/H016805/1;NaturalEnvironmentResearchCouncilStandard,Grant/AwardNumber:NE/J023485/2;NSF/EAR,Grant/AwardNumber:1229370;PackardFellowship

AbstractMultiplesulphur(S)isotoperatiosarepowerfulproxiestounderstandthecomplexityofSbiogeochemicalcyclingthroughDeepTime.Thedisappearanceofasulphurmass-independentfractionation(S-MIF)signal inrocks<~2.4Gahasbeenusedtodateadramaticriseinatmosphericoxygenlevels.However,intricaciesoftheS-cyclebeforetheGreatOxidationEventremainpoorlyunderstood.Forexample,theisotopecom-positionofcoevalatmosphericallyderivedsulphurspeciesisstilldebated.Furthermore,variationinArchaeanpyriteδ34Svalueshasbeenwidelyattributedtomicrobialsul-phate reduction (MSR).While petrographic evidence forArchaean early-diageneticpyrite formation iscommon, texturalevidence for thepresenceanddistributionofMSR remains enigmatic. We combined detailed petrographic and in situ, high-resolutionmultipleS-isotopestudies(δ34S and Δ33S)usingsecondaryionmassspec-trometry(SIMS)todocumenttheS-isotopesignaturesofexceptionallywell-preserved,pyritisedmicrobialites in shales from the ~2.65-Ga Lokammona Formation, GhaapGroup,SouthAfrica.ThepresenceofMSRinthisNeoarchaeanmicrobialmatissup-ported by typical biogenic textures including wavy crinkled laminae, and early-diagenetic pyrite containing <26‰ μm-scale variations in δ34S and Δ33S=−0.21±0.65‰(±1σ).These largevariations inδ34Svalues suggestRayleighdistillationofalimitedsulphatepoolduringhighratesofMSR.Furthermore,weiden-tifiedasecond,morphologicallydistinctpyritephasethatprecipitatedafterlithifica-tion,withδ34S=8.36±1.16‰andΔ33S=5.54±1.53‰(±1σ).WeproposethattheS-MIFsignatureof thissecondarypyritedoesnot reflectcontemporaneousatmos-phericprocessesatthetimeofdeposition; instead, itformedbytheinfluxof later-stagesulphur-bearingfluidscontaininganinheritedatmosphericS-MIFsignaland/orfrommagnetic isotopeeffectsduringthermochemicalsulphatereduction.These in-sightshighlightthecomplementarynatureofpetrographyandSIMSstudiestoresolvemultigenerationalpyriteformationpathwaysinthegeologicalrecord.

1  | INTRODUCTION

The sulphur isotope record has played an integral role in shapingour understanding of key events in Earth’s geological and biolog-ical history. Surficial S-cycling principally involves biological andabiotic mass-dependent fractionation (MDF) processes, which

followathermodynamicallydetermined,linearδ33S/δ34Srelationship(δ33S = 0.515 × δ34S).However, the geological S-isotope record alsocaptures evidence of mass-independent fractionation (MIF), whereδ33S and δ34Sdeviatefromthepredictedterrestrialmassfractionationline, quantified by the capital-delta (Δ) notation. Prior to theGreatOxidation Event (GOE) at ~2.4Ga, S-bearing minerals show large

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium,providedtheoriginalworkisproperlycited.©2017TheAuthorsGeobiologyPublishedbyJohnWiley&SonsLtd.

354  |     MEYER Et al.

positiveandnegativeΔ33Svalues.After~2.4Ga,Δ33S ratiosdimin-ishtowardsvaluesthattightlyclusteraroundzero(Δ33S=0±0.2‰).Archaean sulphur MIF has been widely attributed to atmosphericphotochemical reactions involving SO2; these reactionswould havebeenblockedinthePalaeoproterozoicduetotheUV-shieldingeffectscaused by increased atmospheric O2 and O3 concentrations (e.g.,Farquhar,Bao,&Thiemens,2000).Inaddition,thedeliveryofS-MIFtotheEarth’ssurfacerequiressulphurtoleavetheatmosphereviamul-tipleexitchannelsatdifferentredoxstates,whicharehomogenisedwhenatmosphericoxygenexceeds10−5ofpresentatmosphericlev-els (Pavlov&Kasting, 2002).Therefore,measuringmultiple sulphurisotopes(δ34S,Δ33S)inArchaeansedimentscanprovideinformationonbothatmosphericchemistryandbiogeochemicalsulphurcyclinginArchaeanpalaeoenvironments.

ThebiogeochemicalcyclingofsulphurintheArchaeanwasfunda-mentallydifferenttothepresent-daycycle.Thelargestfluxofsulphurintothemodernoceansisriverinesulphate,derivedfromtheoxidativeweatheringofpyrite.However,thisfluxwaslesssignificantbeforetheGOEduetolowatmosphericpO2;therefore,thetwomostsignificantSfluxesintotheArchaeanoceanswerelikelyhydrothermallysourcedand atmospherically derived sulphur species (e.g., Fike, Bradley, &Rose,2015).TheArchaeanS-MIFsignal isthoughttobedominatedbyphotochemicalreactionsinvolvingatmosphericSO2(e.g.,Farquharetal.,2000;Pavlov&Kasting,2002).Incontrast,theshieldingeffectofatmosphericoxygenandozonepreventsthecreationofanS-MIFsignalinthecontemporaryS-cycle,withtheexceptionofphotochem-icalreactionsgeneratingstratosphericsulphateaerosolsfromvolcaniceruptions(Baroni,Thiemens,Delmas,&Savarino,2007).Furthermore,with low atmospheric and marine pO2, homogenisation of discreteatmosphericsulphurspecies(e.g.,SO2−

4,Sn,SO2)waslimited,further

favouringpreservationoftheArchaeanS-MIFsignal.After formation, the photochemically derived sulphur species,

generally assumed to be sulphate (SO2−

4) and elemental sulphur (Sn),

weredepositedintheoceans.Sulphateissolubleinwaterandwouldhomogenise rapidly. It could form sulphate-bearing phases, such ascarbonate-associatedsulphate(CAS),gypsum(CaSO4.2H2O)orbarite(BaSO4);oritcouldbereducedtohydrogensulphide(H2S)viamicrobi-allymediatedMDFredoxreactionsorthermochemicalsulphatereduc-tion(TSR)beforecaptureintherockrecordaspyrite.TheseS-cyclingprocessescansignificantlychangetheisotopecompositionofδ34Swithsmall(<0.2‰)differentialeffectsonΔ33S(Johnstonetal.,2005).ThefateofSn islesscertain,butoncedepositedontheseafloor,itcouldreactwithH2Sproducedbymicrobialsulphatereduction(MSR)toformareactive,mobile,solubleformofpolysulphideS2−

n.Reactivepolysul-

phideandhydrogensulphidecanthenreacttoformsulphide-bearingminerals,suchaspyrite(Farquharetal.,2013).Therefore,theS-isotopecompositionofpyritegenerallyrecordsthebiogeochemicalMDFpro-cessesthatgeneratedS2−

n and H2Sfromatmosphericprecursors.

Despite the narrative outlined above, there remain several poorlyconstrained aspects of the Archaean sulphur cycle. Firstly, the Δ33S signatures of the atmospheric products (e.g., Sn and SO2−

4) produced

throughphotochemicalreactionsremainasubjectofdebate.ArchaeanrocksfromtheHamersleyBasin,WesternAustralia,containpyritewith

S-isotope ratios that largely follow a linear trend in δ34SversusΔ33S alonganArchaeanreferencearray.Combiningthesedatawithphoto-chemicalmodels,Ono etal. (2003) suggested that theΔ33S signwaspositiveforelementalsulphurandnegativeforsulphate.Furthermore,Palaeoarchaean barites (BaSO4) also show a 0 to −1.5‰Δ

33S signal(e.g.,Roerdink,Mason,Farquhar,&Reimer,2012;Ueno,Ono,Rumble,&Maruyama,2008),butthesecouldhaveformedinthemarinewatercolumn, inthesedimentsor inhydrothermalenvironmentsnotneces-sarilyreflectiveoftheseawatersulphatepool(Paytan,Mearon,Cobb,&Kastner,2002;VanKranendonk,2006).Additionalphotochemicalexper-imentsproduceS-isotopeslopesthatcontrasttoOnoetal.’s(2003)re-sults(e.g.,seereviewofParis,Adkins,Sessions,Webb,&Fischer,2014);moreover, recent photochemicalmodels demonstrate that the sign inΔ33SofexperimentallyproducedsulphurwillbehighlydependentonthewavelengthofUVlightusedtosimulatephotolysis(Claireetal.,2014).

Furtherpoorlyconstrainedcomponentsof theArchaeanS-cyclearetheconcentrationandδ34Sofseawatersulphate.Archaeanseawa-tersulphateconcentrationswereexpectedtobelowerthanmodernvaluesbecauseofthelowatmosphericpO2beforetheGOE.Habicht,Gade, Thamdrup, Berg, and Canfield (2002) used culturing experi-mentsofmodernsulphate reducers toobservedecreasing fraction-ationfactorswithdecreasingsulphateconcentrations.Assumingthisrelationship can be extrapolated into theArchaean, they estimatedseawater sulphate concentrations at <200μM. However,MSR frac-tionation factors aredependentonmicrobial species (Bradleyetal.,2016) and other factors including reduction rate, type of substrateandtemperature(e.g.,Fikeetal.,2009;andreferencestherein).Morerecent results have suggestedArchaean seawater sulphate concen-trations could have been even lower, <10μM (Crowe etal., 2014;Zhelezinskaia,Kaufman,Farquhar,&Cliff,2014).

Theδ34Sratioofseawatersulphatedependsontheisotopecompo-sitionofthesulphateinputintotheoceans,thesteadystateburialfluxofsulphidesandsulphateandthefractionationfactorbetweencoevalsulphate-andsulphide-bearingspecies(Fikeetal.,2015).Particularly,ifthesulphateconcentrations intheArchaeanoceanswere lowandbasinswererestricted,MSRcouldhavechangedtheresidualsulphatecomposition of seawater through Rayleigh distillative fractionation(Roerdinketal.,2012).Therefore,seawatersulphateS-isotopecompo-sitionsmayhavebeenheterogeneous,bothspatiallyandtemporally,asproposedfromthescatterofδ34SvaluespreservedinNeoarchaeancarbonates(e.g.,Parisetal.,2014;Zhelezinskaiaetal.,2014).

Inaddition,thedistributionandoccurrenceofmicrobialsulphatereductionintheArchaeanisdebated.Bulkδ34SdatafromAustralianPalaeoarchaeanbarites(Shen&Buick,2004;Shen,Buick,&Canfield,2001)alongwithphylogeneticstudies(Wagner,Roger,Flax,Brusseau,&Stahl,1998) suggest thatmicrobial sulphate reduction is likelyanancientmetabolism.However,thelinkbetweenMSRS-isotopefinger-printsandpyritetexturesremainsenigmatic.Ono,Beukes,andRumble(2009)sampledmultiplepyritephasesin~2.5-GaupperPrieskafaciesfrom the GKP01 Agouron drill core. They correlated different py-ritephaseswiththeArchaeanreferencearray(Onoetal.,2003)andsuggestednodularand layeredpyritehadS-isotopesignaturesmostconsistentwithMSR.Similarlymatchingpyritecrystalmorphologyto

     |  355MEYER Et al.

the array, Kamber andWhitehouse (2007) proposed spheroidal py-riteconcretionscaptureanMSRS-isotopefingerprintinthe2.52-GaUpper Campbellrand Subgroup, Transvaal, South Africa. Moreover,Fischeretal.(2014)showedthatδ34Sinnodularpyritewerecharac-terisedbysystematicenrichment towards the rims (with the lowestδ34Ssignaturepresentinthecentreofthenodules).AdditionalstudiesintheGriqualandWestBasinarerequiredtotestwhetherthesetex-tural andgeochemical interpretations canbeapplied topyrite fromotherpalaeoenvironmentsanddepositionalages.

To address these gaps in our current understanding ofArchaeansulphur cycling,wemeasured in situ, μm-scale, multiple sulphur iso-tope ratios (δ34S and Δ33S) in exceptionally well-preserved pyritisedmicrobialites in shales (sample-3184) from the ~2.65-Ga LokammonaFormation,GhaapGroup,SouthAfrica.BycombiningpetrographywithS-isotopedatafromsecondaryionmassspectrometry(SIMS),weaimtoincreaseourunderstandingofsulphurcyclinginaNeoarchaeanmicro-bialmat,andthesecondary(late-diagenetic/post-lithification)processesthatmay impact the primary (depositional/early-diagenetic) S-isotopesignal.Bypetrographicallycorrelatingpyritephaseswithsulphurisotopefingerprints,wecandeterminetheMDFandMIFprocessesthatcontrib-utedtotheformationofmultiplepyritegenerationsinthesesediments.

2  | MATERIALS AND METHODS

2.1 | Geological setting and core material

Our sample (3184)was collected from theBH1-SACHA core, drilledthroughtheNeoarchaeanTransvaalSupergroupintheGriqualandWestBasin(Figure1).ThecorematerialwasobtainedfromtheNationalCoreLibraryatDonkerhoek(Pretoria,SouthAfrica).Thecorecontainscarbon-ates,siliciclasticsandironformationsaswellasseveraligneousintrusivebodies(Figure2).Sample-3184wasobtainedfromadepthof3184mfromtheLokammonaFormation,SchmidtsdrifSubgroup,Ghaapgroup,TransvaalSupergroup.TheLokammona(Clearwater)Formationisdomi-natedbyshale,tufflayersintercalatedwithblackshale,andminordo-lomites (Figure2). The high proportion of fine-grained sediments isconsistentwithdepositioninalow-energyenvironment;AltermannandSiegfried(1997)suggestedsedimentationoccurredinadeep,shelfen-vironmentwithshale-carbonate/shalecycles representingshallowing-upward cycles. They also noted the presence of microbial laminites,which suggested deposition in the photic zone. The interpretation issupportedbystudiessuggestingthatmodernmicrobialmatsthatpro-ducemicrobiallyinducedsedimentarystructures(MISS)aredominatedbybenthicphotoautotrophs(Noffke,2009).TheLokammonaFormationhasamodelageof2.650±0.008GafromSHRIMPU-Pbanalysesofzirconsfromtufflaminae(Knoll&Beukes,2009).

TheareasampledbytheBH1-SACHAcoreischaracterisedbylittlesubsequenttectonicdeformation(Beukes,1987),andithasbeensub-jectedtosub-greenschistfaciesmetamorphism(Button,1973;Miyano&Beukes, 1984).The core has been penetrated by dykes and sills;thelargestcontinuousigneousintrusionoccursatadepthof922.3–1201.27m.IntheLokammonaFormation,thereissomeevidenceofsecondarymineralisation,particularlyat~3,165mwheregalenahas

beenidentified(Altermann&Siegfried,1997).Wewillconsidertheseigneous andmetasomatic processeswhen interpreting geochemicaldatafromBH1-SACHA.

2.2 | Imaging and EPMA

WeusedtheVHX-2000super-resolutiondigitalmicroscopehousedintheSchoolofEarthandEnvironmentalSciencesattheUniversityof St Andrews (Scotland, UK) to image thick sections of sample-3184.ThisfacilitatedthedetailedmappingofstructuresandfabricswithinthesampleandselectionoftargetareasforSIMS.Backscatterelectron (BSE) imagingwas carried out using the School of Earthand Environmental Sciences’ Jeol JCXA-733 Superprobe electronmicroprobe analyser (EPMA). Prior to analysis, sample-3184 thicksectionswerecoatedwitha~40-nm-thickgraphitelayer.Theanaly-siswasperformedusingaprimaryioncurrentof~17–21nAandanaccelerationvoltageof15kV.

F IGURE  1 SimplifiedgeologicalmapofTransvaalandGriqualandWestsediments,theapproximatelocationofBH1-SACHA(star)andtheinferredfaulttraceoftheKheissolethrustfault.Insert:theredrectangleshowsthelocationofthemainmaprelativetootherTransvaalsedimentsandtheKaapvaalcraton(beige).AdaptedfromAltermannandWotherspoon(1995)[Colourfigurecanbeviewedatwileyonlinelibrary.com]

356  |     MEYER Et al.

2.3 | Multiple sulphur isotope analyses via SIMS

ThesulphurisotopedataarereportedrelativetotheViennaCanyonDiabloTroilite(V-CDT)internationalreferencestandard.Thedelta(δ)notationdenotesthedeviationofthesamplefromV-CDTinpermil(‰;Equations1–2).Thecapital-delta (Δ)reflectsmass-independentfraction,quantifyingthedeviationofasamplefromtheexpectedter-restrialmassfractionationline(Equation3).

Multiplesulphurisotopeanalyses(δ34S and δ33S)wereconductedusing the CAMECAIMS 7f-GEO secondary ion mass spectrome-ter housed in the Department of Earth and Planetary Sciences atWashingtonUniversity(StLouis,MO,USA).Secondaryionmassspec-trometry(SIMS)enablestheinsitumeasurementofS-isotoperatiosto

ahighspatialresolution,precisionandmassresolutionwhileminimis-ingdestructionto thesample.Epoxythicksectionsofsample-3184witha2.54cmdiameterweremade,polishedto<1μmsurfacerough-nessandcoveredwitha~50-nmgoldcoating.

Samples were outgassed in an ancillary chamber at <1×10−8 mB prior to analysis, and the pressure in the sample chamberwasallowed to equilibrate for >1hr after sample introduction, prior toanalysis.Thicksectionswereanalysedinavacuumwithapressureof3 × 10−9mB.Afocused~2μmCs+primaryionbeamof0.9–1.8nAwasrasteredovera10μm by 10 μmareaof interest.Eachmeasurementwas divided into 15 cycles and required 7–8min to complete. Thesecondaryionswerecollectedusingfaradaycups(FC).The7f-GEOisspecificallydesignedforpreciseisotoperatiomeasurementusingtwoFCsoptimisedformajor(FC1)andminor(FC2)isotopedataacquisitionin “chargemode” (Peres,deChambost,&Schuhmacher,2008).Eachisotope isselectedbymagnetswitching,andthecorrespondingsec-ondaryioncountswerecollectedinsequenceofascendingmass.ThedesiredFCisselectedusingadeflectorsituatedinthedetectorassem-bly.ThisconfigurationallowsfortheinitiationofdataacquisitionfromFC2priortocompletedissipationofthesignalfromFC1.Thisalternat-ing collectionarrangement results in improved signal-to-noise ratiosandtherebygreaterprecisionperunittime. Inthisstudy, twominor

(1)δ33S=

((33S∕32S)sample

(33S∕32S)V−CDT−1

)×1000

(2)δ34S=

((34S∕32S)sample

(34S∕32S)V−CDT−1

)×1000

(3)Δ33S=δ33SV−CDT−1000×

⎛⎜⎜⎝1−

�δ34SV−CDT

1000

�0.515⎞⎟⎟⎠

F IGURE  2 BH1-SACHAbulksulphurisotopedataplottedagainstcoredepth(m)withthecorrespondingsedimentarylog.Brecciatedzoneslargelyreflectpost-depositionaltectonicdeformation.Thepyritisedmicrobialitesanalysedinthisstudyweresampledatadepthof3,184m(greyband).Lokam.,representsLokammona.DatafromIzonetal.(2015)andstratigraphiclogadaptedfromAltermann&Siegfried,1997[Colourfigurecanbeviewedatwileyonlinelibrary.com]

     |  357MEYER Et al.

isotopeswere tobeanalysed.Asback-to-backdata collectionusingFC2woulddefeatthesignal-to-noiseadvantagedescribedabove;thus,apseudomeasurement,atmass33.5,wasmadeonFC1sothateachcycle became 32S(FC1),33S(FC2),33.5(FC1),34S(FC2).Thebackgroundlevel forFC1was~3×104 counts/sand forFC2~3×103 counts/s.Typically, secondary ion yields were: 32S~2×108 counts/s, 33S~106 counts/sand34S~106counts/s.

Backgroundandionyieldswererecalibratedonadailybasis.Masscalibration,aswellasautomaticpeakcentringofthefieldandcontrastapertures,wasperformedatthestartofeachmeasurement;thiscorrectsforanydriftinthesecondaryionbeamorthemagneticfield(Fikeetal.,2009).Toremovethe50-nmgoldcoating,eachmeasurementwaspre-sputteredfor2minpriortothestartofthe15cycles.Aminimummassresolvingpower(MRP)of3900isrequiredtoseparatethe33S and 32SH peaks,andtherefore,aMRPof~4300wasused.Analysiswaspreferen-tiallycarriedoutinregularspacedintervals inagrid-likepattern(Fike,Gammon,Ziebis,&Orphan,2008).Generally, theanalysisspotswereplacedat~50-μmintervals.However,deviationsfromtheidealgridwerenecessarytoconsistentlysamplepyriteinsteadofsiliciclasticmatrix.

Theinternalerror(standarderrorofn=15cycles)variedinverselywiththe32Scountrate(FigureS1),suggestingthatprecisionwaslim-ited by counting statistics (Fike etal., 2009).Anomalous datawereomitted from further analysis if any of the following applied to themeasurement:

1. The sputter area clearly targeted only matrix as subsequentlydetermined from BSE and super-resolution digital microscopeimages.

2. Theuncorrectedrelativestandarderrorfor32Scountswas>1%.3. Theuncorrectedrelativestandarderrorfor34S/32Swas>0.1%.

After excluding the anomalous measurements, the internal errorwas typically <1‰ for δ34S and <0.5‰ for δ33S (1SE). An in-houseWashingtonUniversitypyritestandardwasmountedinapolishedandgold-coatedthinsectionandplacedinaseparateholder.ItscompositionwasdeterminedbymeasuringtheS-isotopecompositionoftheinternalstandard12timesandtheBalmatstandard20times.TheBalmatstan-dard(δ34S=15.1‰andδ33S=7.7‰)analysesbracketedthein-housestandard.The in-housestandardhasaknown isotopecompositionofδ34S=0.13±0.30‰andδ33S=0.13±0.20‰(1SE).Theexternalerrors(1standarddeviationofmultipleadjacentpointsonthein-housestan-dard)weretypically0.38‰and0.32‰forδ34S and δ33S,respectively(n=35).Withinsample-3184,therewere26‰variabilityinmeasuredδ34Svaluesand14‰variability inδ33S.Thisvariationfarexceedsthecalculatedinternalandexternalerrors.

On average, a session consisted of ~50 measurements of un-knownsandwasbracketedby≥4measurementsofthein-housestan-dardonboth sidesof theunknowns (FigureS2). Instrumentalmassfractionation(IMF)wascorrectedforbystandard-samplebracketing.Themean,uncorrectedisotopecompositionofrepeatedanalysesofthein-housepyritestandardontheSIMSwasδ34S=−0.05±0.38‰and δ33S=1.00±0.32‰ (1σ; n=35; relative toV-CDT).Therefore,theIMF(IMF=Rraw/Rknown)wastypically1.000and1.001for

34R and

33R, respectively.Although IMF can vary according to instrumentalconditions,theS-isotopeoffsetbetweenbracketingstandardsandun-knownsshouldnotvary.Furthermore,themeasuredS-isotoperatiosofasession’sbracketingstandardswereobservedtodeterminedrift.Thedriftvaluewasusually<1‰over24hr.Tocorrectforthis,alinearandconstantdriftwasassumedwithina session.Themagnitudeofthisdriftwasmuchsmallerthanthesignalsobservedandwasthere-foreunlikelytohaveimpactedtheresults.

3  | RESULTS

3.1 | Textural analysis

Sample-3184isablackshalecomposedof~1-mm-thicklaminaewithvaryingproportionsofpyriteandmatrix(Figure3).Thematrixiscom-posed of silt-sized particles (~40μm) of aluminosilicates (clay min-erals; 80%modal abundance) and quartz (20%).Uncommon<5-μm detritalgrainsofrutile(TiO2)andapatite(Ca10(PO4)6(F,Cl,OH)2)makeup<1%ofthematrixvolume.Laminaecontainvariableproportionsofpyrite-to-matrixratios,rangingfrom10%–70%pyriteto30%–90%matrix.Therearethreedistinguishablepyritephases:A)disseminatedpyritecrystals,<5μminsize;B)type 1 pyrite:irregularaggregatesofcoalescedpyritecrystals.Clusterscanrangefrom10μmto500μm across,andarecomposedofanhedraltosubhedral,1-to20-μmpyritecrystals.(Figure4);C)type 2 pyrite:euhedralcubicpyritewithacrystalsizerangingfrom10to100μm(Figure4).Someeuhedralpyritecrys-talsareisolatedwithinthematrix,whileothercrystalsovergrowandencrusttype1aggregates.Overgrowthsoccurontheedgesoftype1pyriteaggregatesandwithinclusterswherematrixlensesoccur.Type1 and type 2 pyrites are chemically distinct phases, as determinedfromBSEimages(Figures4andS6).

Laminae containing type 1 pyrite show sedimentary structuresthatsuggestitprecipitatedduringearlydiagenesis,priortosoftsed-imentdeformation.Theseincludewavycrinkledlaminaewithtypicalwavelengthsof~500μmandwaveheightsof~200μm,isoclinalfolds,1-mmrip-upstructuresandover-foldedlaminae.Particularly,<2-mmirregular,ellipsoidalconcretionsarecomposedofpolycrystallinetype1pyrite.Theycross-cutlaminationbutalsocausedrapingoflaminaeoneithersideoftheconcretion(Figures4andS4);thissuggeststype1pyriteprecipitatedafterdepositionbutbefore lithification,duringearlydiagenesis.

A normal microfault cross-cuts the entire core section of sam-ple-3184(Figure3).The~50-μm-widemicrofaulthasbeeninfilledbyaluminosilicatesandsilica.Themicrofaultcross-cutstype1pyrite;type2pyritecrystalscross-cutthemicrofaultitself(Figure4).Furthermore,type2pyriteovergrowsthesedimentarystructurescomposedoftype1pyrite.Accordingtothesecross-cuttingrelationships,type2pyriteprecipitatedpost-lithification.

3.2 | Multiple sulphur isotope data

All the SIMS data measured in sample-3184 are presented inFigure5. There is no obvious stratigraphic trend; however, there is

358  |     MEYER Et al.

a relationshipbetweenpyrite crystal shapeandS-isotopecomposi-tions.Thefirstdatasetshowsameanofδ34S=12.54±4.98‰andΔ33S=−0.21±0.65‰(±1σ,n=177)andcorrespondstotype1py-rite(Figures6and7).Insample-3184,type1pyriteδ34Svaluesvaryby 26‰; within a ~500×500μm area, the range of δ34S is 15‰(Figure6).Wavycrinkledlaminaeshowδ34S and Δ33Svaluesconsist-entwithtypicaltype1pyrite(FigureS3).Theseconddatasethasamean of δ34S=8.36±1.16‰andΔ33S=5.54±1.53‰(±1σ,n=18)andwasmeasuredintype2pyrite(Figures7andS5).Toa95%confi-dencelevel,theS-isotopevaluesfortype1andtype2pyritearesignif-icantlydifferentintheirmediansandvariances(p = .000 for δ34S and Δ33Sintwo-sampleWilcoxonandLevene’stests).SomedatapointscorrespondtoanintermediaryS-isotopesignaturebetweenthetype1

andtype2pyriteendmembers.Whenthesemeasurementsweresub-sequentlyexaminedintheirpetrographiccontexts,photomicrographssuggesttheanalysisspotstargetedamixtureofbothpyritephases.

4  | DISCUSSION

4.1 | Early- diagenetic (type 1) pyrite

Laminae composed of type 1 pyrite are wavy and crinkled on asub-mm scale (Figures3 and 4) and are consistent with sedimen-tary structures thathavebeenmicrobially induced (Noffke,2009).Prokaryotes and eukaryotes in microbial mats produce a matrixof extracellular polymeric substances (EPS), which are composed

F IGURE  3  (a)Reflectedlightscannerimageofsample-3184fromtheLokammonaFormationshowingtheareasofSIMSanalyses(whitedots).Bluedot=Figure6analysisarea;greendot=Figure7analysisarea.Divisionsonthescalebarrepresent1mm.(b)Atraceoftheimagein(a)showinglaminaecompositions,typeexamplesofsedimentarystructuresdiscussedinthetextandthelocationofthenormalmicrofault.Modalpercentagesoflaminaecompositionsareshowninbrackets(pyrite/matrix)[Colourfigurecanbeviewedatwileyonlinelibrary.com]

(a) (b)

     |  359MEYER Et al.

of polysaccharides, proteins, humic substances and nucleic acids(Nielsen, Jahn, & Palmgren, 1997). The matrix has several physi-cal functions, including adhesion to surfaces, aggregation of cells,stabilisationofmicrobialmat structure, sorptionof exogenousor-ganicmoleculesandretentionofwater(Laspidou&Rittmann,2002).Importantly,EPS is responsible for the formationofwavycrinkledlaminationduetoitsabilitytotrapdetritalgrainsanditshighcohe-siveproperties.Thiscohesion isalsoresponsiblefortheformationofover-foldedlayers(Figure3),whichoccurwhenthematsurface

iserodedand turnedoverat itsedges (Schieber,1999).However,wavycrinkled laminaecanalso form through thedifferential com-pactionofphyllosilicatesaround,forexample,microconcretions,siltlensesorsilicaspherules(Schieber,2007).ExamplesofcompactionaroundsuchstructureswereabsentinBSEimagesofsample-3184(Figure4).Thus, thesedimentarystructures indicatehighcohesionofthedepositionallayersandareconsistentwithtypicalmicrobialitetextures,supportingabiogeniccontrolontheformationof type1pyrite.

F IGURE  4 BSEimagesofdifferentsedimentarystructuresinsample-3184.(a)Wavycrinkledinternallaminationcomposedoftype1pyriteanddisseminatedpyrite.(b)Wavycrinkledinternallaminationcomposedoftype1pyritewithtype2pyriteovergrowths.(c)Pyriteconcretion.Notethattheconcretionbothcross-cutsandcausesdeformationofthelamination.(d)Thenormalmicrofaultthatisinfilledbyquartzandclayminerals.Notethemicrofaultandveincross-cuttype1pyrite;type2pyritecross-cutsthemicrofaultandvein.Therefore,therelativeorderofformationisasfollows:pyrite1precipitatedfirst,brittledeformationcausedmicrofaultformation,themicrofaultwasinfilled,andfinallytype2pyriteprecipitated.py,py1,py2andqtzrepresentpyrite,type1pyrite,type2pyriteandquartz,respectively[Colourfigurecanbeviewedatwileyonlinelibrary.com]

F IGURE  5 PlotofmultipleS-isotopedata(δ34S and Δ33S)measuredviaSIMS.SIMSerrorbarsare1SE for each measurement(n=15cycles).Greentrianglesrepresentamixedsignal,wheretheareaofanalysissampledbothtype1andtype2pyrite.TheorangelineistheArchaean reference arrayasdescribedbyOnoetal.(2003)[Colourfigurecanbeviewedatwileyonlinelibrary.com]

360  |     MEYER Et al.

TheS-isotopecompositionof type1pyrite isdistinguishablebya Δ33Svalueof−0.21±0.65‰(±1σ)anda26‰rangeinδ34Swithinsample-3184.TheΔ33Serrorbarsfortype1pyritewithinanareaofanalysisoverlap(Figure6);therefore,theΔ33Svariationwithinalaminaisstatisticallyinsignificant.Theconsiderablevariationinδ34SvaluesisconsistentwithMSR—modernmicrobialmatswithsulphatereducersshow~15–53‰variationsinδ34Sona1-mmscale(Fikeetal.,2009;Wilbanks etal., 2014), similar towhatweobservehere.Thus, bothtexturalevidenceandtheS-isotopecompositionareconsistentwiththe expected biosignatures of an ancient microbial mat containingsulphatereducers.The26‰rangeinδ34Swithinsample-3184couldbeexplainedbynon-uniformity inMSR fractionation factors,whichdependonthemicrobestrain,typeofelectrondonors,reductionrate,temperatureandthepresenceofadditionalSmetabolismssuchasox-idationanddisproportionation(e.g.,Fikeetal.,2009).

Alternatively,thescatterandenrichmentinδ34SsulphiderelativetocoevalsulphatecouldbeproducedbyfractionationduringMSRina (partially)closedsystem,analogoustoamatenvironmentperiodi-callyflushedwithseawatersulphate.WeusedaRayleighdistillationmodeltoexplorethisscenario,usingasuiteofrealisticfractionationfactors for MSR (αMSR=1,000×[

34Rsulphate/34Rsulphide − 1], where

34R = 34S/32S)andvariableseawaterδ34Ssulphatevalues.Figure5showsthat few data occur above 18‰, and thus, the curve showing theδ34Ssulphideoftheinstantaneousproductmustbeinthedistillativetailwhen δ34Ssulphide>18‰.Valuesof6‰–12‰forαMSR,4‰–16‰forδ34Ssulphate and fvalues from~0.5 to0.9 are all consistentwith thedata(Figure8).

This 6‰–12‰ estimate of αMSR is at the lower end of thespectrum for themodernMSRcell-specific fractionation factorsofαMSR=2‰–66‰(Fikeetal.,2015).SmallMSRisotopefractionationcanoccurwhensulphateconcentrationsarelow,whensulphatere-ductionratesorsulphideoxidationratesarehigh,whenH2 isusedasanelectrondonorinsteadoforganiccarbon,and/orwhensulphurdisproportionationlevelsarelow(e.g.,Fikeetal.,2009andreferencestherein).Although the relationshipbetweensulphateconcentrationandMSR fractionation factor is complex (seeBradley etal., 2016),ourestimatesofαMSRareconsistentwithotherstudiesthatsuggestlowseawatersulphateconcentrationsandhighratesofMSRintheArchaean(e.g.,Croweetal.,2014;Habichtetal.,2002;Zhelezinskaiaetal.,2014).Furthermore,theseawaterδ34Scompositionsweesti-matefromclosed-systemmodellingoftheSIMSdataareconsistentwith estimates of seawater sulphate δ34S from Neoarchaean CAS

F IGURE  6 CombinedpetrographywithSIMSS-isotopedatashowsthetypicaltextural,δ34S and Δ33Scharacteristicsoftype1pyrite.(a)Reflectedlightphotomicrographoftheanalyticalgridlocation.(b)BSEimageoftheanalysisareaoverlainbyaSIMSδ34Simageconstructedbysplineinterpolationoftheanalyticalgrid(n=31).Notethattheanalysisareaiscomposedofa~500-μmaggregateofanhedraltosubhedral,1-to20-μmpyritecrystals.(c)PlottoshowSIMSδ34SagainstΔ33Sdata(‰).Notethevariationinδ34S dataissignificant;thevariationinΔ33S dataisnotsignificant.Errorbarsrepresent1SEforeachmeasurement(n=15cycles)(d)BSEimageoftheanalysisareaoverlainbyaSIMSΔ33Simageconstructedbysplineinterpolationoftheanalyticalgrid(n=31)[Colourfigurecanbeviewedatwileyonlinelibrary.com]

     |  361MEYER Et al.

(e.g.,Domagal-Goldman,Kasting,Johnston,&Farquhar,2008;Guoetal.,2009;Parisetal.,2014).However,thesmallMSRfractionationfactorsandvariationinΔ33SvaluesintheliteratureimplyArchaeanseawatersulphatehadashort residencetimeduetoasmall reser-voirsizerelativetothefluxesintoandoutofthereservoir;thus,theS-isotopecompositionofseawatersulphatewas likelyspatiallyandtemporallyheterogeneous (Fischeretal.,2014;Zhelezinskaiaetal.,2014).

Wethereforeconcludethattype1pyritereflectsthemorphologi-calandgeochemicalsignaturesofsulphatereducersinaNeoarchaeanmicrobial mat, as inferred from (i) the biogenicity of sedimentarystructureslikewavycrinkledlamination,(ii)texturalevidenceofearly-diagenetic precipitation of type 1 pyrite, and (iii) models of type 1pyriteδ34SvaluessupportingplausibleMSRfractionationfactorsex-pressedwithinarestrictedseawatersulphatepool.

4.2 | Secondary (type 2) pyrite

Asdiscussedabove,simplecross-cuttingrelationssuggesttype2py-riteprecipitatedafter type1pyrite formationandpost-lithification,because it overgrows deformed aggregates of type 1 pyrite, and itcross-cuts a normal microfault with vein infill (Figure4). The lowstandard deviationof type 2 pyrite S-isotope ratios suggests these

crystals have all been formed from the same, uniform S-isotopesource.Theδ34S and Δ33Ssignatureoftype2pyritefallswithinOnoetal.’s (2003)estimatesof thecompositionofatmosphericelemen-talsulphurasinferredfromArchaeanpyritedataandphotochemicalexperiments(Figure5),andParisetal.’s(2014)NeoarchaeanCASval-ues.Similarδ34S and Δ33SratiosineuhedralgrainssampledfromGKF-01 (a core sampling deeperwater equivalents of the BH1-SACHA;Schröder, Lacassie,&Beukes, 2006)were interpreted as having anatmosphericelementalsulphurorigin(Farquharetal.,2013).Farquharetal.(2013)hypothesisedthatsolidatmosphericallyderivedSnparti-clescouldremaininanunreactiveformastheyfellthroughthewatercolumn.Afterdeposition,theycouldthenreactwithH2Sinthesedi-ment,producingreactivepolysulphide.Finally,thepolysulphidecouldreactwithFeStoformpyrite.Farquharetal.(2013)predictedthatthepyritewould closely reflect the S-isotope signature of atmosphericelementalsulphurifthemasscontributionofH2Swassmallrelativetopolysulphideinthepyriteproduct.

Asimilarinterpretationoftype2pyriteformationinsample-3184is less likely, due to its inferred later timing of formation based onthe petrographic relationshipswe describe above.As type 1 pyritelikely formed from sulphidegeneratedviamicrobial sulphate reduc-tion,H2Swasabundantduringearlydiagenesis.Therefore,anyatmo-spheric,unreactiveelementalsulphurparticlesthatdepositedintothe

F IGURE  7 ThecontrastingtexturalandS-isotopesignaturesoftype1andtype2pyrite.(a)Reflectedlightphotomicrographoftheareaofanalysis.(b)BSEimageoftheanalysisareaoverlainbyaSIMSδ34S image constructedbysplineinterpolationoftheanalyticalgrid(n=15).Notethe~300-μm cubictype2pyritegrain,overgrowingtype1pyriteaggregates.(c)PlottoshowSIMSδ34SagainstΔ33Sdata(‰).NotethesignificantdifferenceinΔ33Sfortype1and2pyrite.Errorbarsrepresent1SE for eachmeasurement(n=15cycles).Bluecircles=type1pyrite;greencircles=type2pyrite.(d)BSEimageoftheanalysisareaoverlainbyaSIMSΔ33Simageconstructedbysplineinterpolationoftheanalyticalgrid(n=15)[Colourfigurecanbeviewedatwileyonlinelibrary.com]

362  |     MEYER Et al.

NeoarchaeanmicrobialmatwouldhavesimultaneouslyreactedwithH2Stoformasoluble,reactiveformofpolysulphide.Duringearlydia-genesis,themobilepolysulphide,aprecursortopyritesynthesis,wouldhavereactedwithanironmonosulphidetoformFeS2.However,petro-graphic relationshipssuggest type2pyrite formedpost-lithification,notcoevallywithtype1pyriteduringdiagenesis(Figure4).Therefore,type2pyriteinsample-3184unlikelyformedfromanelementalsul-phur source by themechanism proposed by Farquhar etal. (2013).Hence,wesuggesttheΔ33S=5.5‰signatureoftype2pyritewasnotproducedfromcontemporaneousatmosphericSnproductsatthetimesample-3184wasformed.Instead,weproposelater-stagealterationprocessescouldexplaintheΔ33Svaluesmeasuredintype2pyrite.

The BH1-SACHA core has been altered by several post-depositional hydrothermal events that could have caused multi-generational pyrite genesis. Stratigraphic logs indicate evidence ofsecondary mineralisation of galena, ~20m stratigraphically higherthan sample-3184. Furthermore, intrusion of the igneous body at922.3–1201.27mformedacontactmetamorphicaureole(Altermann&Siegfried,1997).To thebestofourknowledge, thisdykehasnotbeendated;however, itmaybeco-genetictoothermaficintrusionsrelatedtotheemplacementoftheBushveldcomplex,2.06–2.05Ga(Hartzer,1995).Associated igneousandmetamorphicfluidscanca-talysemetasomaticreactions,causingthe lossandgainofelementsincirculatingfluids.TheoccurrenceofmetasomatisminBH1-SACHAis supported by regional geological evidence; a platform-wide fluidfloweventat~2.0GacouldhavecausedthePb-Zn-Cu-mineralisationin the Campbellrand Subgroup, Ghaap Group. Huizenga, Gutzmer,Greyling,andSchaefer(2006)sampledMississippiValley-type(MVT)deposits in the Griqualand West area, ~100–120km straight-linedistancefromKathu.Theyestimatedaregionalfluidtemperatureof200–240°Candpressureof0.8–1.5kbarduringmineralisationoftheMVT deposits. It is possible that BH1-SACHA experienced similarmetamorphic conditions. However, Altermann and Siegfried (1997)

notedevidenceofthrustfaultsinBH1-SACHAatdepths>2,800mintheMontevilleFormation.Theysuggestedthis isthe intersectionofBH1-SACHAwiththeKheissolethrustfault(Figure1).Allochthonoussections of Archaean Transvaal Supergroup rocks (including rocksthatwere sampledby theBH1-SACHAcore) aswell asProterozoicWaterbergandOlifantshoekGroupswerehorizontallydisplacedandnow lie unconformablyon the autochthonousArchaeanpackageofrocks(Altermann&Wotherspoon,1995;Martini,Eriksson,&Snyman,1995). The age assigned to thrusting ranges from 2.20 to 1.75Ga(Grobbelaar, Burger, Pretorius,Marais, &VanNiekerk, 1995). If thethrust isolder than~2.0Ga, itsuggests that thehangingwall rockswereexhumedbeforethekeyigneousandmetasomaticeventsasso-ciatedwiththeBushveldComplex.Hence,themetamorphicgradeofBH1-SACHAinthehangingwallmaybelowerthanthefootwallMVTdeposits examined by Huizenga etal. (2006). In addition, De Kocketal.(2009)showedthatsimilarstratapreservedwithintheGKP-01drillcorethroughtheGhaapGroupshowpervasiveremagnetisationby2.5-to1.8-Gananoscalepyrrhotite.Furthermore,high-resolutionpalaeomagneticandgeochemicalevidencefromGKF-01NeoarchaeannodularpyritesuggeststhatsomeoftheTransvaalSupergrouppyriteshavebeenpost-depositionallyaltered,~0.5Gaaftertheirdeposition(Fischeretal.,2014).Thus,evenundeformedGhaapGroupstratacanshowacomplexhistoryofironsulphidemineralprecipitation.

The geological evidence therefore suggests sulphur could havecirculated influids as soluble sulphur species andparticipated inhy-drothermalreactionsintheGriqualandWestBasin.TherearemultiplesourcesofsubsurfaceS-richfluids (Figure9): Inmodernsettings,sul-phatecanbeacquiredfrom(buried)seawater(Ohmoto,1972).Ifthehy-drothermalalterationsoccurredaftertheGOE,seawaterwasapotentialsource(A,Figure9).Furthermore,porewatersexpelledfromsedimen-tary rocks during compaction, fluids formed by magmatic/metamor-phicprocessesormeteoric-orseawater-derivedfluidscancausethedissolutionofS-bearingminerals.Dissolutionofsulphate,suchasCAS

F IGURE  8 Graphsshowingthecalculatedδ34Svaluesofthehydrogensulphideinstantaneousproductrelativetotheproportionofsulphateconsumed(f),theinitialδ34SsulphatecompositionandtheMSRfractionationfactor(αsource-product).Thegreyrectanglescorrespondtotherangeofδ34Ssulphidecompositionsmeasuredintype1pyriteinsample-3184.(a)α=6‰,(b)α=12‰[Colourfigurecanbeviewedatwileyonlinelibrary.com]

     |  363MEYER Et al.

orsulphate-bearingminerals(e.g.,gypsum,anhydriteorbarite),occursreadily,whilepyritedissolutioncanalsooccurinthepresenceofanoxi-dant,forexample,O2orFe

3+(Descostes,Vitorge,&Beaucaire,2004;B,Figure9).Moreover,bothsulphateandsulphidecanbeproducedfromthedisproportionationofmagmaticSO2.Forexample,duringthecon-densationofamagmaticplume,H2SO4 and H2Sformfromthedispro-portionationofmagmaticSO2attemperatures200–400°CandpH<3.Sulphatecanbeleachedbyfluids,causingtheproductionofanacidic,sulphate-bearinghydrothermalfluidthatcaninfiltratesurroundingwallrock (C,Figure9;Rye,2005).The intrusionof theBushveldComplexat~2.06Gacouldhaveproducedmagmaticsulphurspecies (Hartzer,1995).Finally,thesesubsurfaceS-richfluidscouldhaveinfiltratedsur-roundingrocksviafracturefloworporousflow,andprecipitatedpyrite.

WesuggesttwoscenarioswherebythecirculationofS-richfluidsfromtheabovesourcescouldhavecontributed to the formationoftype2pyriteand itsassociatedS-MIFsignal:oneviathermochemi-calsulphatereductionofS-bearingfluidsfromamorerecentsulphursource, and oneviamigration of S-rich fluids from stratigraphicallyoldersedimentscarryinganinheritedArchaeanS-MIFsignal.

Ifthesubsurfacesulphur-richfluidweresourcedfrommodernorpost-GOEseawater,thedissolutionofpost-GOEsulphurmineralsormagmaticsulphate,itwouldnotcarryaprimaryMIFsignal.Therefore,toprecipitatetype2pyritewithaΔ33Sfingerprintof+5.5‰fromoneof thesesources,apost-depositionalprocesswouldbenecessary tocause S-MIF. Experimental investigations have shown that thermo-chemicalsulphatereduction(TSR)canproducea<13‰enrichmentinΔ33Srelativetothesulphatesource(Oduroetal.,2011).Ananomalousenrichmentof33Srelativeto34SduringTSRoccursduetomagneticisotopeeffects (MIE). In the sulphur isotope system, the 33S-isotopecontainsmagneticnuclei;duringTSR,33Sundergoesthespin-selectivereactionfasterthan34S and 32S.Thisresultsinananomalousenrichment

of 33S relative to34S in thepolysulphideproduct,and therefore, theTSRproductcarriesapositiveΔ33Ssignal(Oduroetal.,2011).

Thermochemical sulphate reduction occurs at 100–300°C(Johnston, 2011); these temperatures are equivalent to sub-greenschisttogreenschistfacies,whichcorrespondtothemaximummetamorphic grade that the lower Transvaal Supergroup, includingBH1-SACHA,experiencedduringintrusionoftheBushveldComplexin the Palaeoproterozoic (Sumner & Beukes, 2006). Therefore, thetemperature regime during contact metamorphism of BH1-SACHAandthesurroundingrockswassufficientlyhightosupportTSR.Thepetrographicevidenceand isotopefingerprintof type2pyrite,cou-pled with the history of hydrothermal alteration described above,couldbeconsistentwithMIEduringTSRcausingananomalousen-richmentin33Srelativeto34S.However,thereiscurrentlynoevidenceofanomalousΔ33SsignalsassociatedwithTSRinbulkrockanalysesfromthegeologicalrecord.Furthermore,TSRcausesaMIFeffect in33S,butnot36S(Oduroetal.,2011);therefore,thishypothesiscannotbeconclusivelydemonstratedwithouttheinclusionofΔ36Sdata.

Morelikely,theMIFsignalmeasuredintype2pyritecouldhaveformedfromthedissolutionandreprecipitationofArchaeanS-bearingmineralsoriginallydepositedwithapositiveΔ33Ssignature(Figure9).Notably,bulkpyriteshowsaΔ33Scompositionwithinerrorofthetype2 pyrite S-isotope fingerprint, 20–30m stratigraphically lower thansample-3184(Figure2).Therefore,dissolutionofthesepyritephasesandmigrationofΔ33S=+5.5‰sulphur-richfluidsupwardsarealikelysourceofthetype2pyriteMIFsignal.Therefore,grain-scaleArchaeanMIFsignaturesmayreflectsecondarypyriteformationmechanismsaswellasatmosphericprocessesatthetimeofdeposition.Thesemul-tiplepyritephasescanreflectaprotractedhistoryofpyriteprecipita-tion,whichcanberecognisedusinghigh-resolution,insituS-isotopegeochemistry.

F IGURE  9 Hypothesisedformationofpost-lithificationpyritewithananomalousΔ33Ssignal.Sulphurcanbesourcedviathreepathways:(A)seawater,(B)dissolutionofsulphate-orsulphide-bearingmineralsor(C)thedisproportionationofSO2inmagmatic-hydrothermalormagmaticsteamenvironments[Colourfigurecanbeviewedatwileyonlinelibrary.com]

364  |     MEYER Et al.

5  | CONCLUSIONS

Sulphate reduction in Neoarchaean microbial mats:Ourtexturalandsul-phur isotopedataechootherstudiesthatsuggestMSRinfluencedS-cyclingintheNeoarchaean(e.g.,Habichtetal.,2002;Uenoetal.,2008;Zerkle,Claire,Domagal-Goldman,Farquhar,&Poulton,2012).Astypi-calmicrobialitetexturesaregenerallyassociatedwithphotoautotrophs(Noffke,2009),thisstudyproposesthatmicrobialsulphatereducersmayhavecommonlyoccurredinthesamecommunityasphotosynthesisers,formingamicrobialmatecosystemsimilartothoseofthemodern.

Neoarchaean seawater sulphate concentrations and S- isotope compo-sition:SIMSS-isotopedataandclosed-systemmodellingsuggestthelocalsulphurreservoirhadaΔ33S=−0.21±0.65‰(±1σ,n=177)andδ34S=4‰–12‰duringformationoftype1pyritesinsample-3184(Figure8). These values are consistent with previous estimates ofthe δ34S-isotope composition of Archaean sulphate (e.g., Domagal-Goldmanetal.,2008;Guoetal.,2009;Parisetal.,2014).However,theS-isotopecompositionofseawatersulphatemayhavebeenspa-tiallyandtemporallyvariable,particularlyat the lowsulphate levelsestimated for theNeoarchaean (Zhelezinskaia etal., 2014). Closed-systemmodellingofsample-3184indicatessmallMSRfractionationfactors (αMSR=6‰–12‰),consistentwithhighreductionratesandlowsulphateconcentrations.ThesmallMSRfractionationfactorsandscatterinΔ33Svaluesintheliteratureimplysulphatehadashortresi-dencetimeinArchaeanoceansduetoasmallreservoirsizeandcom-parativelylargefluxesintoandoutofthereservoir.Additionalstudiesassessing the S-isotope compositionofCASwill help constrain theevolutionofseawatersulphateδ34S and Δ33SduringtheArchaean.

The utility of combining petrography and SIMS: This study demon-stratesthatgrain-scaleMIFsignalscouldrecordsecondaryprocessesinadditiontocontemporaneousatmosphericprocessesatthetimeofdeposition.MIFsignalscanbegeneratedbyMIEduring laterTSRor,moreprobably,leachedfromsurroundingArchaeansedimentaryrocksandreprecipitatedfollowinglithification.Furthermore,itdemonstratesthatArchaeanrockscanpreserveacomplex,protractedhistoryofpyriteformation.Pyritegenesiscanoccuroverseveralgenerations;precipita-tioncanrangefrompre-softsedimentdeformationtopost-lithification.Thisstudyhighlightstheeffectivenessofpairingpetrographywithhigh-resolutionSIMSsulphurisotopestudies,tounravelthediageneticpro-cessesthathavecontributedtomultigenerationalpyritegenesis.

ACKNOWLEDGMENTS

WethankDr.CliveJonesforhelpingwithSIMSanalysis,Mr.DonaldHerd forassistancewith imagingandthestaffat theSouthAfricanNational Core Library in Donkerhoek for facilitating access to thecorematerials.TheauthorsthankTonyPrave,TimRaub,GarethIzon,MarkClaireandJamesFarquharfortheir inputanddiscussion.ThisstudywasfinanciallysupportedbytheGeologicalSocietyofLondon’sAlan & Charlotte Welch Fund, A Palaeontological AssociationUndergraduate Research Grant (PA-UB201504) and the Universityof St Andrew’s Zannah Stephen Memorial Travel Scholarship(all to N.M.), Natural Environment Research Council Fellowship

NE/H016805/1andNaturalEnvironmentResearchCouncilStandardGrantNE/J023485/2 (to A.Z.) and anNSF/EARGrant (#1229370)andaPackardFellowship(toD.F.)

REFERENCES

Altermann,W.,&Siegfried,H.P.(1997).Sedimentologyandfaciesdevel-opment of an Archaean shelf: Carbonate platform transition in theKaapvaalCraton, asdeduced fromadeepborehole atKathu, SouthAfrica.Journal of African Earth Sciences,24,IN1–IN4.

Altermann, W., & Wotherspoon, J. M. (1995). The carbonates of theTransvaal and Griqualand West Sequences of the Kaapvaal craton,withspecialreferencetotheLimeAcreslimestonedeposit.Mineralium Deposita,30,124–134.

Baroni,M.,Thiemens,M.H.,Delmas, R. J., & Savarino, J. (2007).Mass-independent sulfur isotopic compositions in stratospheric volcaniceruptions.Science,315,84–87.

Beukes,N.J.(1987).Faciesrelations,depositionalenvironmentsanddia-genesis inamajorearlyProterozoicstromatoliticcarbonateplatformto basinal sequence, Campbellrand Subgroup,Transvaal Supergroup,SouthernAfrica.Sedimentary Geology,54,1–46.

Bradley,A. S., Leavitt,W.D., Schmidt,M., Knoll,A. H., Girguis, P. R., &Johnston,D.T.(2016).PatternsofsulfurisotopefractionationduringMicrobialSulfateReduction.Geobiology,14,91–101.

Button, A. (1973). Algal stromatolites of the early Proterozoic WolkbergGroup,Transvaalsequence.Journal of Sedimentary Research,43,160–167.

Claire,M.W.,Kasting,J.F.,Domagal-Goldman,S.D.,Stüeken,E.E.,Buick,R.,&Meadows,V.S. (2014).Modeling the signatureof sulfurmass-independent fractionation produced in the Archean atmosphere.Geochimica et Cosmochimica Acta,141,365–380.

Crowe, S. A., Paris, G., Katsev, S., Jones, C., Kim, S. T., Zerkle, A. L., …Farquhar,J.(2014).SulfatewasatraceconstituentofArcheanseawa-ter.Science,346,735–739.

DeKock,M.O.,Evans,D.A.D.,Kirschvink,J.L.,Beukes,N.J.,Rose,E.,&Hilburn,I.(2009).PaleomagnetismofaNeoarchean-Paleoproterozoiccarbonate ramp and carbonate platform succession (TransvaalSupergroup)fromsurfaceoutcropanddrillcore,GriqualandWestre-gion,SouthAfrica.Precambrian Research,169,80–99.

Descostes, M., Vitorge, P., & Beaucaire, C. (2004). Pyrite dissolution inacidic media. Geochimica et Cosmochimica Acta,68,4559–4569.

Domagal-Goldman, S. D., Kasting, J. F., Johnston, D. T., & Farquhar, J.(2008).Organic haze, glaciations andmultiple sulfur isotopes in theMid-ArcheanEra.Earth and Planetary Science Letters,269,29–40.

Farquhar, J., Bao, H., & Thiemens, M. (2000). Atmospheric influence ofEarth’searliestsulfurcycle.Science,289,756–758.

Farquhar, J., Cliff, J., Zerkle, A. L., Kamyshny, A., Poulton, S.W., Claire,M., …Harms, B. (2013). Pathways for Neoarchean pyrite formationconstrained bymass-independent sulfur isotopes.Proceedings of the National Academy of Sciences,110,17638–17643.

Fike,D.A.,Bradley,A.S.,&Rose,C.V.(2015).Rethinkingtheancientsulfurcycle. Annual Review of Earth and Planetary Sciences,43,593–622.

Fike,D.A.,Finke,N.,Zha,J.,Blake,G.,Hoehler,T.M.,&Orphan,V.J.(2009).HighresolutionSIMS-basedsulfideδ34S:Anewtoolforcharacterizingmicrobialactivityinavarietyofdepositionalenvironments.Geochimica et Cosmochimica Acta,73,6187–6204.

Fike,D.A.,Gammon,C.L.,Ziebis,W.,&Orphan,V.J.(2008).Micron-scalemappingofsulfurcyclingacrosstheoxyclineofacyanobacterialmat:ApairednanoSIMSandCARD-FISHapproach.ISME Journal,2,749–759.

Fischer,W.W.,Fike,D.A.,Johnson,J.E.,Raub,T.D.,Guan,Y.,Kirschvink,J.L.,&Eiler,J.M.(2014).SQUID–SIMSisausefulapproachtouncoverprimarysignalsintheArcheansulfurcycle.Proceedings of the National Academy of Sciences,111,5468–5473.

Grobbelaar,W.S.,Burger,M.A.,Pretorius,A.I.,Marais,W.,&VanNiekerk,I.J.M. (1995). Stratigraphic and structural settingof theGriqualand

     |  365MEYER Et al.

West and theOlifantshoekSequences atBlackRock,Beeshoek andRooinekkeMines,GriqualandWest,SouthAfrica.Mineralium Deposita,30,152–161.

Guo,Q.,Strauss,H.,Kaufman,A.J.,Schröder,S.,Gutzmer,J.,Wing,B.,…Farquhar,J.(2009).ReconstructingEarth’ssurfaceoxidationacrosstheArchean-Proterozoictransition.Geology,37,399–402.

Habicht, K. S., Gade,M., Thamdrup, B., Berg, P., & Canfield, D. E. (2002).CalibrationofsulfatelevelsintheArcheanocean.Science,298,2372–2374.

Hartzer, F. J. (1995).Transvaal Supergroup inliers:Geology, tectonic de-velopmentandrelationshipwiththeBushveldComplex,SouthAfrica.Journal of African Earth Sciences,21,521–547.

Huizenga,J.M.,Gutzmer,J.,Greyling,L.N.,&Schaefer,M.(2006).Carbonicfluid inclusions inPaleoproterozoiccarbonate-hostedZn-PbdepositsinGriqualandWest,SouthAfrica.South African Journal of Geology,109,55–62.

Izon,G.,Zerkle,A.L.,Zhelezinskaia,I.,Farquhar,J.,Newton,R.J.,Poulton,S.W.,…Claire,M.W.(2015).MultipleoscillationsinNeoarchaeanatmo-sphericchemistry.Earth and Planetary Science Letters,431,264–273.

Johnston,D.T.(2011).MultiplesulfurisotopesandtheevolutionofEarth’ssurfacesulfurcycle.Earth- Science Reviews,106,161–183.

Johnston,D.T.,Farquhar,J.,Wing,B.A.,Kaufman,A.J.,Canfield,D.E.,&Habicht,K.S.(2005).Multiplesulfurisotopefractionationsinbiologicalsystems:Acasestudywithsulfatereducersandsulfurdisproportion-ators.American Journal of Science,305,645–660.

Kamber,B.S.,&Whitehouse,M.J.(2007).Micro-scalesulphurisotopeevi-denceforsulphurcyclinginthelateArcheanshallowocean.Geobiology,5,5–17.

Knoll,A.H.,&Beukes,N.J.(2009).Introduction:InitialinvestigationsofaNeoarcheanshelfmargin-basintransition(TransvaalSupergroup,SouthAfrica).Precambrian Research,169,1–14.

Laspidou,C.S.,&Rittmann,B.E.(2002).Aunifiedtheoryforextracellularpolymericsubstances,solublemicrobialproducts,andactiveandinertbiomass.Water Research,36,2711–2720.

Martini,J.E.J.,Eriksson,P.G.,&Snyman,C.P.(1995).Theearlyprotero-zoicMississippiValley-typePb-Zn-FdepositsoftheCampbellrandandMalmanisubgroups,SouthAfrica.Mineralium Deposita,30,135–145.

Miyano, T., & Beukes, N. J. (1984). Phase relations of stilpnomelane,ferri-annite, and riebeckite in very low-grade metamorphosed iron-formations.South African Journal of Geology,87,111–124.

Nielsen,P.H.,Jahn,A.,&Palmgren,R.(1997).Conceptualmodelforpro-ductionandcompositionofexopolymersinbiofilms.Water Science and Technology,36,11–19.

Noffke,N.(2009).Thecriteriaforthebiogeneicityofmicrobiallyinducedsedimentarystructures(MISS)inArcheanandyounger,sandydeposits.Earth- Science Reviews,96,173–180.

Oduro,H.,Harms,B.,Sintim,H.O.,Kaufman,A.J.,Cody,G.,&Farquhar,J.(2011).Evidenceofmagneticisotopeeffectsduringthermochemicalsulfatereduction.Proceedings of the National Academy of Sciences,108,17635–17638.

Ohmoto,H.(1972).Systematicsofsulfurandcarbonisotopesinhydrother-maloredeposits.Economic Geology,67,551–578.

Ono,S.,Beukes,N.J.,&Rumble,D.(2009).Originoftwodistinctmultiple-sulfurisotopecompositionsofpyriteinthe2.5GaKleinNauteFormation,GriqualandWestBasin,SouthAfrica.Precambrian Research,169,48–57.

Ono,S.,Eigenbrode,J.L.,Pavlov,A.A.,Kharecha,P.,Rumble,D.,Kasting,J.F.,&Freeman,K.H.(2003).NewinsightsintoArcheansulfurcyclefrommass-independent sulfur isotope records from theHamersley Basin,Australia.Earth and Planetary Science Letters,213,15–30.

Paris,G.,Adkins,J.F.,Sessions,A.L.,Webb,S.M.,&Fischer,W.W.(2014).Neoarchean carbonate–associated sulfate records positive Δ33S anomalies.Science,346,739–741.

Pavlov,A.A.,&Kasting, J. F. (2002).Mass-independent fractionationofsulfur isotopes inArchean sediments: Strongevidence for an anoxicArcheanatmosphere.Astrobiology,2,27–41.

Paytan,A.,Mearon,S.,Cobb,K.,&Kastner,M.(2002).Originofmarinebar-itedeposits:SrandSisotopecharacterization.Geology,30,747–750.

Peres,P.,deChambost,E.,&Schuhmacher,M.(2008).CAMECAIMS7f-Geo: Specialized SIMS tool for geosciences.Applied Surface Science,255,1472–1475.

Roerdink,D. L.,Mason, P. R., Farquhar, J., & Reimer,T. (2012).Multiplesulfur isotopes inPaleoarcheanbarites identifyan importantroleformicrobialsulfatereductionintheearlymarineenvironment.Earth and Planetary Science Letters,331,177–186.

Rye,R.O.(2005).Areviewofthestable-isotopegeochemistryofsulfateminerals in selected igneousenvironmentsand relatedhydrothermalsystems.Chemical Geology,215,5–36.

Schieber,J.(1999).Microbialmatsinterrigenousclastics;thechallengeofidentificationintherockrecord.Palaios,14,3–12.

Schieber,J.(2007).Microbialmatsonmuddysubstrates—examplesofpos-siblesedimentaryfeaturesandunderlyingprocesses.Atlas of Microbial Mat Features Preserved within the Siliciclastic Rock Record Atlases in Geoscience,2,117–133.

Schröder,S.,Lacassie,J.P.,&Beukes,N.J.(2006).Stratigraphicandgeo-chemicalframeworkoftheAgourondrillcores,TransvaalSupergroup(Neoarchean–Paleoproterozoic,SouthAfrica).South African Journal of Geology,109,23–54.

Shen,Y.,&Buick,R. (2004).Theantiquityofmicrobialsulfatereduction.Earth- Science Reviews,64,243–272.

Shen,Y.,Buick,R.,&Canfield,D.E.(2001).IsotopicevidenceformicrobialsulphatereductionintheearlyArchaeanera.Nature,410,77–81.

Sumner,D.Y.,&Beukes,N.J.(2006).Sequencestratigraphicdevelopmentof the Neoarchean Transvaal carbonate platform, Kaapvaal Craton,SouthAfrica.South African Journal of Geology,109,11–22.

Ueno,Y.,Ono,S.,Rumble,D.,&Maruyama,S.(2008).Quadruplesulfuriso-topeanalysisofca.3.5GaDresserFormation:Newevidenceformicro-bialsulfatereductionintheearlyArchean.Geochimica et Cosmochimica Acta,72,5675–5691.

VanKranendonk,M.J. (2006).Volcanicdegassing,hydrothermal circula-tionandtheflourishingofearlylifeonEarth:Areviewoftheevidencefromc.3490–3240MarocksofthePilbaraSupergroup,PilbaraCraton,WesternAustralia.Earth- Science Reviews,74,197–240.

Wagner,M.,Roger,A.J.,Flax,J.L.,Brusseau,G.A.,&Stahl,D.A.(1998).Phylogenyofdissimilatorysulfitereductasessupportsanearlyoriginofsulfaterespiration.Journal of Bacteriology,180,2975–2982.

Wilbanks,E.G.,Jaekel,U.,Salman,V.,Humphrey,P.T.,Eisen,J.A.,Facciotti,M.T.,…Orphan,V.J.(2014).Microscalesulfurcyclinginthephototro-phicpinkberryconsortiaoftheSippewissettSaltMarsh.Environmental Microbiology,16,3398–3415.

Zerkle,A.L.,Claire,M.W.,Domagal-Goldman,S.D.,Farquhar,J.,&Poulton,S.W.(2012).Abistableorganic-richatmosphereontheNeoarchaeanEarth.Nature Geoscience,5,359–363.

Zhelezinskaia,I.,Kaufman,A.J.,Farquhar,J.,&Cliff,J.(2014).Largesulfurisotope fractionations associatedwith Neoarcheanmicrobial sulfatereduction.Science,346,742–744.

SUPPORTING INFORMATION

AdditionalSupportingInformationmaybefoundonlineinthesupport-inginformationtabforthisarticle.

How to cite this article:MeyerNR,ZerkleAL,FikeDA.SulphurcyclinginaNeoarchaeanmicrobialmat.Geobiology. 2017;15:353–365.https://doi.org/10.1111/gbi.12227