Earth and Planetary Science Letters 306 (2011) 11–22

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Sedimentary talc in Neoproterozoic carbonate successions

Nicholas J. Tosca a,⁎, Francis A. Macdonald b, Justin V. Strauss b, David T. Johnston b, Andrew H. Knoll b,c

a Department of Earth Sciences, University of Cambridge, Cambridge, CB2 3EQ, UKb Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA 02138, USAc Department of Organismic & Evolutionary Biology, Harvard University, Cambridge, MA 02138, USA

⁎ Corresponding author. Tel.: +44 1223 333442; fax:E-mail address: njt41@cam.ac.uk (N.J. Tosca).

0012-821X/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.epsl.2011.03.041

a b s t r a c t

a r t i c l e i n f o

Article history:Received 13 January 2011Received in revised form 30 March 2011Accepted 31 March 2011Available online 15 April 2011

Editor: P. DeMenocal

Keywords:Proterozoicgeobiologymineralogygeochemistrycarbonatesilica

Mineralogical, petrographic and sedimentological observations document early diagenetic talc in carbonate-dominated successions deposited on two early Neoproterozoic (~800–700 million years old) platformmargins. In the Akademikerbreen Group, Svalbard, talc occurs as nodules that pre-date microspar cementsthat fill molar tooth structures and primary porosity in stromatolitic carbonates. In the upper FifteenmileGroup of the Ogilvie Mountains, NW Canada, the talc is present as nodules, coated grains, rip-up clasts andmassive beds that are several meters thick. To gain insight into the chemistry required to form earlydiagenetic talc, we conducted precipitation experiments at 25 °Cwith low-SO4 synthetic seawater solutions atvarying pH, Mg2+ and SiO2(aq). Our experiments reveal a sharp and reproducible pH boundary (at ~8.7) onlyabove which does poorly crystalline Mg-silicate precipitate; increasing Mg2+ and/or SiO2(aq) alone isinsufficient to produce the material. The strong pH control can be explained by Mg-silica complexingactivated by the deprotonation of silicic acid above ~8.6–8.7. FT-IR, TEM and XRD of the synthetic precipitatesreveal a talc-like 2:1 trioctahedral structure with short-range stacking order. Hydrothermal experimentssimulating burial diagenesis show that dehydration of the precipitate drives a transition to kerolite (hydratedtalc) and eventually to talc. This formation pathway imparts extensive layer stacking disorder to the synthetictalc end-product that is identical to Neoproterozoic occurrences. Early diagenetic talc in Neoproterozoiccarbonate platform successions appears to reflect a unique combination of low Al concentrations (and, byinference, low siliciclastic input), near modern marine salinity and Mg2+, elevated SiO2(aq), and pHN~8.7.Because the talc occurs in close association with microbially influenced sediments, we suggest that solublespecies requirements weremost easily met throughmicrobial influences on pore water chemistry, specificallypH and alkalinity increases driven by anaerobic Fe respiration.

+44 1223 333450.

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Talc, Mg3Si4O10(OH)2, is typically interpreted as a high-tempera-ture mineral that forms from hydrothermal alteration or metamor-phism of Mg-rich and ultrabasic rocks (Evans and Guggenheim, 1988;Marumo and Hattori, 1999). Lower temperature reactions can alsoproduce talc-bearing assemblages, for example through the weath-ering of serpentinite deposits (Velde and Meunier, 2008); however,this class of reactions is far less common than its high temperaturecounterpart. As such, outside of the scattered reports of talc inevaporite or carbonate rich deposits (Bodine, 1983; Braitsch, 1971;Calvo et al., 1999; Friedman, 1965; Millot and Palausi, 1959; Noacket al., 1989), talc is rarely reported as a sedimentary mineral. Here wedocument the unusual occurrence of early diagenetic talc associatedwith Neoproterozoic (~800–700 Ma) carbonates deposited on twoseparate platform margins: the Akademikerbreen Group in Svalbard

and the Fifteenmile Group in the Ogilvie Mountains of northwesternCanada. The formation of sedimentary talc and its mineralogicalprecursors requires a specific set of chemical conditions; its presenceplaces tight quantitative constraints on Neoproterozoic ocean chem-istry and provides additional insight into the biogeochemistry ofmarine sediments at that time.

At low temperatures (i.e., less than ~30 °C), theMg-silicate system iscontrolled mainly by kinetic phenomena (Evans and Guggenheim,1988; Jones, 1986; Wollast et al., 1968), and so constraints on earlydiagenetic chemistry are difficult to derive based on thermodynamicsalone. To address this problem, we designed a series of experiments toevaluate the effects of changingpH,Mg2+andSiO2(aq) on the formationof Mg-silicates from low-SO4 (~2.8 mmol/kg), Neoproterozoic-likeseawater solutions. When these experiments and companion modelingresults are coupledwith stratigraphically constrained geochemical data,we are able to posit that the early diagenetic talc in Neoproterozoiccarbonate successions reflects a unique combination of low siliciclasticinput, nearmarine salinity andMg2+, elevated SiO2(aq), and pH greaterthan ~8.6–8.7. In what follows, we discuss: (1) the sedimentology,geochemistry and mineralogy of the Akademikerbreen and Fifteenmile

Fig. 1. Bedded and nodular talc interbedded with microbialite in the Callison LakeDolostone, Ogilvie Mountains, Yukon.

12 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

talc occurrences, (2) experimental constraints onMg-silicate formationfrom modified Neoproterozoic-like seawater solutions, and (3) reportsof similar mineral assemblages from other Neoproterozoic successions.Finally, we consider both the specific constraints these data place on the

Fig. 2. Stratigraphic sections of the Fifteenmile a

chemistry of waters bathing Neoproterozoic carbonate platforms andwhy this interval in Earth history may have favored sedimentary talcformation.

2. Geologic setting

Sedimentary talc occurs in Neoproterozoic strata that crop outdiscontinuously over a distance of more than 140 km in the Coal Creekand Hart River inliers of the Ogilvie Mountains, Yukon (for exactlocations see Macdonald and Roots, 2009). The talc is stratigraphicallyconfined to the lower Callison Lake Dolostone (unit PF2 of theFifteenmile Group), which consists predominantly of dolomite withinterbedded shale (Figs. 1 and 2) (Macdonald et al., 2010; Macdonaldand Roots, 2009). A tuff within the lowermost map unit of theFifteenmile Group provides a maximum age constraint on the talc of811.51±0.25 Ma; a minimum constraint comes from a 717.43±0.14 Ma quartz–phyric rhyolite flow in the overlying Mount Harpervolcanic complex (Macdonald et al., 2010).

On a ridge ~10 km to the northwest of Mt. Harper, the lowerCallison Lake Dolostone is well exposed and measures 77 m thick(Macdonald and Roots, 2009). Here, these strata rest unconformablyon a brecciated surface of PF1 platformal carbonate and consist of abasal 2 m of channelized, sub-rounded quartz gravel conglomerate.These beds are overlain by 17 m of green and red shale and siltstone

nd Akademikerbreen talc-bearing deposits.

Fig. 3. Void-filling talc nodules (in outcrop (A) and thin section (B) under cross-polarized light) formed in dolomitic stromatolite of the Hunnberg Fm, Nordaustlandet,Svalbard. The talc in (B) (orange-brown) appears penecontemporaneous with void-filling microspar cement.

13N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

with dolomite lenses containing microbialite textures. The micro-bialite is draped with a ~10 cm thick bed of hematite iron formation.Above the varicolored shale and siltstone is an additional 60 m ofblack shale with laterally discontinuous dolomitic bioherms andabundant microbialite. However, the black shale has a peculiar waxyluster, reminiscent of phosphorite, and XRD analyses indicate that thisshale consists almost exclusively of talc. The black talc-bearing shalealso contains lenses of redeposited talc rip-up clasts suspended infine-grained dolomite matrix and abundant black chert nodules.These strata are succeeded gradationally by over 300 m of silicifieddolostone (Unit PF3), dominated by stromatolites, microbialaminite,and edgewise conglomerate with cm-sized coated grains, occasionalexposure surfaces, and patchy silicification. These facies all suggestdeposition in a shallow-water, episodically exposed, marginal marineenvironment.

The talc is stratigraphically confined and laterally persistent alongthe outcrop belt for over 30 km to the east to Mt. Gibben, and occursagain an additional ~110 km to the east in the Hart River inlier(Abbott, 1997). This lateral extent suggests that talc deposition was atleast a basin-wide phenomenon. The Fifteenmile Group has beencorrelated with the Little Dal Group in the Mackenzie Mountains(Macdonald et al., 2010), which has been interpreted as a marginalmarine carbonate bank deposited on a rifted passive margin (Turnerand Long, 2008). However, the significant unconformity recentlyidentified at the base of the Callison Lake Dolostone suggests that itmay have formed in an additional distinct basin-forming episode(Macdonald and Roots, 2009), perhaps coeval with the Coates LakeGroup, which lies stratigraphically above the Little Dal Group. TheCoates Lake Group was deposited in narrow, restricted grabens(Jefferson and Parrish, 1989). Although neither evaporitic mineralsnor pseudomorphs have been identified in the Fifteenmile Group,evaporites are present in both the Little Dal and Coates Lake groups(Jefferson and Parrish, 1989), and we cannot rule out restriction onthe platform.

In Svalbard, mm to cm scale talc nodules occur within shallowmarine carbonates of the Hunnberg Formation, Nordaustlandet, andits lateral equivalents in Spitsbergen, the upper Grusdievbreen andSvanbergfjellet formations (Figs. 2 and 3; Knoll and Swett, 1990). As inCanada, the talc deposits occur in settings ranging from coastal tosubtidal environments below storm wave base and can be tracedalong strike for several hundred kilometers, suggesting a basin-scalephenomenon. Also as observed in Canada, cm-scale hematite layersdrape microbial dolomites in the talc-bearing interval. In theSvanbergfjellet Formation, talc forms cm-scale nodules withinrelatively deep carbonaceous shales and carbonates. More extensiveand striking, however, are rounded nodules that formed within molartooth structures and primary voids in microbialites, in both casesbefore penecontemporaneous void-filling microspar cements weredeposited (Fig. 3; Knoll, 1984, who mistakenly identified the talcnodules as phosphorite).

No radiometric dates closely constrain the age of talc-bearingSvalbard carbonates, but C and Sr isotopic chemostratigraphy andbiostratigraphy suggest an age broadly comparable to that of theOgilvie succession (Halverson et al., 2007; Knoll et al., 1986;Macdonald et al., 2010).

3. Analytical and experimental methods

Our sample suite was selected from the lower Callison LakeDolostone of the Fifteenmile Group and from throughout theAkademikerbreen Group (Fig. 2; Table 1). The Fifteenmile Groupsamples, all from outcrop, include dolostone hosting nodular talc,bedded talc deposits, and a small horizon of iron formation. With theexception of some visible surface staining, late stage alteration/oxidation of the samples generally appears minor. Samples from theAkademikerbreen Group are predominantly from the Svanbergfjellet

Formation and correlative horizons in the Hunnberg Formation,where nodular talc was identified in the field, but they also includecarbonates and shales from units above and below this interval. AllSvalbard samples were collected from outcrop and were chosen tocapture the range of carbonate lithofacies reported in Knoll and Swett(1990). Again, late stage alteration/oxidation is minor in Svalbardsamples, as documented by geochemical screens (e.g., Sr concentra-tion, Mn/Sr, Sr/Ca and δ

18O; Derry et al., 1989).

Mineralogical analyses included X-ray diffraction (XRD) of b2 μm(and in some cases, b0.2 μm) oriented aggregates from decarbonatedsamples, and bulk XRD of unfractionated samples. Petrographicanalysis and electron microprobe analysis of selected samples wereperformed using polished, carbon-coated thin sections. Further detailsof sample preparation and analysis are given in the Supplementaryfile.

Precipitation experiments were performed by the addition of SiO2

(aq) to 1 kg batches of synthetic low-SO4 seawater at variousMg2+(aq)concentrations and pH. Experiments were run in water baths at 25±0.1 °C for a minimum of 4–5 weeks to a maximum of 7 months,depending on conditions. Filtered aqueous samples were collectedperiodically and analyzed by ICP–AES, and residual solid precipitateswere collected at experiment termination, washed and analyzed byXRD, FT-IR and TEM. Selected solid precipitates were extracted andreacted with deionized water at 180 °C and 400 °C to evaluatemineralogical changes in response to heating/dehydration. Further

Table 1Sample details, locations and b2 μm mineralogy.

Sample ID Age (Ma) Formation/unit Location Illitea I/Sb Kaol Chl Green Talc Corr Sap C/S

F927-3.8 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada X – – – – X – – –

F927-16 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – – – – – X – – –

F927-27 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – X – – – X – – –

F927-58 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – – – – – X – – –

J908-25 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada X X – – – – – – –

J908-65.4 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada X – – – – X – – –

F926-16.6 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – – – – X – – – –

F926-17.1 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – – – – X – – – –

F926-18.1 811.5–717.4 Callison Lake DL Ogilvie Mts; NW Canada – – – – X – – – –

86-M-31 ca 800–700 Backlundtoppen Spitsbergen, Svalbard X – X – – – – – –

81-P-4984 ca 800–700 Backlundtoppen Spitsbergen, Svalbard X – – – – − − − −P-4843 ca 800−700 Backlundtoppen Spitsbergen, Svalbard X X − − − − − − −86-M-15 ca 800–700 Draken Spitsbergen, Svalbard X X X − − − − − −81-P-4187 ca 800–700 Draken Spitsbergen, Svalbard X X X – – – – – –

81-P-4353 ca 800–700 Draken Spitsbergen, Svalbard X X X − − − − − −81-P-4378 ca 800–700 Draken Spitsbergen, Svalbard X X X − − − − − −86-P-9 ca 800–700 Draken Spitsbergen, Svalbard X X X – – – – – –

SV-1646 ca 800–700 Draken Spitsbergen, Svalbard X X X – – – – – –

K2016 ca 800–700 Hunnberg Nordaustlandet, Svalbard – – – – – X – X XHU 1246 ca 800–700 Hunnberg Nordaustlandet, Svalbard – X – – – – – X –

86-G-1 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – – – – – X – – –

86-G-3 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – – – – – X X – –

86-G-7 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – X X – – X – – –

86-G-11 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – X X – – – – – –

86-G-32 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – X – – – – –

86-G-41 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – – – – – – –

82-Sv-465 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X – – – – – – – –

86-SP-4 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – – – – – – –

86-SP-8 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – X – – – – – – –

86-P-2664 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – – – X – – –

86-P-73 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – – – – – X – – –

86-P-78 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard – – – – – X – – –

SV-1 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – X – – – – –

86-G-28 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X X X – – – – –

86-G-34 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X – – – – – – –

86-G-63 ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X X X – – – – – –

L12B ca 800–700 Svanbergfjellet Spitsbergen, Svalbard X – X – – – – – –

86-P-66 ca 800–700 Grusdievbreen Spitsbergen, Svalbard X – – – – – – – –

Kaol: kaolinite; Chl: chlorite; Green: greenalite; Corr: corrensite; Sap: saponite; C/S: chlorite/smectite.a All illite is discrete illite and/or muscovite.b All I/S is approximately equal to illite 0.7/smectite.

14 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

details of the experimental and analytical procedures are given in theSupplementary file.

4. Mineralogy and petrography of Neoproterozoic talc

4.1. Fifteenmile group

Nodular talc samples from the Callison Lake Dolostone have threecomponents in varying proportions: dolomite, chert, and talc. Detritalsiliciclastic input is apparent only in the lower 40 m of varicoloredshale and siltstone with interbedded microbial dolomite (Fig. 2). Thethin horizon (~10 cm) of iron formation that drapes the microbialiteis mineralogically simple, consisting of euhedral hematite grainsdispersed in a dolomite matrix, with minor silicification and raredetrital quartz grains. b2 μm size fractions reveal a single 7 Å kaolin–serpentine group phase with trioctahedral occupancy, consistent withgreenalite. The dolomite is present both as primary microbiallaminations, coated grains, and as interstitial microspar cement.Talc both drapes and fills laminae within the microbialites and,along with the black chert, forms early diagenetic nodules. Beddedchert is also common. In silicified dolostones of the Callison LakeDolostone that host nodular talc, carbonate mineralogy consistsexclusively of dolomite, with no calcite indicated from powderXRD. Aside from talc, the clay mineralogy of decarbonated samplesindicates lesser amounts of either discrete highly crystalline illiteand/or mixed layer illite/smectite. The composition of the mixed

layer phases is highly consistent from sample to sample, and analysesafter ethylene glycol treatment are consistent with R0 (disordered)illite(0.7)/smectite.

The bedded talc horizons consist predominantly of black, talc-richshale. Thin lenses of sub-angular to sub-rounded rip-up clasts of talcsupported by very fine-grained dolomite matrix are interbedded withthe talc horizons. The interstitial dolomite cements between thereworked talc clasts suggest that at least some of the talc formationpredates the formation of these particular cements. The mineralogy ofbedded talc samples is also simple; talc overwhelms bulk analyses andclay fractions in abundance. However, two samples of bedded talcrevealed subordinate amounts of discrete highly crystalline illite andmixed layer R0 illite (0.7)/smectite (Table 1). In contrast to bulksamples, oriented b2 μm size fractions of bedded talc indicate that nomixed layering is apparent. In other words, the talc is present as adiscrete mineral phase. No other Mg-silicates were identified in anybulk or b2 μm fractions from Fifteenmile Group samples (Fig. 4).Randomly oriented specimens of the b2 μm clay fraction furtherindicate that the talc exhibits a high degree of stacking disorder; it isessentially turbostratic. A number of hkl peaks are significantlymodulated, particularly in the 20–25° and 30–45° ranges. This typeand extent of disordering is present in all of the Fifteenmile talcoccurrences.

The carbonate overlying the talc horizons consists of silicifieddolomite with microbial lamination, coated grains, and abundantstromatolites. Void-filling cements consist predominantly of dolomite

Fig. 4. XRD patterns of oriented aggregates of the b2 μm fraction extracted from talc-bearing rocks. Black lines indicate scans acquired in the air-dried (Ca-saturated) stateand red lines indicate scans acquired in the EG-solvated state. (A) Talc from unitPF2 (U. Fifteenmile Group; F927-58 m) (B) Talc, saponite, and chlorite/smectite fromthe Hunnberg Fm (K2016). (C) Saponite and illite/smectite from the Hunnberg Fm(HU1246). (D) Trioctahedral low-charge corrensite from the Svanbergfjellet Fm (86-G-3).

15N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

and silica. Decarbonated samples from these strata also yielded talcwith lesser amounts of discrete illite andmixed layered illite/smectite.The talc in these samples is distributed throughout the carbonatematrix and not as nodules or other conspicuous sedimentary featuresotherwise resolvable by petrographic analysis.

4.2. Akademikerbreen group

Akademikerbreen talc occurs as mm- to cm-scale nodules anddisseminated particles in the b2 μm fraction of the carbonate matrix.

Talc nodules fill molar tooth structures, primary void space inmicrobialites and intergranular space, predating precipitation ofpenecontemporaneous microspar cement. In general, nodules arepetrologically homogeneous; there are no discernable nuclei nor ischemical zoning evident (Fig. 3).

Powder XRD of isolated and cleaned nodules handpicked fromAkademikerbreen carbonates shows the same type and extent of layerstacking disorder as in the Fifteenmile talc samples (Fig. 5). That is, theAkademikerbreen talc is also highly turbostratic. XRD patterns againdisplay modulated and/or extinguished hkl reflections in the 20–25and 30–45° 2Θ ranges. Aside from talc, the carbonate mineralogyfrequently includes ankerite in addition to dolomite and lesseramounts of calcite.

The clay mineralogy of decarbonated Akademikerbreen samplesshows more variation among the Mg-silicates. Analyses of decarbo-nated residues reveal talc in the carbonate matrix in some sampleswhere nodules are not apparent. The talc again shows no evidence ofmixed layering or occurrences with other low temperature Mg-silicates such as sepiolite or palygorskite. However, saponite (a Mg-rich trioctahedral smectite) and its burial diagenetic equivalents,mixed layered chlorite/smectite and corrensite, have all beenidentified (Fig. 4; Table 1; Supplementary file).

In addition to Mg-silicates, Akademikerbreen clay fractionscontain varying amounts of discrete highly crystalline illite, illite/smectite (most of which corresponds to R1 illite(0.7)/smectite),kaolinite, and, in one sample only, chlorite. In general, the resultssummarized in Table 1 show that samples containing talc and otherassociated Mg-silicates consistently lack illite, I/S, and kaolinite. Thereverse is also true: aluminous clay mineral assemblages are rarelyaccompanied by talc.

The Svanbergfjellet, Draken and Backlundtoppen Formationsalso contain discrete siliciclastic horizons that represent detritalpulses into the carbonate platform (Butterfield et al., 1994; Knoll andSwett, 1990). The clay mineralogy of these horizons is dominated byillite and kaolinite in the b2 μm fraction; the b0.2 μm fraction iscomposed almost entirely of discrete illite, with significant defectbroadening.

4.3. Fifteenmile and Akademikerbreen group Mg-silicate occurrences:summary

Sedimentology and petrology indicate that talc formed duringearly diagenesis, perhaps penecontemporaneously, in both theAkademikerbreen and the Fifteenmile Groups. In both environmentsthe clay mineralogy also indicates that talc generally occurs inenvironments that received low background siliciclastic flux, andwhere alumino-silicates are present, talc has not been identified.

Saponite, chlorite/smectite and corrensite in Akademikerbreencarbonates, all indicate that saponite formed during early diagenesisand was transformed to chlorite-rich mineral species upon burial andlate-stage alteration. Because there is little evidence for mafic input tothe Akademikerbreen platform, saponite formation suggests that theinitial deposition of smectite (originally as a dioctahedral smectitesuch as montmorillonite), or some other detrital precursor, initiated arecrystallization/transformation to trioctahedral saponite, likely inresponse to favorable aqueous chemistry associated with theplatformal carbonates (discussed below).

Samples from the Fifteenmile Group, on the other hand, yield verylittle decarbonated residue and show little variation in backgroundsiliciclastic component; they are dominated by highly crystalline illite(muscovite) and illite/smectite. Together this implies very lowbackground detrital fluxes and negligible input of terrigenous claybeyond detrital muscovite.

The sedimentology, petrography and mineralogy of these depositsare indicative of unusual aqueous chemistry associated with platfor-mal carbonates that facilitated the early diagenetic precipitation of

Fig. 5. Representative powder XRD patterns of randomly-oriented experimentalprecipitates formed by (A) low temperature precipitation from modified seawater,(B) heating a low temperature precipitate at 180 °C for 4 weeks, (C) heating theprecipitate from (B) at 400 °C for 4 days. For comparison, a cleaned talc nodule from theSvanbergfjellet Fm is shown (D), indicating extensive layer stacking disorder. “C” =cristobalite; “D” = dolomite.

16 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

Mg-silicates, whether talc or a precursor phase, from solution. Whatdrove these reactions during Akademikerbreen and Fifteenmile Groupdeposition?

5. Constraints from Mg-silicate precipitation experiments

To better understand what chemical environment is required toinitiate the precipitation of Mg-silicates, we performed a systematicseries of experiments interrogating the role of SiO2(aq) and pH ontalc formation. We incubated one liter batches of standard, “modern”synthetic seawater at various SiO2(aq) concentrations and pH,monitoring the bottles for both precipitation of solid phases andchanges in aqueous chemistry. In certain experiments, a visibleprecipitate formed within the first 48 h. The precipitate often cloudedthe solution before flocculating and, eventually, settled to the bottom of

the reaction vessel. In other experiments, no precipitate was formedover N7 months.

When the experimental conditions favored precipitation, XRDanalyses on randomly oriented powders showed broad and poorlyresolvable hkl bands and no 001 reflections (Fig. 5). All XRD analysesshowed weak bands at d=1.55 to 1.52 Å, indicating the presence oftrioctahedral layers, and all hkl bands were reproducible fromexperiment to experiment, displaying little variation in their relativeintensity and d-spacing. In only one case, the experiment conducted atthe highest pH, a precipitate gave broad low-angle reflectionscorresponding tod-spacings of 28.7, 13.9 and 10.4 Å.With the exceptionof the highest pH experiment, the structure of the precipitate is bothinsoluble and reproducible from experiment to experiment. Perhapsmost striking, however, is the observation that the precipitate onlyformed at a pH of 8.66 and above, implicating pH as a major control onMg-silicate precipitation.

Although XRD data show poorly crystalline products, FT-IR spectraindicate that the experimental precipitates are almost identical to talc.The close correspondence between Mg–O vibrations from the pre-cipitates and lattice vibrations from talc (including the diagnostic535 cm−1 (Mg–O)–OH feature (Farmer, 1974; Zhang et al., 2006))reflect a similar bonding environment for Mg (Fig. 6). Absorptions at~1020 cm−1 show the samebonding arrangement for SiO4 tetrahedra: a2:1 layered configuration of tetrahedral silicate layers and octahedralMgO6 layers (Fig. 6). A trioctahedral occupancy (consistent with XRDpeaks at 1.55–1.52 Å), manifested by the OH-stretching vibration at~3676 cm−1, is also evident for both the precipitates and for talcstandards. However, the lack of an observable basal 001 reflection inoriented XRD patterns suggests that there is little to no stacking orderbetween TOT sheets, perhaps due to variable interlayer/surfacehydration (consistent with absorptions at ~1630 and ~3400 cm−1

indicating appreciable boundmolecular H2O). Taken together, XRD, FT-IR and TEM data (shown in the Supplemental file) indicate theexperimental precipitates are simply a hydrated, disordered form oftalc; they are composed of 2-dimensional talc TOT sheets not neatlystacked as in true talc, but organized similar to a pile of bricks looselycemented with water of hydration.

To ascertain the effects of mild heating on poorly crystalline Mg-silicate precipitated from modified seawater solutions, we alsoperformed experiments involving the hydrothermal treatment ofpoorly crystalline Mg-silicate (the products of precipitation experi-ments) in the presence of deionized water. After 4 weeks of reactionat 180 °C, the poorly crystalline Mg-silicate transformed to kerolite(Mg3Si4O10(OH)2∙nH2O), a distinct mineral phase and hydratedstructural analog of talc (Fig. 5). The final product exhibited anapparent d(001) of 9.9 Å and hkl peaks that correspond to the kerolitestructure discussed in Brindley et al. (1977). Combined XRD, FT-IR andTEM data indicate that simple dehydration – in this case, in responseto temperature increase – drives the transformation from poorlycrystalline Mg-silicate to the mineral kerolite. Further heating atelevated temperature or over long timescales in turn causesadditional dehydration and experiments conducted at 400 °C and1 kbar show that this dehydration drives the transition from keroliteto talc (and minor cristobalite) (Fig. 5). The talc produced from thisprocess exhibits significant turbostratic ordering and powder diffrac-tion data show similar peak modulation to Neoproterozoic talcsamples discussed above.

6. Discussion

6.1. Mg-silicate system at low temperature

Present-day seawater is supersaturated with respect to crystallinetalc, yet modern and even ancient environments lack authigenic talc(Fig. 7). Like in the case of carbonate precipitation, there must exista kinetic barrier to talc precipitation; a barrier that we can address

Fig. 6. FT-IR spectra of experimental Mg-silicate precipitates (black) in comparison to talc collected from the Svanbergfjellet Fm (red). (A) Lattice vibration region, (B) Si–Ostretching region, (C) H2O and CO3 region, (D) OH stretching region.

17N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

through experimental interrogation. Our data and ensuing discussionallow talc precipitation to be reconstructed in an environmentallytuned, step-wise fashion.

Previous experimental work has targeted talc precipitation frommodern-like seawater, and presents contrasting results to thosedescribed herein. Wollast et al. (1968) spiked filtered modernseawater with sodium metasilicate (Na2SiO3:9H2O) at 25 °C andprecipitated a Mg-silicate phase at pHN8.3. The Mg-silicate wasdescribed as “poorly crystalline sepiolite” on the basis of FT-IR andXRD — a phase not identified in any of our experiments. In addition,the apparent solubility calculated from their experiments is lowerthan the experiments described here (Fig. 7), implying a lower degreeof supersaturation needed for precipitation. Upon closer inspection,however, the Mg-silicate precipitated by Wollast et al. (1968) is notinconsistent with our results. In fact, their FT-IR and XRD analyses alsolack the most diagnostic features of sepiolite, including the 12 Å and7.5 Å diffraction peaks and a Si–O feature in FT-IR attributed to Q3

bonding of silica tetrahedra in chain silicates (Russell and Fraser,1994). However, their material also appears to exhibit FT-IR featuresthat ours does not.

Aside from the identity of the initial precipitate, there also remainsa clear discrepancy in the literature in the aqueous conditions neededto initiate precipitation. To address this, we duplicated some of theexperiments conducted by Wollast et al. (1968) using a synthetic“modern” seawater solution at elevated SiO2(aq). The only differenceis the nature of the silica source: TEOS solutions in experimentsdescribed here versus sodium metasilicate solutions in Wollast et al.(1968). The results show that Mg-silicate formation does not occuruntil higher levels of supersaturation than those reported by Wollastet al. (1968) and that the initial precipitate is a known keroliteprecursor. We argue that the difference in silica source may well becritical. We have taken precautions to ensure that SiO2(aq) added toseawater solutions was equilibrated in the monomeric form prior tothe addition of MgCl2 (Dietzel, 2000; Iler, 1979). In contrast, sodiummetasilicate solutions are known to polymerize rapidly and formcolloidal material upon dilution (Iler, 1979), perhaps in responseto decreasing silica levels and/or rapid changes in pH associatedwith their preparation. Thus, it is possible that the addition ofmetasilicate solutions to seawater solutions co-precipitated silicawith Mg2+ in a way that may not be truly representative of natural

Fig. 7. Solubility diagram showing the equilibrium solubilities of crystalline talc (Jones, 1986), kerolite and sepiolite (Jones and Galan, 1988), “amorphous sepiolite” from Wollast et al.(1968) and the equilibriumboundary betweenmontmorillonite and saponite fromWeaver andBeck (1977). Data points represent apparent solubilities calculated from solutions collectedupon experiment termination. Experiments that have resulted inMg-silicate precipitate form an apparent boundary approximately parallel to kerolite and/or sepioilite, but these data donot represent true equilibrium conditionswith respect to either of these crystalline phases. Green point represents experiment at elevatedMg and pHf=8.0where no precipitate formed.

Fig. 8. SiO2(aq) loss as a function of final pH showing the sharp pH control on Mg-silicate precipitation. Process blank experiments conducted with no MgCl2(aq) indicatethat one experiment (pHf ~9.2) has resulted in SiO2(aq) loss that cannot be explainedby experimental artifacts alone and suggests the formation of a precipitate that wasunable to be extracted for analysis.

18 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

surface environments. This is a complication avoided in our experi-ments through the use of TEOS as a SiO2(aq) source.

6.2. Mg-silicate formation: pH as a master variable

To assess what special combination of Mg2+, SiO2(aq) and pHresults in Mg-silicate precipitation, we ran experiments varying Mg2+

concentrations (by 10 times over modern seawater) and at variablepH (Fig. 7). As discussed above, raising SiO2(aq) alone was insufficientto drive precipitation. In fact, our experiments identify pH as theprimary control on Mg-silicate formation. At a pHf of 8.0, we observedno precipitation over the course of the experiment. Within error, SiO2

(aq) levels remained unchanged for N5 months with no sign ofprecipitation. In contrast, the same Mg2+ and SiO2(aq) at elevated pHresulted in the rapid precipitation of Mg-silicate on the timescale ofdays, suggesting that despite exceedingly high Mg2+ levels, pHappears to be the primary control on Mg-silicate formation fromseawater-derived solutions.

Solution chemistry at experiment termination tells a similar story.Relationships between SiO2(aq) loss and pHf reveal a sharp specia-tion boundary for SiO2(aq) at ~8.6–8.7 (Fig. 8). Below this thresh-old, SiO2(aq) is unchanged within analytical error (compared to“blank” experiments conducted in the absence of MgCl2). At SiO2(aq)concentrations between amorphous silica and quartz saturation, thedominant silica species in seawater is monomeric silica; H4SiO4

0.As pH is increased above ~8.5, H4SiO4

0 deprotonates, favoring H3SiO4−

(Fig. 9). As a result, at these higher pH values, the concentration of aMg-silica complex (MgH3SiO4

+) increases sharply. The MgH3SiO4+

species alone is probably not representative of all possible aqueous

Fig. 9. Thermodynamic calculations of SiO2(aq) species distribution in “Neoproterozoic-like” seawater as a function of pH. The sharp pH control on the MgH3SiO4

+ speciesrepresents the effect of pH on potential nucleation of Mg-silicate from seawater atelevated SiO2(aq). Initial SiO2(aq)=60 mg/kg and initial SO4T=2.8 mmol/kg. Chargebalance was satisfied by readjustment of CO2–HCO3–CO3 equilibria in an open systemwith respect to CO2(g). All mineral precipitation was suppressed.

19N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

Mg-silica complexes in solution, especially given the complexity ofsilica speciation revealed by NMR (Felmy et al., 2001). Nevertheless,the stability constant for this species (derived from actual potentio-metric measurements of SiO2(aq) andMg-bearing solutions at variouspH and high salinity) shows that regardless of actual speciesdistribution, the pH dependence of Mg-silica complexing is wellrepresented by MgH3SiO4

+ (Santschi and Schindler, 1974).It is possible that at low pH, the rate of crystallization is too slow to

be observed. However, the idea of a minimum pH for Mg-silicateformation is well supported by previous synthesis experiments. Forexample, Siffert and Wey (1962), using dilute MgCl2 solutions andmonomeric SiO2(aq), precipitated “true” sepiolite at pH N8.5. At pHN9, talc and stevensite formed. Abtahi (1985) and La Iglesia (1978)obtained similar results. The same trends have been observed byworkers synthesizing “Mg-silicate hydrate” for industrial applications,with some citing a canonical pH for “Mg-silicate scale” formation of~8.5 (Brew and Glasser, 2005a, 2005b; Delacaillerie et al., 1995;Mizutani et al., 1990; Packter, 1986; Strese and Hofmann, 1941;Takahashi et al., 1994; Wei and Chen, 2006).

MgH3SiO4+ and similar species are, in turn, the solution precursors

to Mg-silicate precipitates (Packter, 1986). Theoretical treatments oflayer silicate nucleation from solution require that an excess surface

energy barrier be overcome by a critical supersaturation of a pre-existing solution complex (Carrado et al., 2006). The criticalsupersaturation barrier to initiate Mg-silicate formation, then, mustoccur close to pH 8.6–8.7. Below this point, speciation calculationsshow that aqueous Mg-silicate complexes, while present, are unlikelyto be concentrated enough to initiate the nucleation of a silicatephase; deprotonated silica species would be kept to scant amounts insolution, preventing appreciable interaction with Mg2+. This strongkinetic control onMg-silicateprecipitation fromseawater is reminiscentof the more familiar problem of CaCO3 precipitation from modernsurface seawater. For both systems, despite a clear thermodynamicdriving force, a critical level of supersaturation is required. For Mg-silicates, deprotonation of H4SiO4

0 at elevated pH appears to be the“switch” that enables Mg-silica complexing, in turn driving Mg-silicatenucleation and precipitation.

6.3. Burial diagenesis and turbostratic stacking disorder in talc

The sedimentary record of authigenic Mg-silicates can be relatedto solution chemistry only with knowledge of thermal transforma-tions possible during burial diagenesis. Our hydrothermal experi-ments run with poorly crystalline Mg-silicates precipitated fromseawater show that upon dehydration, the material readily trans-forms to kerolite. Previous work has shown that with continueddehydration, either with time or at increased temperature, kerolite,in turn, transforms to crystalline disordered talc (Mitsuda andTaguchi, 1977). Our heating experiments with kerolite confirmthese observations and show that the talc formed by dehydratingkerolite exhibits severe layer stacking disorder identical to thedisorder observed in Akademikerbreen and Fifteenmile samples(Fig. 5).

Stacking disorder in talc is not uncommon. In fact, it results froman inherently weak electrostatic charge between 2:1 layers (Giese,1975). The disorder observed in our samples is best matched byordered stacking of 2:1 layers by a/3 along one of the pseudo-hexagonal axes, with random displacements of ±b/3 along another(Gualtieri, 1999). It is not clear why this type of disordering results,but as others have noted, it may well be controlled by the pathway ofcrystallization (in this case, dehydration) rather than solid statetransformation (Baronnet, 1992; Kogure and Kameda, 2008). Specif-ically, the 2:1 layer dehydration mechanism of the low temperatureprecipitate, which initiates stacking to form kerolite and, in turn, talc,may well exert the most control on the type of disorder that resultsfrom this pathway. Nevertheless, this poorly understood feature ispresent in all Neoproterozoic samples, consistent with low temper-ature precipitation and subsequent dehydration during burial.

Although there are no direct temperature constraints on thesamples described herein, Neoproterozoic strata in the Yukon arerelatively low-grade for rocks of their age. Regionally, conodontalteration indices and vitrinite reflectance from the overlyingPaleozoic strata suggest these rocks have not seen more than 140 °C(Van Kooten et al., 1997). A similar thermal history of the Svalbardsamples is suggested by paleomagnetic studies in which demagneti-zation experiments removed a Caledonian thermochemical remnantmagnetization at 150 °C (Maloof et al., 2004). Moreover, the orange-brown color of preserved organic matter in the AkademikerbreenGroup indicates that the degree of graphitization andmaximumburialtemperatures were modest. In reconciling these constraints withexperimental data discussed above, it is important to note that wehave used elevated temperature (up to 400 °C) purely to observe thedehydration of poorly crystalline Mg-silicate to kerolite and eventu-ally to talc on convenient laboratory timescales. However, thistransition is known to occur in response to lower temperature andlonger reaction time (Mitsuda and Taguchi, 1977). Indeed, if thekinetics of the Mg-silicate dehydration reaction were to behave in asimilar fashion to other burial diagenetic reactions involving clay

20 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

minerals, the dehydration of poorly crystalline Mg-silicate to talccould, for example, take as little as 105 years at ~80 °C (Velde, 1985).

6.4. Mg-silicate formation on Neoproterozoic carbonate platforms

Taken together, experimental data indicate that low temperaturetalc formation in association with carbonates requires: (1) elevatedSiO2(aq), between quartz and amorphous silica saturation; (2)sufficient (at least “modern”) Mg2+; (3) near marine salinity; (4)low Al (i.e., a low background detrital flux, see below) and (5)elevated pH of at least ~8.7.

One key requirement for talc precipitation that we have yet todiscuss is low Al, requiring that detrital fluxes were low duringdeposition of the talc-bearing intervals in the Akademikerbreen andFifteenmile groups. As Al3+ increases in concentration, talc, anominally Al-free mineral, is no longer stable, being replaced insteadby Al-bearing smectites, chlorites, palygorskite or other mineralsdepending on local pore water chemistry and sediment supply (Jones,1986; Jones and Galan, 1988; Weaver and Beck, 1977). Low detritalfluxes have been episodic features of carbonate platform environ-ments throughout recorded Earth history, and don't place a uniquechemical constraint beyond that discussed. Thus, our mineralogicalobservations of low siliciclastic flux in talc-bearing facies of theAkademikerbreen and Fifteenmile successions help to explain thedistribution of early diagenetic talc in space but not in time.

The requirement of elevated SiO2(aq) concentration was alsoreadily met in Neoproterozoic seawater. Indeed, prior to the evolutionof silica skeletons in sponges, radiolarians and diatoms, Precambrianseawater should have carried higher silica concentrations relative tomodern levels (Maliva et al., 2005). Consistent with this expectation,early diagenetic chert in Neoproterozoic carbonate successions,largely from peritidal or possibly mildly evaporitic environments,provides evidence for local pore water enrichment in SiO2(aq) onshallow water platforms (Maliva et al., 2005). The AkademikerbreenGroup abounds with evidence for early diagenetic silicification (e.g.,Fairchild et al., 1991; Knoll, 1984; Knoll and Swett, 1990). FifteenmileGroup samples are also heavily silicified, but record a more complexhistory of silicification (Mustard and Donaldson, 1990).

The Akademikerbreen and Fifteenmile Groups were likely depos-ited on marginal to isolated marine platforms (Knoll and Swett, 1990;Macdonald et al., 2010). Near marine salinity (~35‰) and Mg2+

concentrations are consistent with these settings, yet the salinity andMg2+ content of Proterozoic seawater is poorly constrained. Oolites inthe Akademikerbreen Group provide petrographic and geochemical(high Sr) evidence for an aragonite sea at the time of their deposition,suggesting that contemporaneous seawater may have had Mg2+

concentrations broadly similar to those observed today (Swett andKnoll, 1989). Although no evaporite minerals or their associatedpseudomorphs have been identified in either succession, episodicrestriction in the Akademikerbreen and Fifteenmile Groups cannot beruled out. On the other hand, seawater evaporation should decreasepH (e.g., McCaffrey et al., 1987), which is inconsistent with thestability of talc.

Talc formation by meteoric mixing faces a similar constraint, asfreshwater influx would tend to drive pH below marine values.Additionally, salinity would further decrease in response to post-depositional meteoric influx. Higher water activity (or lower salinity)favors the precipitation of Mg-silicates of higher hydration, namelysepiolite. Lower water activity (or near normal marine salinity), onthe other hand, favors kerolite and/or other 2:1 silicates, and synthesiswork reinforces these constraints (Jones and Galan, 1988; Siffert andWey, 1962; Stoessell and Hay, 1978). Mg-silicate occurrences in peri-marine and lacustrine environments also conform to this trend, sowell, in fact, that Mg-silicate assemblages have been used to tracksalinity fluctuations and the influence of meteoric water in a numberof modern and ancient settings (Calvo et al., 1999; Jones, 1986;

Stoessell and Hay, 1978; Weaver and Beck, 1977; Webster and Jones,1994).

Of all the requirements for early diagenetic talc formation,elevated pH (≥~8.7) is the most problematic, given the difficulty ofincreasing seawater pH, which is largely buffered by the carbon cycle,to the required level. The more realistic alternative is to increase pH inpore waters via microbial processes. Elevated pH can be achievedthrough certain microbial reactions that remineralize organic matter,a set of processes that need not have decreased salinity, Mg2+ and/orSiO2(aq) levels. Both the Akademikerbreen and Fifteenmile talcdeposits are intimately associated with microbial facies: bedded talcdrapes occur with microbialaminites and stromatolites in Fifteenmilesamples, and both localities host talc nodules within microbialites.

Anaerobic respiration is a particularly attractive mechanism forincreasing pH. In a Neoproterozoic world with low pO2 and nomacrometazoanbioturbation, anoxiawouldhavebeena common featureof early diagenetic carbonate sediments. Given the additional that nitratelevels were also low (Fennel et al., 2005), ferric iron and sulfate wouldhave been themost energetically favorable oxidants available for organiccarbon remineralization, with Fe3+ favored over SO4

2− (Canfield et al.,2005). Under conditions where the reactive Fe flux was high, (perhapsultimately sourced fromhydrothermal Fefluxes rather thandetrital inputin these sediment-starved environments) dissimilatory Fe reductionwould have been particularly effective at increasing pH through thefollowing reaction:

2CH2O þ 8FeðOHÞ3 ➔ 2CO2 þ 8Fe2þ þ 6H2O þ 16OH

−

When the reactive Fe pool was kept low, either by exhaustion ofreactive Fe3+ or by pyrite precipitation (Johnston et al., 2010; Lyonsand Severmann, 2006), sulfate reduction would have contributed to apH increase in solution:

2CH2O þ SO2−4 ➔ 2CO2 þ H2Sþ 2OH

−

Fe-speciation and isotopic analyses of other early Neoproterozoicsuccessions indicate that these two metabolic pathways stronglyinfluenced marine chemistry in both basinal and platform settings,with organic carbon fluxes and variations in basinal Fe and S cyclescontrolling the balance between them (2010; Canfield et al., 2008,Johnston et al., 2009). But regardless of the interplay betweendifferent types of anaerobic respiration, both pathways would haveresulted in a pH increase without changes in other conditions neededfor talc precipitation. Pore waters in microbially-dominated sedi-ments would have been especially susceptible to pH increase drivenby anaerobic respiration, where higher sediment to pore-water ratiosfavored more effective changes in chemistry.

In addition to the Svalbard and Canadian occurrences describedhere, sedimentary (oolitic) talc has been reported in association withstromatolitic dolomites in an Infracambrian (Neoproterozoic) succes-sion of the Volta Basin of West Africa (Millot and Palausi, 1959). Theauthors argue from petrographic analyses that talc formationpreceded dolomitization and silicification. Noack et al. (1989) alsoreported oolitic talc in the Schisto-Calcaire Group of the West CongoBasin. Overlying the Upper Diamictite Formation, Schisto-Calcairecarbonates have strontium isotope values of 0.7073 (Frimmel et al.,2006; Poidevin, 2007), consistent with a Cryogenian age (716–635 Ma). The sedimentology of the Schisto-Calcaire Group has beeninterpreted to reflect alternately shallow marine and lagoonalconditions, separated from outer subtidal regions of the platform bystromatolitic reefs (Trompette and Boudzoumou, 1988). Noack et al.(1989) interpret the talc to have formed by transformation fromstevensite or sepiolite on the basis of thermodynamic considerations.This interpretation is not incompatible with ours; stevensite precip-itation is generally only observed from synthesis studies at high pHand near marine salinity (La Iglesia, 1978; Siffert and Wey, 1962). Wehave not conclusively identified stevensite in our experiments, but

21N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

have identified a corrensite-like phase forming with talc at high pH.Corrensite formation at low temperature requires the initial forma-tion of a stevensite-like smectite into which brucite-like layers areprecipitated (Reynolds, 1988).

The apparent concentration of low temperature talc deposits inmid-Neoproterozoic carbonate platforms may find explanation in thesuperposition of Neoproterozoic-specific environmental conditionson the broader set of requirements (low detrital influx, high silica, lowoxygen, sufficient Mg, and efficient sulfide scavenging (Ben-Yaakov,1973)) likely to have characterized Precambrian carbonate platformsin general. Fe speciation chemistry indicates that iron played a moreimportant role in organic remineralization after 800 million years agothan it did earlier in the Proterozoic Eon (Canfield et al., 2008;Johnston et al., 2010). Also, the hypothesized shift from a Mesopro-terozoic biosphere with warm, high pCO2 atmosphere with highseawater DIC to a more glacial-prone later Neoproterozoic world withlower pCO2 and DIC would further have eased chemical resistance toincreases in pore-water pH in Neoproterozoic carbonate platforms(Kah and Bartley, 2004). During deposition of the Akademikerbreenand Fifteenmile Groups, anaerobic respiration in microbial sedimentreceiving little terrigenous input provided a fortuitous combination ofchemistry and depositional conditions to increase pH and precipitateauthigenic Mg-silicate at near marine salinity. The subsequent burialof this material would have driven a thermal transformation tokerolite and eventually to turbostratically-disordered talc.

Talc is not the only product to result from unusual pore waterchemistry on Neoproterozoic platforms. As discussed above, theidentification of saponite, an Al-bearing trioctahedral smectite,suggests additional diagenetic reactions. In the absence of kineticconstraints, but instead using thermodynamic considerations basedon field data, Weaver and Beck (1977) suggest modern seawater needonly increased SiO2(aq) and/or pH to drive a diagenetic conversionfrom detrital montmorillonite to saponite. Thus, the same chemicalconditions that led to talc precipitation were also capable of drivingsmectite recrystallization. Bristow et al. (2009) identified saponite(and corrensite) in Member 2 of the Ediacaran Doushantuo Formationand suggested a lacustrine depositional environment. Our data,however, show that early diagenetic modification of Neoproterozoicseawater is also capable of producing saponite. Thus, in the absence ofindependent sedimentological evidence of non-marine deposition,saponite cannot be taken as evidence for lacustrine deposition inNeoproterozoic successions. Indeed, the common occurrence ofsaponite in modern lacustrine settings rather than marine maysimply reflect decreased seawater silica levels after the radiation ofsiliceous plankton in addition to lower modern marine pH. Thisevolutionary innovation alone would have been sufficient to leavemodern saline lakes among the few environments able to concentrateand drive saponite formation from a dioctahedral precursor.

7. Conclusions

Sedimentology, petrography and mineralogy all indicate thatsedimentary talc reflects distinct aqueous chemistry recorded ontwo separate Neoproterozoic carbonate platform margins. Precipita-tion experiments withmodified “Neoproterozoic-like” seawater showthat elevated SiO2(aq), normal marine salinity and Mg2+ levels, andlow Al3+ all favor talc formation. However, the switch that enablestalc nucleation is elevated pH. Above pH 8.6–8.7, the deprotonation ofH4SiO4

0 initiates Mg-silica complexation and nucleation, which leadsto the subsequent formation of a kerolite precursor composed ofhydrated 2:1 layers with little to no stacking order. The progressivedehydration of this material, driven by low water activity in solutionor by burial diagenesis, leads to the formation of kerolite andeventually turbostratic talc.

Our analyses, together with other reports of sedimentary talcthrough this interval, suggest that the later Neoproterozoic Era

seawater is unusually favorable for early diagenetic talc precipitation.Although we argue that talc precipitation was occurring within thesediment, contemporaneous seawater must have been speciallypoised, only requiring reasonable changes to chemistry within thepore-waters. Following from this, we suggest that anaerobic respira-tion may have provided the alkalinity pump that tipped the scalestoward talc formation in soft sediments. The degree to whichsedimentary processes altered overlying seawater, neither of whichare known uniquely, hinders direct interpretation of paleo-seawatercompositions and pH. Further, whether the Neoproterozoic was thefirst and last time talc formed with marine carbonates at lowtemperature is unclear, but its identification of multiple continentsspeaks to the state of the Cryogenian oceans. This study does elucidatethat insight into seawater and pore-water chemistry is clearlyrecorded by early diagenetic Mg-silicates. An improved understand-ing of the kinetic and thermodynamic controls behind their formationwill continue to exploit these minerals as recorders of early diageneticchemistry in modern and ancient marine and lacustrine settings.

Acknowledgments

NJT acknowledges support from Royal Society Research Grant2009/R2 and from Churchill College, Cambridge. NJT also thanks T.Abraham, M. Zhang, R. Harrison, R. Parsons and M. Walker foranalytical assistance, and C. Jeans for fruitful discussions. Research byAHK and DTJ was supported, in part, by NASA Exobiology grantNNX07AV51G. FAM and JVS thank Charlie Roots and the YukonGeological Survey for logistical support. The authors thank David Fikefor constructive reviews.

Appendix A. Supplementary data

Supplementary data to this article can be found online atdoi:10.1016/j.epsl.2011.03.041.

References

Abbott, G., 1997. Geology of the Upper Hart River Area, Eastern Ogilvie Mountains,Yukon Territory (116A/10, 116A/11). Exploration and Geological Services Division,Yukon Region, Bulletin, vol. 9, pp. 1–76.

Abtahi, A., 1985. Synthesis of sepiolite at room-temperature from SiO2 and MgCl2solution. Clay Miner. 20, 521–523.

Baronnet, A., 1992. Polytypism and stacking disorder. In: Buseck, P.R. (Ed.), Mineralsand Reactions at the Atomic Scale: Transmission Electron Microscopy, Vol. 27.Mineralogical Society of America, Washington, DC, pp. 231–288.

Ben-Yaakov, S., 1973. pH buffering of pore water of recent anoxic marine sediments.Limnol. Oceanogr. 18, 86–94.

Bodine,M.W., 1983. Trioctahedral claymineral assemblages in Paleozoicmarine evaporiterocks. Sixth International Symposium on Salt, vol. 1. Salt Institute, pp. 267–284.

Braitsch, O., 1971. Salt Deposits, Their Origin and Composition. Springer-Verlag, Berlin.Brew, D.M.R., Glasser, F.P., 2005a. The magnesia-silica gel phase in slag cements: alkali

(K, Cs) sorption potential of synthetic gels. Cement Concrete Res. 35, 77–83.Brew, D.R.M., Glasser, F.P., 2005b. Synthesis and characterisation of magnesium silicate

hydrate gels. Cement Concrete Res. 35, 85–98.Brindley, G.W., Bish, D.L., Wan, H.M., 1977. Nature of kerolite and its relation to talc and

stevensite. Mineralogical Mag 41, 443–452.Bristow, T.F., Kennedy, M.J., Derkowski, A., Droser, M.L., Jiang, G.Q., Creaser, R.A., 2009.

Mineralogical constraints on the paleoenvironments of the Ediacaran DoushantuoFormation. Proc. Natl. Acad. Sci. U. S. A. 106, 13190–13195.

Butterfield, N.J., Knoll, A.H., Swett, K., 1994. Paleobiology of the NeoproterozoicSvanbergfjellet Formation, Spitsbergen. Fossils Strata 34, 1–84.

Calvo, J.P., Blanc-Valleron, M.M., Rodriguez-Arandia, J.P., Rouchy, J.M., Sanz, M.E., 1999.Authigenic clay minerals in continental evaporitic environments. Special Publica-tion IAS, vol. 27. IAS, pp. 129–151.

Canfield, D.E., Kristensen, E., Thamdrup, B., 2005. Aquatic Geomicrobiology. Elsevier,Amsterdam. 636 pp.

Canfield, D.E., Poulton, S.W., Knoll, A.H., Narbonne, G.M., Ross, G., Goldberg, T., Strauss,H., 2008. Ferruginous conditions dominated later neoproterozoic deep-waterchemistry. Science 321, 949–952.

Carrado, K.A., Decarreau, A., Petit, S., Bergaya, F., Lagaly, G., 2006. Synthetic clayminerals and purification of natural clays. In: Bergaya, F., Theng, B.K.G., Lagaly, G.(Eds.), Handbook of Clay Science, vol. 1. Elsevier, pp. 115–139.

Delacaillerie, J.B.D., Kermarec, M., Clause, O., 1995. Si-29 NMR observation of anamorphous magnesium-silicate formed during impregnation of silica with Mg(II)in aqueous-solution. J. Phys. Chem. 99, 17273–17281.

22 N.J. Tosca et al. / Earth and Planetary Science Letters 306 (2011) 11–22

Derry, L.A., Keto, L.S., Jacobsen, S.B., Knoll, A.H., Swett, K., 1989. Sr isotopic variations inUpper Proterozoic carbonates from Svalbard and East Greenland. GeochimicaCosmochimica Acta 53, 2231–2239.

Dietzel, M., 2000. Dissolution of silicates and the stability of polysilicic acid. GeochimicaCosmochimica Acta 64, 3275–3281.

Evans, B.W., Guggenheim, S., 1988. Talc, pyrophyllite, and related minerals. Rev.Mineralogy 19, 225–294.

Fairchild, I.J., Knoll, A.H., Swett, K., 1991. Coastal lithofacies and biofacies associatedwith syndepositional dolomitization and silicification (Draken Formation, UpperRiphean, Svalbard). Precambrian Res. 53, 165–197.

Farmer, V.C., 1974. The layer silicates. In: Farmer, V.C. (Ed.), The Infrared Spectra ofMinerals:Mineralogical Society Monograph. Mineralogical Society, London, pp. 331–364.

Felmy, A.R., Cho, H., Rustad, J.R., Mason, M.J., 2001. An aqueous thermodynamic modelfor polymerized silica species to high ionic strength. J. Solut. Chem. 30, 509–525.

Fennel, K., Follows, M., Falkowski, P.G., 2005. The co-evolution of the nitrogen, carbonand oxygen cycles in the Proterozoic ocean. Am. J. Sci. 305, 526–545.

Friedman, G.M., 1965. Occurrence of talc as a clay mineral in sedimentary rocks. Nature207, 283–284.

Frimmel, H.E., Tack, L., Basel, M.S., Nutman, A.P., Boven, A., 2006. Provenance andchemostratigraphy of the Neoproterozoic West Congolian Group in the DemocraticRepublic of Congo. J. Afr. Earth Sci. 46, 221–239.

Giese, R.F., 1975. Interlayer bonding in talc and pyrophyllite. Clays Clay Miner. 23,165–166.

Gualtieri, A.F., 1999. Modelling the nature of disorder in talc by simulation of X-raypowder patterns. Eur. J. Mineralogy 11, 521–532.

Halverson, G.P., Dudas, F.O., Maloof, A.C., Bowring, S.A., 2007. Evolution of the Sr-87/Sr-86 composition of Neoproterozoic seawater. Palaeogeography PalaeoclimatologyPalaeoecology 256, 103–129.

Iler, R.K., 1979. The Chemistry of Silica: Solubility, Polymerization, Colloid and SurfaceProperties, and Biochemistry. Wiley, New York. 866 pp.

Jefferson, C.W., Parrish, R., 1989. Late Proterozoic stratigraphy, U/Pb zircon ages and rifttectonics, Mackenzie Mountains, northwestern Canada. Can. J. Earth Sci. 26,1784–1801.

Johnston, D.T., Wolfe-Simon, F., Pearson, A., Knoll, A.H., 2009. Anoxygenic photosyn-thesis modulated Proterozoic oxygen and sustained Earth's middle age. Proc. Natl.Acad. Sci. U. S. A. 106, 16925–16929.

Johnston, D.T., Poulton, S.W., Dehler, C., Porter, S., Husson, J., Canfield, D.E., Knoll, A.H.,2010. An emerging picture of Neoproterozoic ocean chemistry: insights from theChuar Group, Grand Canyon, USA. Earth Planetary Sci. Lett. 290, 64–73.

Jones, B.F., 1986. Claymineral diagenesis in lacustrine sediments. In: Mumpton, F.A. (Ed.),Studies in Diagenesis, vol. 1578: USGS Bulletin. U.S. Geological Survey, pp. 291–300.

Jones, B.F., Galan, E., 1988. Sepiolite and palygorskite. Rev. Mineralogy 19, 631–674.Kah, L.C., Bartley, J.K., 2004. Marine carbon reservoir, C-org-C-carb coupling, and the

evolution of the Proterozoic carbon cycle. Geology 32, 129–132.Knoll, A.H., 1984. Microbiotas of the Late Precambrian Hunnberg Formation,

Nordaustlandet, Svalbard. J. Paleontology 58, 131–162.Knoll, A.H., Swett, K., 1990. Carbonate deposition during the Late Proterozoic Era — an

example from Spitsbergen. Am. J. Sci. 290A, 104–132.Knoll, A.H., Hayes, J.M., Kaufman, A.J., Swett, K., Lambert, I.B., 1986. Secular variation in

carbon isotope ratios from Upper Proterozoic successions of Svalbard and EastGreenland. Nature 321, 832–838.

Kogure, T., Kameda, J., 2008. High-resolution TEM and XRD simulation of stackingdisorder in 2 : 1 phyllosilicates. Z. Kristallographie 223, 69–75.

La Iglesia, A., 1978. Sintesis de la sepiolita a temperatura ambiente por precipitacionhomogenea. Bol. Geologico Minero 89, 258–265.

Lyons, T.W., Severmann, S., 2006. A critical look at iron paleoredox proxies: new insightsfrommodern euxinic marine basins. Geochimica Cosmochimica Acta 70, 5698–5722.

Macdonald, F.A., Roots, C.F., 2009. Upper Fifteenmile Group in the Ogilvie Mountainsand correlations of early Neoproterozoic strata in the northern Cordillera. In:McFarlane, K.E., Weston, L.H., Blackburn, L.R. (Eds.), Yukon Exploration andGeology 2009. Yukon Geological Survey, Whitehorse, YT, pp. 237–252.

Macdonald, F.A., Schmitz, M.D., Crowley, J.L., Roots, C.F., Jones, D.S., Maloof, A.C., Strauss,J.V., Cohen, P.A., Johnston, D.T., Schrag, D.P., 2010. Calibrating the Cryogenian.Science 327, 1241–1243.

Maliva, R.G., Knoll, A.H., Simonson, B.M., 2005. Secular change in the Precambrian silicacycle: insights from chert petrology. Geological Soc. Am. Bull. 117, 835–845.

Maloof, A.C., Halverson, G.P., Kirschvink, J.L., Hoffman, P.F., Schrag, D.P., 2004. Inertialinterchange true polar wander: combined paleomagnetic, isotopic and strati-graphic evidence from the Neoproterozoic Akademikerbreen Group, Svalbard.Geological Soc. Am. Bull. 118, 1099–1124.

Marumo, K., Hattori, K.H., 1999. Seafloor hydrothermal clay alteration at Jade in theback-arc Okinawa Trough: mineralogy, geochemistry and isotope characteristics.Geochimica Cosmochimica Acta 63, 2785–2804.

McCaffrey, M.A., Lazar, B., Holland, H.D., 1987. The evaporation path of seawater and thecoprecipitation of Br− and K+ with halite. J. Sedimentary Petrology 57, 928–937.

Millot, G., Palausi, G., 1959. Sur un talc d'origine sedimentaire. C. R. Soc. GeologiqueFrancais 45–47.

Mitsuda, T., Taguchi, H., 1977. Formation of magnesium-silicate hydrate and itscrystallization to talc. Cement Concrete Res. 7, 223–230.

Mizutani, T., Fukushima, Y., Kamigaito, O., 1990. Mechanism of the copolymerization ofsilicic-acid and metal-ions in aqueous-media. Bull. Chem. Soc. Jpn. 63, 618–619.

Mustard, P.S., Donaldson, J.A., 1990. Paleokarst breccias, calcretes, silcretes and faulttalus breccias at the base of upper Proterozoic "Windermere" strata, northernCanadian Cordillera. J. Sedimentary Petrology 60, 525–539.

Noack, Y., Decarreau, A., Boudzoumou, F., Trompette, R., 1989. Low-temperature oolitictalc in Upper Proterozoic rocks, Congo. J. Sedimentary Petrology 59, 717–723.

Packter, A., 1986. Precipitation of alkaline-earth metal silicate hydrates from aqueoussolution — ionic equilibria, crystalline phases and precipitation mechanisms.Crystal Res. Technol. 21, 575–585.

Poidevin, J.L., 2007. Sr-isotope stratigraphy and dating of Neoproterozoic carbonatesand glacials from the northern and western parts of the Congo Craton. C. R.Geoscience 339, 259–273.

Reynolds Jr., R.C., 1988. Mixed layer chlorite minerals. In: Bailey, S.W. (Ed.), HydrousPhyllosilicates, vol. 19. Mineralogical Society of America, Washington, DC, p. 725.

Russell, J.D., Fraser, A.R., 1994. Infraredmethods. In:Wilson, M.J. (Ed.), ClayMineralogy:Spectroscopic and Chemical Determinative Methods. Chapman and Hall, London,pp. 11–67.

Santschi, P.H., Schindler, P., 1974. Complex-formation in ternary-systems Ca(II)–H4SiO4–H2O and Mg(II)–H4SiO4–H2O. J. Chem. Soc. Dalton Trans. 181–184.

Siffert, B., Wey, R., 1962. Synthese d'une sepiolite a temperature ordinaire. C. R. Acad.Sci. 253, 142–145.

Stoessell, R.K., Hay, R.L., 1978. Geochemical origin of sepiolite and kerolite at Amboseli,Kenya. Contrib. Mineralogy Petrology 65, 255–267.

Strese, H., Hofmann, U., 1941. Synthese von Magnesiumsilikat-Gelen mit zweidimen-sional regelmaessiger struktur. Z. Anorganische Allg. Chem. 247, 65–95.

Swett, K., Knoll, A.H., 1989. Marine pisolites from Upper Proterozoic Carbonates of EastGreenland and Spitsbergen. Sedimentology 36, 75–93.

Takahashi, N., Tanaka, M., Satoh, T., Endo, T., 1994. Study of synthetic clay-minerals.3.Synthesis and characterization of 2-dimensional talc. Bull. Chem. Soc. Jpn. 67,2463–2467.

Trompette, R., Boudzoumou, F., 1988. Paleogeographic significance of stromatoliticbuildings on late Proterozoic platforms: the example of the West Congo Basin.Palaeogeography Palaeoclimatology Palaeoecology 48, 101–112.

Turner, E.C., Long, D.G.F., 2008. Basin architecture and syndepositional fault activityduring deposition of the Neoproterozoic Mackenzie Mountains Supergroup,Northwest Territories, Canada. Can. J. Earth Sci. 45, 1159–1184.

Van Kooten, G.K., Watts, A.B., Coogan, J., Mount, V.S., Swenson, R.F., Daggett, P.H.,Clough, J.G., Roberts, C.T., Bergman, S.C., 1997. Alaska Division of Geological andGeophysical Surveys, Report of Investigations 96-6A. Alaska Division of Geologicaland Geophysical Surveys. Fairbanks, AK, 3 sheets, scale 1:125,000.

Velde, B., 1985. Clay Minerals: A Physico-Chemical Explanation of their Occurrence.Elsevier, Amsterdam. 427 pp.

Velde, B., Meunier, A., 2008. The Origin of Clay Minerals in Soils and Weathered Rocks.Springer-Verlag, Berlin. 406 pp.

Weaver, C.E., Beck, K.C., 1977. Miocene of SE United-States — model for chemicalsedimentation in a peri-marine environment. Sedimentary Geology 17, 1–234.

Webster, D.M., Jones, B.F., 1994. Paleoenvironmental implications of lacustrine clayminerals from the Double Lakes Formation, Southern High Plains, Texas.Sedimentology and Geochemistry of Modern and Ancient Saline Lakes, vol. 50:SEPM Special Publication. SEPM, pp. 159–172.

Wei, J.X., Chen, Y.M., 2006. Research on the cementious materials in MgO–SiO2–H2Osystem at room temperature. In: Tongbo, S., Rongxi, S., Wensheng, Z. (Eds.),Proceedings of the 6th International Symposium on Cement & Concrete andCANMET/ACI International Symposium on Concrete Technology for SustainableDevelopment, vols. 1 and 2, pp. 594–598.

Wollast, R., Mackenzie, F.T., Bricker, O.P., 1968. Experimental precipitation and genesisof sepiolite at Earth-surface conditions. Am. Mineralogist 53, 1645–1661.

Zhang, M., Hui, Q., Lou, X.-J., Redfern, S.A.T., Salje, E.K.H., Tarantino, S.C., 2006.Dehydroxylation, proton migration, and structural changes in heated talc: aninfrared spectroscopic study. Am. Mineralogist 91, 816–825.