Earth and Planetary Science Letters 389 (2014) 209–220

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Trace element content of sedimentary pyrite as a new proxy for

deep-time ocean–atmosphere evolution
Ross R. Large a,∗ , Jacqueline A. Halpin a , Leonid V. Danyushevsky a ,
Valeriy V. Maslennikov b , Stuart W. Bull a , John A. Long c,d , Daniel D. Gregory a ,
Elena Lounejeva a , Timothy W. Lyons e , Patrick J. Sack f , Peter J. McGoldrick a ,
Clive R. Calver g
a
Centre for Ore Deposit and Exploration Science (CODES), University of Tasmania, Private Bag 126, Hobart, Tasmania 7001, Australia
b
Institute of Mineralogy, Russian Academy of Science, Urals Branch, Miass, Russia
c
School of Biological Sciences, Flinders University, PO Box 2100, Adelaide, South Australia 5001, Australia
d
Museum Victoria, PO Box 666, Melbourne, Victoria 3001, Australia
e
Department of Earth Sciences, University of California, Riverside, CA 92521, USA
f
Yukon Geological Survey, PO Box 2703 (K-102), Whitehorse, Yukon Y1A 2C6, Canada
g
Mineral Resources Tasmania, PO Box 56, Rosny Park, Tasmania 7018, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Sedimentary pyrite formed in the water column, or during diagenesis in organic muds, provides an
Received 29 July 2013 accessible proxy for seawater chemistry in the marine rock record. Except for Mo, U, Ni and Cr,
Received in revised form 8 December 2013 surprisingly little is known about trace element trends in the deep time oceans, even though they
Accepted 15 December 2013
are critical to developing better models for the evolution of the Earth’s atmosphere and evolutionary
Available online xxxx
pathways of life. Here we introduce a novel approach to simultaneously quantify a suite of trace elements
Editor: G. Henderson
in sedimentary pyrite from marine black shales. These trace element concentrations, at least in a first-
Keywords: order sense, track the primary elemental abundances in coeval seawater. In general, the trace element
palaeo-oceanography patterns show significant variation of several orders of magnitude in the Archaean and Phanerozoic,
sedimentary pyrite but less variation on longer wavelengths in the Proterozoic. Certain trace elements (e.g., Ni, Co, As,
trace elements Cr) have generally decreased in the oceans through the Precambrian, other elements (e.g., Mo, Zn, Mn)
ocean chemistry have generally increased, and a further group initially increased and then decreased (e.g., Se and U).
oxygenation proxy
These changes appear to be controlled by many factors, in particular: 1) oxygenation cycles of the Earth’s
selenium
ocean–atmosphere system, 2) the composition of exposed crustal rocks, 3) long term rates of continental
erosion, and 4) cycles of ocean anoxia. We show that Ni and Co content of seawater is affected by global
Large Igneous Province events, whereas redox sensitive trace elements such as Se and Mo are affected
by atmosphere oxygenation. Positive jumps in Mo and Se concentrations prior to the Great Oxidation
Event (GOE1, c. 2500 Ma) suggest pulses of oxygenation may have occurred as early as 2950 Ma. A flat to
declining pattern of many biologically important nutrient elements through the mid to late Proterozoic
may relate to declining atmosphere O2 , and supports previous models of nutrient deficiency inhibiting
marine evolution during this period. These trace elements (Mo, Se, U, Cu and Ni) reach a minimum in
the mid Cryogenian and rise abruptly toward the end of the Cryogenian marking the position of a second
Great Oxidation Event (GOE2).
© 2013 Elsevier B.V. All rights reserved.

1. Introduction of the global ocean (Holland, 1984; Anbar et al., 2007; Scott et
al., 2008; Partin et al., 2013). Most of the redox sensitive TE, as
Whole rock variation of trace elements (TE) — in particular Mo, well as multitude of other TE in black shales, are concentrated in
U, V, Cr and Ni — in black shales deposited in marine environments early-formed pyrite (Huerta-Diaz and Morse, 1992; Large et al.,
have been successfully used to interpret palaeo-redox conditions 2007, 2009; Gregory et al., in press). Laser Ablation-Inductively
on the ocean floor (Lyons et al., 2003; Tribovillard et al., 2006; Coupled Plasma Mass Spectrometer (LA-ICPMS) techniques devel-
Reinhard et al., 2013), including temporal changes in oxygenation oped to study zonation in the composition of pyrite (FeS2 ) in var-
ious ore systems (Danyushevsky et al., 2011) have demonstrated
how this information can be used to study potential sources of
* Corresponding author. metals (Meffre et al., 2008), track changes in the chemistry of hy-
E-mail address: ross.large@utas.edu.au (R.R. Large). drothermal fluids (Large et al., 2009), determine the evolution of

0012-821X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.epsl.2013.12.020
210 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

ore deposits (Thomas et al., 2011) and establish the effectiveness

of pyrite as a trap for TE in contaminated sediments (Gregory
et al., 2013, in press). We suggest here, that in the same way that
hydrothermal pyrite tracks changes in chemistry of ore fluids, sedi-
mentary pyrite, which is a common trace mineral in carbonaceous
marine black shales (Fig. 1a–d), can be used to track first-order
changes in the chemistry of seawater and ultimately oxygenation
of the ocean–atmosphere system. Whereas bulk shale studies re-
quire samples of particular palaeo-redox facies (Lyons et al., 2003;
Scott et al., 2008; Algeo and Rowe, 2012), the focus on early-
formed pyrites developed here allows sampling of a wider range
of redox conditions, enabling a more complete record of seawater
variation. For this study we have analysed over 1880 sedimentary
pyrites from 131 black shale samples covering 77 time intervals
from 3515 Ma to present day (Supplementary Tables 1–4).

2. TE analysis of sedimentary pyrite as a proxy for seawater

chemistry

2.1. Incorporation of TE in sedimentary pyrite

Sedimentary pyrite forms in seafloor muds when H2 S, pro-

duced by the microbial reduction of marine sulfate, reacts with
iron to form pyrite (FeS2 ), arguably by way of an iron mono-
sulfide precursor, such as mackinawite or greigite (Rickard and
Luther, 1997; Rickard, 2012). Experimental studies (Huerta-Diaz
and Morse, 1992; Morse and Arakaki, 1993) have shown that
many TE are incorporated into the precursor iron monosulfide
at an early stage, including (in approximate order of incorpora-
tion) As, Hg, Mo, Co, Cu, Mn, Ni, Cr, Pb, Zn and Cd, which are
absorbed from seawater and local pore waters in seafloor muds
(Larget et al., 2007, 2009; Gregory et al., in press). This TE en-
richment process is controlled by the amount of pyrite produced
and the amount of trace metals available from seawater and pore
waters (Huerta-Diaz and Morse, 1992). However it is recognised
that microbially controlled reduction and sequestration of TE and
sulfate on the seafloor are complex biochemical processes. Vari-
ables like Eh/pH conditions in the sediment/water column, de-
gree of bio-productivity and availability of preferred electron ac-
ceptors will influence which TE the microbes reduce and co-
precipitate with pyrite at any given time (Kulp and Pratt, 2004;
Mitchell et al., 2012).
Various textural forms of pyrite are typical of syngenetic/dia-
genetic sedimentary origins, including framboids, microcrystalline
clusters, small nodules and layers of fine crystals (Fig. 1a–d).
Late diagenetic pyrites typically overgrow earlier pyrite genera-
tions (Thomas et al., 2011; Rickard, 2012) and commonly have
a rounded anhedral surface which may be expressed as elon-
gate nodules or various spheroidal and nodular forms (Large et
al. 2007, 2009). Recent studies have shown that many of the
TE originally thought to be present in the structure of pyrite,
are actually present as micro-inclusions (Ciobanu et al., 2009;
Large et al., 2009; Deditius et al., 2011) abundant in early forms
of pyrite (Fig. 1a–c). Whereas Ni, Co, and Se are commonly present
in the pyrite structure, As, Cu, Zn, Pb, Bi, Sb, Tl, Mo, Ag, Cd, Mn, Fig. 1. a–d, Various textures of sedimentary pyrite; a, cluster of framboids, b, cloud
Hg, and Te may be either in the structure or as sulfide micro- of fine framboids and microcrystals, c, nodule, d, fine microcrystals concentrated
inclusions, and Ti, V, U, Ba, Sn, W, and Cr, are invariably present in a sedimentary bed. e–j, LA-ICPMS images of TE distribution in pyrite nodule
within micro-inclusions of matrix material. This earliest formed (c above) and surrounding matrix. Note the high concentration of Se in pyrite rela-
tive to matrix in (e), compared with low concentration of V in the pyrite relative to
inclusion-bearing sedimentary pyrite is the most TE rich, and sub-
matrix in j. k, time resolved LA-ICPMS spectra of selected TE in a sedimentary pyrite
sequent, more crystalline generations with fewer inclusions (late analysis. The laser was switched on at 31 s after the start of the analysis. Ablation
diagenetic to early metamorphic) may contain lower TE concentra- continued for 60 s. The intensity of the signal for each isotope is taken as the dif-
tions and trend toward a pure FeS2 composition (Huerta-Diaz and ference between the mean signal with laser off (background, the first 30 s of each
acquisition) and the mean signal with the laser on. As the vertical axis is counts
Morse, 1992; Large et al., 2007; Large et al., 2009; Rickard, 2012).
recorded by the detector on ICPMS, differences in intensity of the signals from dif-
Trace elements that are loosely held in the pyrite structure or oc- ferent isotopes do not directly reflect differences in concentrations. The plot shows
cur as nano-inclusions (e.g., Zn, Cu, Pb, Ag, Mo) are readily expelled the differing response from TE in the pyrite structure (Pb, Co, Sb and Se) compared
from the pyrite during metamorphic recrystallisation, whereas TE to TE in micro-inclusions (Ti, Zn and Cr). Copper shows a dual response.
 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220 211

Table 1
Summary of TE variations in the Archean and Proterozoic oceans based on sedimentary pyrite composition.

Eon/Era Trace elements enriched* in ocean Trace elements depleted* in ocean Little change in mean trace element
concentration*
Across GOE2 (c. 660 Ma) Mo, Se, Zn, Cd, Ni, Co, As, Cu, Mn, Bi, Ba, W, Sn Ag, Cr, V, Tl
Hg, Pb, Sb, Te, U
Neoproterozoic to 660 Ma# Cu, Bi, Ba Mo, Zn, Ni, As, Se, Cd, Tl, Hg, Cr, Ag, Co, Mn, Pb, U, V, W, Sn
Sb, Te
Mesoproterozoic As, Tl Mo, Co, Mn, Hg, U, Cr, Ba, V, W, Te Zn, Ni, Se, Cd, Cu, Pb, Bi, Sn, Ag, Sb
Palaeoproterozoic Cu, Bi, U, Cr, V, W Mo, Zn, As, Cd, Co, Tl, Hg, Ba, Ag, Sb, Ni, Se, Mn, Pb, Sn
Te
Across GOE1 (c. 2500 Ma) Mo, Se, Tl, Mn, U Zn, Cd, Ni, Co, As, Cu, Ag, Sb, Hg, Te, V, W
Pb, Bi, Cr, Ba, Sn
Archean Ni, As, Co, Cu, Hg, Bi, Cr, W, Sn, Sb, Te Mo, Zn, Se, Cd, Mn, U, Ba Tl, Pb, V, Ag
*
Compared with the range of average values through time (0 to 3.5 Ga).
#
Neoproterozoic prior to GOE2 (Tonian-Cryogenian).

that are stoichiometric in the pyrite structure (e.g., Co, Ni, As, Se) All LA-ICPMS analyses of pyrite were performed at the Cen-
tend to remain and even become enriched during metamorphism tre for Ore Deposit and Exploration Science (CODES), University of
(Large et al., 2007, 2011b). Tasmania (see Supplementary Information for LA-ICPMS method-
ology). LA-ICPMS element maps clearly show those TE that are
2.2. Sample selection, rationale and analysis evenly distributed through the pyrite (e.g., Se in Fig. 1e), com-
pared with those that are zoned (Cu, Fig. 1h), those with a
patchy distribution (As, Fig. 1i), and those dominated by inclusions
In this study, samples of sedimentary pyrite were selected on
(Mo and Zn, Fig. 1f–g). The time-resolved signal of an individ-
the basis of: (1) metamorphic grade of the host shale, (2) texture
ual LA-ICPMS analysis (Fig. 1k) provides information on the likely
of pyrite and (3) Co/Ni ratio of pyrite, as outlined below.
form of elements incorporation in the pyrite (Large et al., 2009;
Thomas et al., 2011). Removal of the effects of matrix micro-
1. Black shales of mid greenschist facies or below were chosen.
inclusions leads to the pyrite compositions measured on several
Post lithification metamorphism and/or hydrothermal activity
pyrite grains from the same black shale sample commonly exhibit-
leads to recrystallisation of earlier pyrite forms and commonly
ing a variation of less than one order of magnitude (ppm scale),
results in pyrite overgrowths with euhedral outer surfaces and
compared with the overall dataset variation of up to four orders of
well developed crystal growth zones with alternation of TE
magnitude for any given TE (see example in Supplementary Fig. 1).
chemistry (Large et al., 2009, 2013; Thomas et al., 2011). These
latter euhedral and zoned pyrite types have been studiously
avoided in this current study, as their chemistry reflects the 2.3. Proof of concept
composition and conditions of the metamorphic or hydrother-
mal fluids rather than the precursor seawater chemistry during Five separate lines of evidence support our contention that vari-
sedimentation. Samples with pyrrhotite were also rejected, as ations in the TE composition of early-formed sedimentary pyrite in
previous studies have shown a marked loss of TE during con- black shales are a good first order proxy for TE variations in the
version of pyrite to pyrrhotite (Thomas et al., 2011). ocean through time (Table 1). The physical/chemical connection
2. Pyrite was selected that retained textures indicative of syn- between early-formed sedimentary pyrite and seawater underpins
genesis or formation during early diagenesis. Textures indica- our first line of evidence. Syngenetic pyrite which forms in eu-
tive of diagenetic pyrite include; framboids, microcrystalline xinic water columns adsorbs TE from the seawater (Huerta-Diaz
pyrite in clouds in the shale, microcrystalline pyrite aligned and Morse, 1992) is an obvious proxy for seawater TE chem-
parallel to bedding, small cubic euhedral and bladed crystals, istry (Berner et al., 2013). However, diagenetic pyrite in anoxic to
small porous patches and rounded or elongate zoned nod- dysoxic bottom water conditions forms by biogenic reduction of
ules (Fig. 1a–d) (Huerta-Diaz and Morse, 1992; Large et al., sulfate in the pore waters of seafloor muds within the first few
2007, 2009; Guy et al., 2010). Pyrite textures indicative of centimetres of the seafloor (Wilkin et al., 1996). In this environ-
metamorphic and/or hydrothermal growth, including medium ment the muds contain from 60 to 85% seawater (Baldwin and But-
and large sized euhedral crystals (>1 mm), clear crystals with- ler, 1985; Harrold et al., 1999) and thus ensure a strong connection
out inclusions, and pyrite which cuts across the bedding as between pyrite chemistry and seawater chemistry, even though
blebs or veins, were not selected for analysis. Textural dis- sedimentary clays and organic matter are likely to play a role in
crimination between diagenetic and metamorphic pyrite tex- partitioning TE into pyrite. The seawater content of seafloor muds
tures was assisted by etching pyrite with nitric acid (Large remains over 40% in the top 100 to 200 m (Baldwin and Butler,
et al., 2007), coupled with the LA-ICPMS mapping technique 1985) and thus diagenetic pyrites forming at considerable depth
when required (Fig. 1e–j, Supplementary Fig. 7) (Large et al., will maintain a seawater connection influenced by matrix interac-
2009, 2013; Thomas et al., 2011). tions. The significance of matrix/pore water interaction in effecting
3. Chemical screening of the pyrite was undertaken following LA- a change in sedimentary pyrite chemistry can be evaluated from
ICPMS analysis. On a Co versus Ni plot, samples which showed data on pyrite TE depth profiles in recent sedimentary environ-
a linear array parallel to Co/Ni = 1, and with Co/Ni < 2 were ments from Huerta-Diaz and Morse (1990, 1992). Their data, col-
selected as suitable. This is based on previous work (Bajwah lected from a range of anoxic to euxinic sedimentary environments
et al., 1987; Large et al., 2009) that have shown diagenetic in the Gulf of Mexico, shows that most TE in sedimentary pyrite
pyrite commonly has Co/Ni  2, whereas hydrothermal pyrite increase with depth from 0 up to 100 cm depth. However, the in-
has higher Co values that trend away from the Co/Ni = 1 line. crease is variable, depending on sedimentary environment and the
More than 80% of our data was deemed acceptable on this ba- TE measured, and is commonly in the range of 1.4 to 6 times (less
sis. than one order of magnitude), compared to the 2 to 4 orders of
212 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

Fig. 2. a, Changes in mean LA-ICPMS analyses of Mo, As, Ni, Co and Se in the paragenetic stages of sedimentary pyrite from the Neoproterozoic Khomolkho Formation, Siberia
from Large et al. (2007). Py1 = syngenetic to early diagenetic pyrite; py2 = early diagenetic pyrite; py3 = late diagenetic pyrite; py4 = metamorphic pyrite; py5 = late
metamorphic pyrite; pyrrhotite = late metamorphic pyrrhotite replacing pyrite. b, Range in TE content (Se, As, Ni and Mo) plotted against pyrite framboid size for a Late
Permian black shale from the Perth Basin, Western Australia. The range in individual TE for each framboid size is similar.

magnitude variation exhibited by sedimentary pyrite in our sam- The third line of evidence, based on our framboid data from
ples from different epochs and eras through time. the Perth Basin in Late Permian marine black shales (Thomas
The second line of evidence comes from a study of the LA- et al., 2004) shows there is no substantial difference in the TE
ICPMS composition of a range of different textural and paragenetic (range ppm) of small (5 μm) syngenetic framboids, compared
pyrite types, from syngenetic, through diagenetic to metamorphic, with larger (6 to 35 μm) pyrite framboids (Fig. 2b). All sizes of
in the Neoproterozoic Khomolkho Formation in the Lena Gold- framboids in this Late Permian black shale sample show similar TE
field, Siberia (Large et al., 2007). This study demonstrated a general contents (40 to 800 ppm Ni and As; 3 to 80 Mo and Se; Fig. 2b)
trend of steady to decreasing mean TE content in pyrite with time, suggesting that early-formed sedimentary pyrite is a first order
from syngenetic to diagenetic to metamorphic through the para- recorder of TE variations irrespective of its framboid size and mode
genesis (Fig. 2a). The changes during diagenesis for all five TE are of origin (syngenetic or diagenetic).
substantially less than one order of magnitude. This suggests that The fourth line of evidence is exhibited in Fig. 3, which demon-
the first order (2 to 4 orders of magnitude) temporal changes are strates a first order positive relationship between the LA-ICPMS TE
unlikely to relate to diagenesis in the seafloor muds, but more composition of sedimentary pyrite from an actively forming sed-
likely reflect the changes in TE content of seawater within and imentary basin the Cariaco Basin on the Venezuela shelf (Lyons
overlying the muds. et al., 2003; Piper and Perkins, 2004), and the TE composition of
 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220 213

The final proof of concept for the approach proposed here in-
volves comparing our data on the TE composition of pyrite with
the TE content of whole rock black shale samples, particularly
for Mo, which is commonly used to interpret palaeo-ocean con-
ditions and changes in oxygenation of the atmosphere/ocean sys-
tem through time (Lyons et al., 2003; Tribovillard et al., 2006;
Scott et al., 2008). A small group of black shale samples from the
Devonian Popovich Formation, Nevada, reveal similar down-hole
trends (Fig. 4); those TE concentrated in pyrite such as Mo and
Se are commonly between 10 to 100 times the whole rock value,
depending on the percentage of sedimentary pyrite and organic
matter in the sample.
In order to evaluate the deep time variability of TE in sedimen-
tary pyrite in black shales, we compare Mo data on whole rock
euxinic and/or organic-rich black shales from Scott et al. (2008)
with our data on the Mo content of sedimentary pyrite in black
shales over the same time interval (Fig. 5a and b). The two data
patterns are remarkably similar, with several times more Mo in
Fig. 3. Positive correlation (large dashed field) between mean TE concentrations in pyrite compared with the whole rock values, even though Mo is
current ocean and mean TE concentrations in modern sedimentary pyrite from Cari- present both in the sedimentary matrix and in the sedimentary
aco Basin. Vanadium, U and Ba do not fall in the same field as the other traces, as pyrite (Fig. 1) (Chappaz et al., 2014). A second comparison is made
they are concentrated in non-sulfide inclusions within the pyrite.
with the Ni/Fe ratio measured on whole rocks and minerals in ma-
rine Banded Iron Formations (BIFs) (Konhauser et al., 2009) (Fig. 5c
and d). Again the two datasets show similar patterns, although the
Konhauser et al. (2009) plot contains no Phanerozoic data due to
the lack of BIFs in this era. In summary, our pyrite TE plots have
reproduced the temporal trends in the previous whole rock data
and confirm the value of sedimentary pyrite as a proxy for deep
time variations of TE in the oceans.

3. Temporal trends in TE from sedimentary pyrite

Individual TE time curves derived from LA-ICPMS analysis of

1885 pyrite spots on 131 black shale samples are given in the
Supplementary Information (Supplementary Tables 2 and 3; Sup-
plementary Figs. 3–5). We have selected a subset of the TE to
focus our discussion; Mo, Ni, Co, As and Se measured in sedi-
mentary pyrite, and U and Cr measured in the sedimentary matrix
(Figs. 6–8). In the summary figures to follow, we use a rolling TE
mean to highlight the long-term TE trends (Fig. 6b–f, see also Sup-
plementary Fig. 2) at the expense of the variation in individual
sample sets (Fig. 6a). In general terms (see Mo, Ni, Co, As and Se
in Fig. 6b–f), most TE exhibit significant cyclic variation of sev-
eral orders of magnitude in the Archaean and Phanerozoic, but less
variation on longer wavelengths in the Proterozoic. In the Precam-
brian, there are four general TE trends:

Trend 1: a flat-to-broadly increasing pattern through the Archaean

Fig. 4. Down hole comparison between LA-ICPMS Mo and Se analyses from whole and Proterozoic: only Mo, Zn, Mn, and Cd show this pattern.
rock and sedimentary pyrite in black shales from drill hole CD 12c, Carlin District
Nevada (Large et al., 2011a).
Although this pattern is not unidirectional, it suggests that
supply of these TE to the ocean has exceeded drawdown into
seafloor muds through most of the Precambrian.
modern ocean water (http://www.mbari.org/chemsensor/summary.
Trend 2: A variable, but generally decreasing pattern through the
html). Relative to seawater, TE in Cariaco Basin pyrite are enriched
Archaean and Proterozoic: Ni, Co, As, Te, Tl, Hg, Sb, Pb, Bi, Cr,
between 5 and 8 orders of magnitude (Supplementary Table 1).
Sn, and W. This pattern suggests drawdown has exceeded sup-
Trace elements Mo, As, Ni, Zn, Cu and Se which are enriched in
ply for the Precambrian.
the current ocean (>100 ng/kg) are also enriched in the currently Trend 3: a general increase through the Mesoarchaean to early
forming sedimentary pyrite (>100 ppm), whereas Co, Pb, Sb, Tl, Palaeoproterozoic followed by a decrease through the Meso-
Cd and Ag which are in lower abundance in the ocean (>1 and and Neoproterozoic: Se, U, and V. In the first half of the pe-
<100 ng/kg) are also in lower abundance in sedimentary pyrite riod supply exceeded drawdown, with reversal in the second
(>10 and <100 ppm). The least abundant TE in the ocean (Te, half.
Bi, Au) are also the least abundant in Cariaco Basin sedimentary Trend 4: a relatively flat pattern with no obvious trend through
pyrite. Barium, V and U, which are present within silicate inclu- the Archaean and Proterozoic: Cu, Ag, and Ba. In this case TE
sions in the sedimentary pyrite, plot outside the pyrite correlation supply and drawdown were roughly matched through the pe-
field. riod but with perturbations.
214 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

Fig. 5. Comparison of published data on temporal variation in whole rock Mo and Ni analyses on black shales and BIFs with LA-ICPMS analyses of sedimentary pyrite in black
shales. a, Mo whole rock black shale data from Scott et al. (2008) and b, Mo in sedimentary pyrite derived from this study. c, Ni/Fe whole rock and mineral data from BIFs
(Konhauser et al., 2009) and d, Ni in sedimentary pyrite derived from this study.

All TE show high frequency and high amplitude variations through (Trend 2) shows a similar pattern to Ni, but with more varia-
the Phanerozoic suggesting geologically rapid changes in ocean tion (Fig. 6d). Most of the segments on the time curve when Ni
chemistry over this latter period of Earth history. It should be em- and Co in pyrite exceeds 1000 ppm and 500 ppm respectively,
phasised that the four trends discussed above are a first attempt to overlap with episodes of Large Igneous Province (LIP) eruptions of
group the data, but in fact each TE shows its own individual pat- mafic lavas and associated intrusives, both subareal and submarine
tern (Supplementary Figs. 3 and 5), that is probably controlled by (Prokoph et al., 2004) (Fig. 7). The komatiitic LIPs of the Archaean
a number of factors, which are beyond the scope of analysis here. are particularly enriched in Ni and Co, and thus contributed to the
A brief discussion of the temporal trends of the five TE shown in sustained high levels of Ni and Co to the global oceans and sed-
Fig. 6, and comparison between these pyrite TE trends and those imentary pyrite (Fig. 7). The highest Phanerozoic Ni and Co were
previously published for whole rock datasets follows below. measured in sedimentary pyrite of 250 Ma from the Perth Basin
Molybdenum (Trend 1) is a conservative TE in the current Western Australia; values up to 48,000 ppm Ni and 16,000 Co
ocean, present as the soluble MoO4 = species (Helz et al., 1996). were recorded, indicating the likelihood of widespread enrichment
The general rise in the Mo time curve (Fig. 6b) is attributed to of the global ocean in Ni and Co due to the Siberian Trap erup-
oxygenation of the atmosphere, particularly in the Neoarchean and tions. It is noteworthy that these Siberian eruptions are associated
Phanerozoic, leading to increased continental oxidative weathering with the largest nickel ore bodies on Earth, Norilsk and Talnack.
such that supply of Mo to the oceans exceeded the drawdown Arsenic (Trend 2) (Fig. 6e) also exhibits a declining trend
into seafloor sediments (Anbar et al., 2007; Scott et al., 2008; through the Precambrian. However, the As pattern has two seg-
Dahl et al., 2011). Our time curve shows that Mo in pyrite com- ments; 1) a continuous decrease from a mean of 5000 to 300 ppm
menced at a relatively low 0.5 to 1 ppm at 3500 Ma and increased in pyrite over the period 3000 to 1800 Ma; 2) followed by a rise to
through a series of steps to reach 10 to 70 ppm at 1000 Ma. The 10,000 ppm at 1400 Ma and fall to the Cryogenian TE trough, ris-
step up at around 2500 Ma likely corresponds to the Great Oxi- ing again at GOE2 around 660 Ma. The cause for the sharp decline
dation Event (GOE1) (Farquhar et al., 2001; Holland, 2006) when in As through Late Archean and Early Proterozoic is not source-
appreciable O2 first accumulated in the atmosphere. In the Neo- rock controlled, as unlike Ni and Co, As is not enriched in ultra-
proterozoic there is a trough in the Mo time curve (referred to as mafic or mafic rocks (Reimann and Caritat, 1998), and alternative
the Cryogenian TE trough hereafter) that concludes with the Stur- As-bearing Archean source rocks are not known. A possible expla-
tian glaciation and proposed Snowball Earth (Hoffman et al., 1998). nation for the As trend relates to the fact that As forms a soluble
A rapid rise corresponding to a second Great Oxidation Event (re- arsenate complex (HAsO24− ) in oxic to anoxic ocean conditions and
ferred to here as GOE2) occurred during the Neoproterozoic when soluble polysulfide complex (H2 As3 S− 6 ) in euxinic ocean conditions
oxygen rose to significant levels that ultimately supported the rise (see Supplementary Table 5). Thus higher levels of volatile ele-
of animal diversity. Our TE data suggests a threshold was reached ments, S and As, associated with a degassing Earth in the Archean,
by 660 Ma, earlier than recent estimates that place this event ei- may have contributed to stable arsenic polysulfide complexes in
ther shortly after the Marinoan Glaciation (630 Ma) (Sahoo et al., the mainly H2 S-bearing Archean ocean, which declined in concen-
2012) or prior to the late Ediacaran (580–550 Ma) (Fike et al., tration through the Early Proterozoic as H2 S levels in the ocean
2006; Canfield et al., 2007; Scott et al., 2008). dropped and As was drawn-down into seafloor pyrite. An increase
Nickel (Trend 2) is a nutrient TE present as Ni2+ in the mod- in As through the Mesoproterozoic (Fig. 6e) supports a switch to
ern ocean. The Ni time curve (Fig. 6c) shows an overall decrease anoxic conditions as the SO24− /H2 S ratio of the ocean continued to
through the Precambrian to 660 Ma, but most of the drop oc- increase, followed by an As decline as euxinic conditions took hold
curs in the Archaean, decreasing from around 2500–10,000 ppm to again in the Neoproterozoic (Poulton and Canfield, 2011).
just under 1000 ppm through the Palaeoproterozoic. This pattern, Selenium (Trend 3) (Fig. 6f), like Mo, is sourced from the ox-
also seen in Ni/Fe ratio in BIFs (Fig. 5c), is attributed to cooling idative weathering of continental pyrite, with release of selenite
upper mantle and decreased eruption of Ni-rich komatiite ultra- and selenate species (SeO23− and SeO24− ; see Supplementary Fig. 6),
mafic rocks from about 2700 Ma (Konhauser et al., 2009). Cobalt which accumulate in the ocean. Selenium is a nutrient TE, with
 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220 215

Fig. 6. a–f, Temporal trends of Mo, Ni, Co, As and Se measured by LA-ICPMS analysis of sedimentary pyrite in black shales. Each symbol represents a single pyrite analysis in
(a). The curves presented in (b–f) are derived from a 19 point rolling mean (or moving average).

significant biological uptake and eventual accumulation on the 3.1. Selenium as a proxy for deep time atmosphere–ocean oxygenation
seafloor in both organic matter and by substitution for S in sedi-
mentary pyrite (Diener and Neumann, 2011; Mitchell et al., 2012). In Fig. 8 we compare the temporal trend of Se in pyrite with
Se shows strong narrow peaks around 3000 and 2500 Ma, rising to Mo, U and Cr, three of the commonly employed whole rock
a flat trend between 2500 and 1700 Ma, with a major perturbation proxies for atmosphere–ocean oxygenation (Anbar et al., 2007;
Scott et al., 2008; Partin et al., 2013; Reinhard et al., 2013). All
from 1800 to 1300 Ma, then declines to a trough at 900 to 660
four redox-sensitive TE show related trends through the Archean
Ma, before rising again at GOE2. The stoichiometric substitution
and Proterozoic.
of Se for S in pyrite, its homogeneous distribution in sedimentary
The temporal trend of U (Fig. 8c) is broadly similar to Se
pyrite (Fig. 1e, Supplementary Fig. 7), high enrichment in pyrite
(Fig. 8b), stepping up through the late Archean and Early Protero-
versus the black shale matrix (Supplementary Table 4), low vari- zoic to a broad peak around 1700–1400 Ma, and then decreasing
ability in pyrite from the same sample (Supplementary Fig. 1), and into the Cryogenian TE trough from 1000 to 660 Ma. The broad U
tendency to remain in pyrite even after significant recrystallisation peak around the mid Proterozoic may relate to the abundance of
(Large et al., 2007; Thomas et al., 2011), indicate the Se content hot U-bearing granites in the Paleo- and Mesoproterozoic (Wilson
of sedimentary pyrite may be a superior oxygenation proxy com- and Akerblom, 1982; Wyborn et al., 1992) that would have pre-
pared to other commonly used proxies, as further discussed be- sented an enriched source rock of U for the contemporary oceans.
low. Chromium (Fig. 8d), like Ni and Co (Fig. 7), is elevated in the
216 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

Fig. 7. Intervals of major LIPs (Prokoph et al., 2004), juxtaposed with the time curves
for Co and Ni. Most highs for Ni and Co in sedimentary pyrite coincide with the
periods of high numbers of LIPs (green bars). The extreme minimum in Ni and Co
around 400 Ma is a good match for a period of minimum LIP activity from 300 to
500 Ma, whereas the maximum at 250 Ma matches the Siberian Trap LIP eruptions.

Archean due to the abundant ultramafic/mafic crustal components

and shows a decline from 2200 to 1200 Ma before rising to a
plateau around 800 Ma. The Cr data shows a very similar trend
to that recently published for organic-rich shales using whole-rock
analytical techniques (Reinhard et al., 2013).
Through the Archean, Se (Fig. 8b) averages around 10 ppm in
pyrite but with a spike at 2970 Ma and steps upwards between
2650 and 2500 Ma. The latter steps, to over 100 ppm Se, are many
tens of millions years earlier than the accepted position of GOE1
between 2450 to 2300 Ma (Farquhar et al., 2001; Holland, 2006;
Pufahl and Hiatt, 2012) and suggest earlier oxygenation pulses
(Anbar et al., 2007; Wille et al., 2007; Scott et al., 2008). The
spike at 2970 Ma, which is from pyrite in black shales from the
Witwatersrand Basin, may define the first pulse of global oxygena-
tion, and supports other recent evidence for an oxygenation pulse
determined by Cr isotopes and redox sensitive TE in the Pongola
Supergroup (Crowe et al., 2013). From 1700 to 700 Ma, Se shows
a gradual decline (Fig. 8b). This trend suggests the possibility of a
gradual drop in atmosphere oxygen through the Meso- and Neo-
proterozoic but with minor perturbations (particularly at 1700 to
1300 Ma). Our U and Cr curves measured on the black shale matrix
(Fig. 8c and d) show a similar decreasing trend during this period,
consistent with published whole-rock data (Partin et al., 2013).
Reinhard et al. (2013) have calibrated the palaeoredox proxies Cr
and Mo in the Proterozoic to delineate anoxic settings, where Cr is
reduced and buried, compared to euxinic settings, where both Cr
and Mo are reduced/sulfidised and buried. On this basis, our data Fig. 8. Temporal trends of a, Mo in sedimentary pyrite, b, Se in sedimentary pyrite,
(Fig. 8a and d) may be interpreted to suggest that anoxic settings c, U in shale matrix, and d, Cr in shale matrix; measured by LA-ICPMS displayed as
a 19 point rolling mean.
were dominant in the Proterozoic from 2150 to 700 Ma, whereas
euxinic settings (anoxic + H2 S) were dominant from 1200 to 800
Ma. This interpretation fits with the concept of decreasing atmo- 3.2. TE deficiency in the Meso- to Neoproterozoic Oceans
sphere oxygenation through the Meso- and Neoproterozoic, with
minimum O2 levels around 800 Ma, corresponding to the Cryoge- Anbar and Knoll (2002) present the hypothesis that under the
nian TE trough. conditions of a deep sulfidic ocean for most of the Meso- and Neo-
 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220 217

Fig. 9. Schematic of the major findings of this paper related to redox sensitive TE variations and oxygenation of the ocean–atmosphere. The red line is the estimated
oxygenation curve based on the Se content of sedimentary pyrite. The baseline % PAL (Present Atmospheric Level) is from Sahoo et al. (2012).

proterozoic (Canfield, 1998) nutrient TE would be scarce, thus lim- data, and these previous estimates, we speculate that atmospheric
iting the ecological distribution of eukaryotic algae and adversely O2 may have decreased from a level around 10–20% PAL at 1800
effecting marine evolution. Our data on the biologically important Ma to the Cryogenian TE trough at 1–5% PAL between 800 and 660
nutrient TE (Mo, Zn, Co, Cd, Se, Ni, Ag, Mn, U, V and Cr; Sup- Ma (Fig. 9).
plementary Figs. 3–5) supports this hypothesis and shows that
although these TE exhibit various trends in detail, they all show 4. Strengths, weaknesses and caveats of sedimentary pyrite
a relative depletion from 1200 to 700 Ma compared to their early proxies for ocean chemistry
Phanerozoic values, and in most cases their Neoarchean to Palaeo-
proterozoic values. For the redox sensitive TE that are employed The TE variation in sedimentary pyrite, determined by LA-
as atmosphere–ocean oxygenation proxies, Mo, U, Cr and Se there ICPMS, has the potential to enable the development of sev-
are two possible causes for these trends; (1) low values of these eral ocean–atmosphere proxies through time including: ocean TE
TE represent gradual drawdown of these elements in a predom- chemistry, ocean pH, ocean sulfate content, ocean–atmosphere
inantly sulfidic deep ocean (Anbar and Knoll, 2002; Reinhard et oxygenation, ocean anoxic events, ocean productivity and ocean
al., 2013), with limited renewal by continental erosion due to low hydrothermal activity. Although there are strengths associated with
levels of tectonic activity over the period 1800 to 700 Ma. Atmo- this approach, there are also several weaknesses and caveats.
spheric oxygen remains constant in this scenario (Canfield, 2005;
4.1. Strengths
Holland, 2006) Alternatively, (2) the decline in these redox sensi-
tive TE signal a gradual decrease in atmosphere oxygenation from
Sedimentary pyrite concentrates many of the redox sensitive
1800 to 800 Ma, with the lowest O2 levels forming the Cryoge-
and nutrient TE, in particular, Se, Mo, Cu, Ni, Ag, As, Cd, Co, Mn, Sb
nian TE trough between 800 and 660 Ma, prior to the rise at
well above their content in the sedimentary matrix of black shales
GOE2. Other workers (Shen et al., 2003; Planavsky et al., 2011;
(Fig. 4, Supplementary Information). This means, that by using the
Poulton and Canfield, 2011; Partin et al., 2013) provide evidence LA-ICPMS technique, the minimum (down to ppb levels) and max-
for sustained low levels of O2 during the mid Proterozoic, although imum (over wt% levels) of TE in pyrite can be measured accurately,
none have suggested a significant O2 decline during this period. enabling proxies for both maximum and minimum levels of TE in
In considering these two scenarios, there is little evidence to the oceans to be developed. This has not been previously possible.
support a static tectonic regime, as this period of the middle to late The LA-ICPMS technique used here enables 22 TE to be analysed
Proterozoic saw the amalgamation (1300–1000 Ma) and breakup simultaneously. This enables the trends of a suite of redox sensi-
(900–700 Ma) of the super continent Rodinia (Li et al., 2008; tive elements (e.g., Mo, Se, Cu, U and Cr) to be compared on the
Condie et al., 2011). Active erosion during these events would same sample set. Furthermore, TE in both the pyrite and sedimen-
involve a continuous supply of nutrient TE to the oceans. How- tary matrix are quantified, and thus TE concentrated in the matrix,
ever the normalised 87 Sr/86 Sr curve of riverine runoff compared (e.g., U, V, W, Sn and Ba) can also be tracked through time.
with mantle influence (Shields, 2007), indicates a decline in ero- Pyrite analyses involve only one mineral, and thus there are less
sive supply of Sr which parallels the Se and U time curves. On complications than whole rock analyses of black shales, where TE
the other hand, low O2 may have caused the nutrient deficiency, are partitioned amongst a number of minerals and organic matter,
which is corroborated by carbon isotope evidence for low pro- of varying proportions, plus seawater and pore fluids. Sedimentary
ductivity (Anbar and Knoll, 2002). Canfield (2005) suggests that pyrite forms in black shales from sub-oxic to euxinic environments,
atmospheric oxygen rose to a level of 5 to 10% Present Atmo- and thus sampling is not restricted to a particular redox facies in
spheric Level (PAL) around 1800 Ma and remained constant until order to apply the technique. This enables a much larger popula-
about 700 Ma. Planavsky et al. (2011) suggest a lower level of 1 to tion of black shales to be included in deep time research. A de-
10% PAL throughout the Meso- and Neoproterozoic. Based on our tailed petrographic study of sedimentary pyrite allows the effects
218 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

of late diagenesis, metamorphism and hydrothermal overprints to may contain diagenetic nodules or partially recrystallised fram-
be fully assessed and thus avoided in sample selection. boids. Great care is required to record the textural and paragenetic
changes in pyrite in order to undertake analyses on the early-
4.2. Weaknesses formed pyrite or their partially recrystallised equivalents.
In order to place pyrite TE analyses in a temporal context, high
Sedimentary pyrite develops both in the water column (syn- quality geochronology is required. For Phanerozoic samples, fossil
genetic) and in the sea floor muds (diagenetic), and yet the tech- assemblages can yield detailed chronological resolution. In Pre-
nique proposed here does not take into account the modification cambrian settings, the U–Pb analysis of zircon can provide high
of seawater TE chemistry by interaction with minerals, other than quality constraints on the timing of eruption of tuffs and/or max-
pyrite, in the muds. Our analyses are a mixture of syngenetic and imum depositional ages for silt- and sand-dominated strata that
early diagenetic framboidal pyrites plus nodular pyrites that de- may be interbedded with black shale sequences. U-bearing au-
veloped during late diagenesis, probably several meters below the thigenic mineral phases also provide at least a minimum age of
seafloor. Diagenetic nodular pyrite that forms during late diagen- deposition of the host rocks, and in some cases may date diagen-
esis, commonly contains lower TE concentrations (Fig. 2; Large et esis. Where available, we have preferred chronological constraints
al., 2007), but no more than an order of magnitude less than syn- based on these techniques in order to minimise uncertainties re-
genetic and early diagenetic pyrite in the same sedimentary unit. lated to whole rock dating methods, however the age resolution
and uncertainties vary significantly (Supplementary Table 2). A ma-
We argue that seawater is the principal component of pore flu-
jor caveat of any temporal analysis such as this is the accuracy and
ids, and as framboidal pyrite develops in the top few centimetres
resolution of chronological constraints.
of the muds, when seawater is >60% of the muds, then a strong
chemical connection between pyrite and seawater is maintained.
5. Conclusions
These same assumptions apply in the use of strontium, carbon and
oxygen isotopes of sedimentary carbonate as a proxy for their re-
The use of sedimentary pyrite as a proxy for TE trends in
spective isotope variations in the palaeo-ocean (Veizer et al., 1999;
the palaeo-ocean has opened a new window into deep time, and
Shields and Veizer, 2002).
presents exciting possibilities. Several lines of evidence indicate
The TE in early formed (syngenetic and early diagenetic) pyrite
that for any particular black shale depositional environment, syn-
can be strongly modified by later diagenetic, metamorphic or hy-
genetic pyrite that forms in the water column and diagenetic pyrite
drothermal processes and thus have little relationship to the orig-
that forms in the top few meters of the seawater-saturated muds,
inal seawater chemistry. This is a major issue, especially for those
will have a similar first order TE composition that relates to the
elements loosely held in the pyrite structure, such as Mo (Chappaz
TE composition of the seawater at that time. Processes during di-
et al., 2014). However, the approach proposed here is to under-
agenesis do not substantially change the mean first order TE con-
take petrographic studies of pyrite textures in order to screen out
tent of sedimentary pyrite. However, metamorphic processes that
samples with evidence of metamorphic and hydrothermal influ-
cause recrystallisation of pyrite and conversion to pyrrhotite at
ences (e.g., Thomas et al., 2011). Selenium in pyrite is very robust
moderate to high metamorphic grades, may be accompanied by
to metamorphic influences and is the best of the redox sensitive TE
major changes in many TE concentrations. Therefore, this study
(Fig. 9). Black shales and associated pyrite are of little use as sea-
has concentrated on early-formed (syngenetic to diagenetic) pyrite
water proxies for samples above middle greenschist facies due to
and avoided samples with obvious metamorphic or hydrothermal
major TE exchange with metamorphic fluids (Pitcairn et al., 2006;
pyrite and/or pyrrhotite.
Large et al., 2011b).
Our data suggest that for most TE, temporal variations in con-
Our data indicate a wide variation in the TE content of sedi-
centration have been over 2 to 4 orders of magnitude. The concen-
mentary pyrite from the same sample. However, many TE show a
trations of some TE, like Ni, Co and Cr, generally decrease through
variation of less than one to two orders of magnitude. This degree
time, but with significant variation, commonly related to changes
of variation could be considered to relate to ocean TE variation,
in the composition of the exposed crust. Other redox sensitive TE
or to the inherent variability of syngenetic and early diagenetic
like Mo and Se, generally increase through time, and are princi-
pyrite, and the effects of diagenesis and low-grade metamorphism.
pally controlled by oxygenation trends in the ocean–atmosphere
The variation of one order of magnitude in a single sample is
system.
significantly less that the variation of up to 4 orders of magni-
A schematic diagram of our major findings is given in Fig. 9.
tude in the full dataset through time. Studies on single drill core
This figure uses the Se content of sedimentary pyrite as a proxy for
(Supplementary Fig. 1) show the variation in individual samples
variations in atmosphere oxygenation. This proxy is based on the
compared to the mean variation through a stratigraphic section,
premise that oxidative weathering of pyrite in continental rocks
and clearly demonstrate that robust trends in the TE are related to leads to the release of selenium as both the selenate and selen-
stratigraphic position. ite species. Under neutral to alkaline oxygenated conditions the
In addition to the concentrations of TE in seawater, a number selenate species remains highly soluble, where it can be read-
of other factors may have an effect on the TE content of diage- ily transported via river systems to the ocean. However under
netic pyrite, including; the matrix composition of the sediment, low oxygen conditions, continental weathering releases little sol-
the amount of pyrite produced, pH of pore fluids, degree of biopro- uble Se, because the reduced forms of Se (selenide, elemental Se0
ductivity and preferred electron acceptors. Although these factors and organo-selenium complexes) are relatively insoluble. Thus sig-
are assumed here to have a second order effect, they each need nificant increases in atmospheric oxygen, accompanied by active
further evaluation. erosion, can lead to a major increase in the supply of soluble Se
to the ocean, to be subsequently trapped in sedimentary pyrite in
4.3. Caveats seafloor muds.
The Se curve derived from sedimentary pyrite, along with
The most coherent TE data on early-formed sedimentary pyrite curves for other redox sensitive TE (e.g., Mo, U, Tl, Mn, Cd, Ag),
is obtained by the analysis of pyrite framboids, and these must be show significant increases in TE abundance of one to two orders
the first priority in any analytical strategy. However many black of magnitude at GOE1 and GOE2 (Fig. 9). However our data for
shales of greenschist facies do not contain framboidal pyrite, but some TE shows sharp rises prior to GOE1, suggesting pulses of
 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220 219

oxygenation commenced as early as 2970 Ma. Through the Protero- Crowe, S.A., Døssing, L.N., Beukes, N.J., Bau, M., Kruger, S.J., Frei, R., Canfield, D.E.,
zoic, TE variations in sedimentary pyrite are subdued, compared 2013. Atmospheric oxygenation three billion years ago. Nature 501, 535–538.
Dahl, T.W., Canfield, D.E., Rosing, M.T., Frei, R.E., Gordon, G.W., Knoll, A.H., An-
with the Archaean, suggesting steady state or declining oxygen lev-
bar, A.D., 2011. Molybdenum evidence for expansive sulfidic water masses in
els reaching a minimum in the Cryogenian prior to GOE2. From ∼750 Ma oceans. Earth Planet. Sci. Lett. 311, 264–274.
1800 to 700 Ma a number of biologically important TE (e.g., Se, Cr, Danyushevsky, L., Robinson, P., Gilbert, S., Norman, M., Large, R., McGoldrick, P., Shel-
Ag, Mo, Cu, Ni, Mn) show a flat to decreasing trend suggesting that ley, M., 2011. Routine quantitative multi-element analysis of sulphide minerals
nutrient TE deficiency may be the reason for limited eukaryotic by laser ablation ICP-MS: Standard development and consideration of matrix ef-
fects. Geochem., Explor. Environ. Anal. 11, 51–60.
development and radiation, thus adversely effecting marine evolu-
Deditius, A.P., Utsunomiya, S., Reich, M., Kesler, S.E., Ewing, R.C., Hough, R., Walshe,
tion during the middle to late Proterozoic. This period ended with J., 2011. Trace metal nanoparticles in pyrite. Ore Geol. Rev. 42, 32–46.
the Sturtian glaciation, and from 700 to 540 Ma many of the re- Diener, A., Neumann, T., 2011. Synthesis and incorporation of selenide in pyrite and
dox sensitive TE rose by 1 to 2 orders of magnitude in response mackinawite. Radiochim. Acta 99, 791–798.
Farquhar, J., Savarino, J., Airieau, S., Thiemens, M.H., 2001. Observation of
to GOE2, accompanied by diversification of phytoplankton and the
wavelength-sensitive mass-independent sulfur isotope effects during SO2 pho-
appearance of eukaryotic clades (Shields-Zhou and Och, 2011); the tolysis: Implications for the early atmosphere. J. Geophys. Res., Planets 106,
prelude to the Cambrian Explosion. 32829–32839.
Fike, D.A., Grotzinger, J.P., Pratt, L.M., Summons, R.E., 2006. Oxidation of the Edi-
Acknowledgements acaran ocean. Nature 444, 744–747.
Gregory, D., Meffre, S., Large, R.R., 2013. Mineralogy of metal contaminated estuarine
sediments, Derwent estuary, Hobart, Australia: implications for metal mobility.
The following colleagues have provided black shale samples to- Aust. J. Earth Sci. 60, 589–603.
ward this project: Robert Scott, Ray Coveney Jr., Jeffrey Abbott, Gregory, D.D., Meffre, S., Large, R.R., in press. Comparison of metal enrichment in
Bradley Guy, David Huston, Lex Lambeck, Jeffrey Steadman, Charles pyrite framboids from a metal-enriched and metal poor estuary. Am. Mineral.
http://dx.doi.org/10.2138/am.2014.4545.
Makoundi, Sean Johnson, Richard Batchelor, Stuart Smith, Mikhail
Guy, B.M., Beukes, N.J., Gutzmer, J., 2010. Paleoenvironmental controls on the texture
Krupenin, Jerry Sharrock, Peter Sorjonen-Ward, Luba Leonova, Va- and chemical composition of pyrite from non-conglomeratic sedimentary rocks
leriy Murzin and Sergey Karpov. Funding was provided by an ARC of the Mesoarchean Witwatersrand Supergroup, South Africa. S. Afr. J. Geol. 113,
Centre of Excellence grant CE0561595 and AMIRA International 195–228.
Harrold, T.W.D., Swarbrick, R.E., Goulty, N.R., 1999. Pore pressure estimation from
project grant P923 to RRL. Ron Fuge (University of Wales), Thomas
Mudrock porosities in tertiary basins, southeast Asia. Am. Assoc. Pet. Geol.
Kulp (Binghamton University), Bernd Lottermoser, Patrick Quilty, Bull. 83, 1057–1066.
Jocelyn McPhie (University of Tasmania), Brian Skinner (Yale Uni- Helz, G.R., Miller, C.V., Charnock, J.M., Mosselmans, J.F.W., Pattrick, R.A.D., Garner,
versity) and Noel White (Consultant) are thanked for their com- C.D., Vaughan, D.J., 1996. Mechanism of molybdenum removal from the sea
ments on earlier versions of the manuscript. We thank B. Kamber and its concentration in black shales: EXAFS evidence. Geochim. Cosmochim.
Acta 60, 3631–3642.
and three other reviewers, and Editor G. Henderson, for excellent
Hoffman, P.F., Kaufman, A.J., Halverson, G.P., Schrag, D.P., 1998. A Neoproterozoic
comments for revision and improvement of the final manuscript. snowball earth. Science 281, 1342–1346.
This is YGS publication number 019. Holland, H.D., 1984. The Chemical Evolution of the Atmosphere and Oceans. Prince-
ton University Press, Princeton.
Holland, H.D., 2006. The oxygenation of the atmosphere and oceans. Philos. Trans.
Appendix A. Supplementary material
R. Soc. Lond. B, Biol. Sci. 361, 903–915.
Huerta-Diaz, M.A., Morse, J.W., 1990. A quantitative method for determination of
Supplementary material related to this article can be found on- trace metal concentrations in sedimentary pyrite. Mar. Chem. 29, 119–144.
line at http://dx.doi.org/10.1016/j.epsl.2013.12.020. Huerta-Diaz, M.A., Morse, J.W., 1992. Pyritization of trace metals in anoxic marine
sediments. Geochim. Cosmochim. Acta 56, 2681–2702.
Konhauser, K.O., Pecoits, E., Lalonde, S.V., Papineau, D., Nisbet, E.G., Barley, M.E.,
References Arndt, N.T., Zahnle, K., Kamber, B.S., 2009. Oceanic nickel depletion and a
methanogen famine before the Great Oxidation Event. Nature 458, 750–753.
Algeo, T.J., Rowe, H., 2012. Paleoceanographic applications of trace-metal concentra- Kulp, T.R., Pratt, L.M., 2004. Speciation and weathering of selenium in Upper Cre-
tion data. Chem. Geol. 324, 6–18. taceous chalk and shale from South Dakota and Wyoming, USA. Geochim. Cos-
Anbar, A.D., Knoll, A.H., 2002. Proterozoic ocean chemistry and evolution: A bioinor- mochim. Acta 68, 3687–3701.
ganic bridge?. Science 297, 1137–1142. Large, R.R., Maslennikov, V.V., Robert, F., Danyushevsky, L.V., Chang, Z., 2007. Mul-
Anbar, A.D., Duan, Y., Lyons, T.W., Arnold, G.L., Kendall, B., Creaser, R.A., Kaufman, tistage sedimentary and metamorphic origin of pyrite and gold in the Giant
A.J., Gordon, G.W., Scott, C., Garvin, J., Buick, R., 2007. A whiff of oxygen before Sukhoi log deposit, Lena Gold Province, Russia. Econ. Geol. 102, 1233–1267.
the great oxidation event?. Science 317, 1903–1906. Large, R.R., Danyushevsky, L., Hollit, C., Maslennikov, V., Meffre, S., Gilbert, S., Bull, S.,
Bajwah, Z.U., Seccombe, P.K., Offler, R., 1987. Trace element distribution, Co:Ni ratios Scott, R., Emsbo, P., Thomas, H., Singh, B., Foster, J., 2009. Gold and trace element
and genesis of the Big Cadia iron-copper deposit, New South Wales, Australia. zonation in pyrite using a laser imaging technique: Implications for the timing
Miner. Depos. 22, 292–300. of gold in orogenic and carlin-style sediment-hosted deposits. Econ. Geol. 104,
Baldwin, B., Butler, C.O., 1985. Compaction curves. Am. Assoc. Pet. Geol. Bull. 69, 635–668.
622–626. Large, R.R., Bull, S.W., Gilbert, S., 2011a. Multiple syngenetic metalliferous horizons
Berner, Z.A., Puchelt, H., Nöltner, T., Kramar, U., 2013. Pyrite geochemistry in enriched in V–As–Mo–Se–Ni–Ag–Zn–Au in the Northern Carlin Trend: impli-
the Toarcian Posidonia Shale of south-west Germany: Evidence for contrasting cations for the genesis of world class gold-arsenic deposits. In: Steininger, R.,
trace-element patterns of diagenetic and syngenetic pyrites. Sedimentology 60, Pennell, B. (Eds.), Geological Society of Nevada 2010 Symposium: Great Basin
548–573. Evolution and Metallogeny. DEStech Publications Inc., Reno, pp. 1–16.
Canfield, D.E., 1998. A new model for Proterozoic ocean chemistry. Nature 396, Large, R.R., Bull, S.W., Maslennikov, V.V., 2011b. A carbonaceous sedimentary
450–453. source-rock model for carlin-type and orogenic gold deposits. Econ. Geol. 106,
Canfield, D.E., 2005. The early history of atmospheric oxygen: Homage to Robert M. 331–358.
Garrels. Annu. Rev. Earth Planet. Sci. 33, 1–36. Large, R.R., Meffre, S., Burnett, R., Guy, B., Bull, S., Gilbert, S., Goemann, K., Danyu-
Canfield, D.E., Poulton, S.W., Narbonne, G.M., 2007. Late-Neoproterozoic deep-ocean shevsky, L., 2013. Evidence for an intrabasinal source and multiple concentration
oxygenation and the rise of animal life. Science 315, 92–95. processes in the formation of the carbon leader reef, Witwatersrand Supergroup,
Chappaz, A., Lyons, T.W., Gregory, D.D., Reinhard, C.T., Gill, B.C., Li, C., Large, R.R., South Africa. Econ. Geol. 108, 1215–1241.
2014. Does pyrite act as an important host for molybdenum in modern and Li, Z.X., Bogdanova, S.V., Collins, A.S., Davidson, A., De Waele, B., Ernst, R.E., Fitzsi-
ancient euxinic sediments?. Geochim. Cosmochim. Acta 126, 112–122. mons, I.C.W., Fuck, R.A., Gladkochub, D.P., Jacobs, J., Karlstrom, K.E., Lu, S., Nat-
Ciobanu, C.L., Cook, N.J., Pring, A., Brugger, J., Danyushevsky, L.V., Shimizu, M., apov, L.M., Pease, V., Pisarevsky, S.A., Thrane, K., Vernikovsky, V., 2008. Assem-
2009. ‘Invisible gold’ in bismuth chalcogenides. Geochim. Cosmochim. Acta 73, bly, configuration, and break-up history of Rodinia: A synthesis. Precambrian
1970–1999. Res. 160, 179–210.
Condie, K.C., Bickford, M.E., Aster, R.C., Belousova, E., Scholl, D.W., 2011. Episodic Lyons, T.W., Werne, J.P., Hollander, D.J., Murray, R.W., 2003. Contrasting sulfur geo-
zircon ages, Hf isotopic composition, and the preservation rate of continental chemistry and Fe/Al and Mo/Al ratios across the last oxic-to-anoxic transition in
crust. Geol. Soc. Am. Bull. 123, 951–957. the Cariaco Basin, Venezuela. Chem. Geol. 195, 131–157.
220 R.R. Large et al. / Earth and Planetary Science Letters 389 (2014) 209–220

Meffre, S., Large, R.R., Scott, R., Woodhead, J., Chang, Z., Gilbert, S.E., Danyushevsky, T.W., Jiang, G., 2012. Ocean oxygenation in the wake of the Marinoan glaciation.
L.V., Maslennikov, V., Hergt, J.M., 2008. Age and pyrite Pb-isotopic composition Nature 489, 546–549.
of the giant Sukhoi Log sediment-hosted gold deposit, Russia. Geochim. Cos- Scott, C., Lyons, T.W., Bekker, A., Shen, Y., Poulton, S.W., Chu, X., Anbar, A.D.,
mochim. Acta 72, 2377–2391. 2008. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452,
Mitchell, K., Mason, P.R.D., Van Cappellen, P., Johnson, T.M., Gill, B.C., Owens, 456–459.
J.D., Diaz, J., Ingall, E.D., Reichart, G.J., Lyons, T.W., 2012. Selenium as paleo- Shen, Y., Knoll, A.H., Walter, M.R., 2003. Evidence for low sulphate and anoxia in a
oceanographic proxy: A first assessment. Geochim. Cosmochim. Acta 89, mid-Proterozoic marine basin. Nature 423, 632–635.
302–317. Shields, G.A., 2007. A normalised seawater strontium isotope curve: possible impli-
Morse, J.W., Arakaki, T., 1993. Adsorption and coprecipitation of divalent metals with cations for Neoproterozoic-Cambrian weathering rates and the further oxygena-
mackinawite (FeS). Geochim. Cosmochim. Acta 57, 3635–3640. tion of the Earth. eEarth 2, 35–42.
Partin, C.A., Bekker, A., Planavsky, N.J., Scott, C.T., Gill, B.C., Li, C., Podkovyrov, V., Shields, G., Veizer, J., 2002. Precambrian marine carbonate isotope database: Version
Maslov, A., Konhauser, K.O., Lalonde, S.V., Love, G.D., Poulton, S.W., Lyons, T.W., 1.1. Geochem. Geophys. Geosyst. 3, 1–12.
2013. Large-scale fluctuations in Precambrian atmospheric and oceanic oxygen Shields-Zhou, G., Och, L., 2011. The case for a neoproterozoic oxygenation event:
levels from the record of U in shales. Earth Planet. Sci. Lett. 369–370, 284–293. Geochemical evidence and biological consequences. GSA Today 21, 4–11.
Piper, D.Z., Perkins, R.B., 2004. A modern vs. Permian black shale – the hydrog-
Thomas, B.M., Willink, R.J., Grice, K., Twitchett, R.J., Purcell, R.R., Archbold, N.W.,
raphy, primary productivity, and water-column chemistry of deposition. Chem.
George, A.D., Tye, S., Alexander, R., Foster, C.B., Barber, C.J., 2004. Unique marine
Geol. 206, 177–197.
Permian-Triassic boundary section from Western Australia. Aust. J. Earth Sci. 51,
Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S., 2006.
423–430.
Sources of metals and fluids in orogenic gold deposits: Insights from the Otago
Thomas, H.V., Large, R.R., Bull, S.W., Maslennikov, V., Berry, R.F., Fraser, R., Froud,
and Alpine schists, New Zealand. Econ. Geol. 101, 1525–1546.
S., Moye, R., 2011. Pyrite and pyrrhotite textures and composition in sediments,
Planavsky, N.J., McGoldrick, P., Scott, C.T., Li, C., Reinhard, C.T., Kelly, A.E., Chu, X.,
laminated quartz veins, and reefs at Bendigo gold mine, Australia: Insights for
Bekker, A., Love, G.D., Lyons, T.W., 2011. Widespread iron-rich conditions in the
ore genesis. Econ. Geol. 106, 1–31.
mid-Proterozoic ocean. Nature 477, 448–451.
Tribovillard, N., Algeo, T.J., Lyons, T., Riboulleau, A., 2006. Trace metals as paleoredox
Poulton, S.W., Canfield, D.E., 2011. Ferruginous conditions: a dominant feature of the
and paleoproductivity proxies: An update. Chem. Geol. 232, 12–32.
ocean through Earth’s history. Elements 7, 107–112.
Prokoph, A., Ernst, R.E., Buchan, K.L., 2004. Time-series analysis of large igneous Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, F., Garden, G.A.F., Di-
provinces: 3500 Ma to present. J. Geol. 112, 1–22. ener, A., Ebneth, S., Godderis, Y., Jasper, T., Korte, C., Pawellek, F., Podlaha, O.G.,
Pufahl, P.K., Hiatt, E.E., 2012. Oxygenation of the Earth’s atmosphere–ocean system: Strauss, H., 1999. 87 Sr/86 Sr, δ 13 C and δ 18 O evolution of Phanerozoic seawater.
A review of physical and chemical sedimentologic responses. Mar. Pet. Geol. 32, Chem. Geol. 161, 59–88.
1–20. Wilkin, R.T., Barnes, H.L., Brantley, S.L., 1996. The size distribution of framboidal
Reimann, C., Caritat, P.d., 1998. Chemical Elements in the Environment: Factsheets pyrite in modern sediments: An indicator of redox conditions. Geochim. Cos-
for the Geochemist and Environmental Scientist. Springer-Verlag. mochim. Acta 60, 3897–3912.
Reinhard, C.T., Planavsky, N.J., Robbins, L.J., Partin, C.A., Gill, B.C., Lalonde, S.V., Wille, M., Kramers, J.D., Nägler, T.F., Beukes, N.J., Schröder, S., Meisel, T., Lacassie,
Bekker, A., Konhauser, K.O., Lyons, T.W., 2013. Proterozoic ocean redox and bio- J.P., Voegelin, A.R., 2007. Evidence for a gradual rise of oxygen between 2.6 and
geochemical stasis. Proc. Natl. Acad. Sci. USA 110, 5357–5362. 2.5 Ga from Mo isotopes and Re-PGE signatures in shales. Geochim. Cosmochim.
Rickard, D., 2012. Sulfidic sediments and sedimentary rocks. In: Van Loon, A.J. (Ed.), Acta 71, 2417–2435.
Developments in Sedimentology. Elsevier, p. 801. Wilson, M.R., Akerblom, G.V., 1982. Geological setting in geochemistry of uranium-
Rickard, D., Luther, G.W.I.I., 1997. Kinetics of pyrite formation by the H2S oxida- rich granites in the Proterozoic of Sweden. Mineral. Mag. 46, 235–247.
tion of iron (II) monosulfide in aqueous solutions between 25 and 125 ◦ C: The Wyborn, L.A.I., Wyborn, D., Warren, R.G., Drummond, B.J., 1992. Proterozoic gran-
mechanism. Geochim. Cosmochim. Acta 61, 135–147. ite types in Australia: implications for lower crust composition, structure and
Sahoo, S.K., Planavsky, N.J., Kendall, B., Wang, X., Shi, X., Scott, C., Anbar, A.D., Lyons, evolution. Trans. R. Soc. Edinb. Earth Sci. 83, 201–209.