Vol 461 | 10 September 2009 | doi:10.1038/nature08266
LETTERS
Fluctuations in Precambrian atmospheric
oxygenation recorded by chromium isotopes
Robert Frei1, Claudio Gaucher1,2, Simon W. Poulton3 & Don E. Canfield4
Geochemical data1–4 suggest that oxygenation of the Earth’s atmosphere occurred in two broad steps. The first rise in atmospheric
oxygen is thought to have occurred between 2.45 and 2.2 Gyr
ago1,5, leading to a significant increase in atmospheric oxygen concentrations and concomitant oxygenation of the shallow surface
ocean. The second increase in atmospheric oxygen appears to have
taken place in distinct stages during the late Neoproterozoic era
( 800–542 Myr ago)3,4, ultimately leading to oxygenation of the
deep ocean 580 Myr ago3, but details of the evolution of atmospheric oxygenation remain uncertain. Here we use chromium
(Cr) stable isotopes from banded iron formations (BIFs) to track
the presence of Cr(VI) in Precambrian oceans, providing a timeresolved picture of the oxygenation history of the Earth’s
atmosphere–hydrosphere system. The geochemical behaviour of
Cr is highly sensitive to the redox state of the surface environment
because oxidative weathering processes produce the oxidized
hexavalent [Cr(VI)] form. Oxidation of reduced trivalent
[Cr(III)] chromium on land is accompanied by an isotopic fractionation, leading to enrichment of the mobile hexavalent form in
the heavier isotope. Our fractionated Cr isotope data indicate the
accumulation of Cr(VI) in ocean surface waters 2.8 to 2.6 Gyr ago
and a likely transient elevation in atmospheric and surface ocean
oxygenation before the first great rise of oxygen 2.45–2.2 Gyr ago
(the Great Oxidation Event)1,5. In 1.88-Gyr-old BIFs we find that
Cr isotopes are not fractionated, indicating a decline in atmospheric oxygen. Our findings suggest that the Great Oxidation
Event did not lead to a unidirectional stepwise increase in atmospheric oxygen. In the late Neoproterozoic, we observe strong
positive fractionations in Cr isotopes (d53Cr up to 14.9%),
providing independent support for increased surface oxygenation
at that time, which may have stimulated rapid evolution of
macroscopic multicellular life3,4,6.
The mobile Cr(VI) anion (HCrO42) is the most thermodynamically
stable Cr form in equilibrium with present-day air. Oxidation of
Cr(III) to Cr(VI) in soils depends upon the co-occurrence of Cr(III)
(bound most commonly as FeCr2O4) and manganese oxides (catalysing Cr(III) oxidation). Once mobilized during oxidative weathering,
Cr(VI) is mobile as either chromate (CrO422; alkalic pH) or bichromate (HCrO42; acidic pH) ions, entering the oceans via riverine transport7. There is a considerably smaller input of Cr from atmospheric
and hydrothermal vent sources. In today’s oceans, total dissolved Cr
concentrations are in the range of 2 to 10 nM with a relatively short
residence time of ,2.5 to 4 3 104 years8.
Cr(VI) can be reduced to Cr(III) by microbes9 and by aqueous
Fe(II) or Fe(II)-bearing minerals10 (see equation (1)). Indeed, the
oxidation of Fe(II) (aq) by Cr(VI) proceeds faster than with oxygen,
even under well-aerated, high-pH conditions11. This means that in
the presence of Fe(II), Cr(IV) is efficiently reduced to Cr(III). The
Cr(III) is subsequently and effectively scavenged into Fe(III)–Cr(III)
oxyhydroxides12 owing to the very low solubility of Fe,Cr(OH)3
solids13. Some Cr(III) can be regenerated and lost from sediments
as a result of Fe oxide reduction, but, as on land, the Cr(III) is
reoxidized rapidly14 to Cr(VI) in a catalytic reaction with MnO2
(ref. 7):
Cr(VI) (aq) 1 3Fe(II) (aq) R Cr(III) (aq) 1 3Fe(III) (aq) (1)
At equilibrium, the Cr(VI)O42– anion is enriched by up to 7% at
room temperature in 53Cr compared to coexisting compounds containing Cr(III) (we use the delta notation relative to the certified
National Bureau of Standards Cr reference standard SRM 979,
defined as d53Cr 5 1,000 3 [(53Cr/52Cr)sample/(53Cr/52Cr)SRM 979)2 1)
(see ref. 15). Therefore, subsurface aqueous environments will have
positive d53Cr values16. Although the isotopic composition of Cr in sea
water has not yet been measured, the positive groundwater Cr(VI)
signal should be transferred to the sea, because subsequent adsorption
of Cr onto particles (as might occur in soils and rivers) produces no
isotope effect17. The microbial reduction of Cr(VI) generates isotopic
shifts of up to 24.1%, comparable to those produced during abiotic
reduction9,10. This will potentially enrich the heavier isotope in the
remaining, unreacted, dissolved Cr(VI). However, because of the efficient sequestration of Cr(VI) during Cr reduction and subsequent
precipitation of Cr(III) with Fe-oxyhydroxides, the stable Cr isotope
signatures of chemically precipitated Fe(III)-rich sediments should
mirror the sea water from which the Fe oxides precipitated. The
surface chemistry of Cr and its stable isotope geochemistry are
summarized in Fig. 1.
The prerequisite for Cr isotopes to record the presence of Cr(VI) in
sea water is a predominance of dissolved Fe(II), which acts as the
reductant. Therefore, the isotopic composition of Cr in ancient ironrich sediments should provide a first-order proxy for the presence of
Cr(VI) in ancient surface waters, and thus the history of the oxidative
weathering of Cr on land. This approach should be relatively insensitive to the type of iron-rich chemical sediments and the palaeoenvironment in which these were deposited.
Oxidation and solubilization of Cr from soils is strongly dependent on the presence of MnO2, which is stable under elevated oxygen
fugacities, so the pathway of Cr to the oceans in the early Precambrian
would have been limited by the absence of Mn(IV) under low atmospheric oxygen pressures. The geochemical behaviour of Cr in sea
water is therefore highly sensitive to levels of atmospheric oxygen (see
Supplementary Information for further details).
We analysed d53Cr values of numerous Precambrian BIFs
(Supplementary Table 1), from which we delineate six stages of Cr
cycling (Fig. 2). During stage 1, comprising BIFs deposited during
the Archaean from 3.7–2.8 Gyr, the d53Cr values are unfractionated
1
Institute of Geography and Geology and Nordic Center for Earth Evolution (NordCEE), University of Copenhagen, Øster Voldgade 10, 1350 Copenhagen, Denmark. 2Departamento de
Geologı́a, Facultad de Ciencias, Iguá 4225, 11400 Montevideo, Uruguay. 3School of Civil Engineering and Geosciences, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK.
4
Nordic Center for Earth Evolution (NordCEE) and Institute of Biology, University of Southern Denmark, Campusvej 55, 5230 Odense, Denmark.
250
©2009 Macmillan Publishers Limited. All rights reserved
LETTERS
NATURE | Vol 461 | 10 September 2009
compared to the Earth’s high-temperature igneous reservoirs18. This
indicates a lack of oxidative continental weathering during most of the
Archaean. In stage 2, during the latest Archaean and before the Great
Oxidation Event (GOE) in the early Proterozoic (,2.8–2.45 Gyr ago),
four out of seven BIFs show positively fractionated d53Cr values of
10.04 to 10.29% (Fig. 2, Supplementary Table 1). We interpret these
enrichments to reflect transient occurrences of slightly elevated oxygen.
This is consistent with Mo concentration and S isotope evidence for
ephemeral surface water (and possibly atmospheric) oxygenation in
the run-up to the GOE1,5,19–22, although our results would suggest that
Cr(VI) was mobilized up to 300 Myr before the GOE.
Stage 3, from ,2.45 to 1.9 Gyr, is defined by rare BIF deposition,
particularly during the GOE itself. This apparent absence of BIFs could
reflect a transition from Fe-rich to Fe-poor oceans. It is interesting that
the BIFs just predating the GOE (the youngest samples in stage 2
deposited about 2.5–2.4 Gyr ago) show little Cr enrichment, and were
presumably deposited just before the GOE and the associated rise in
atmospheric O2. Major BIFs were again deposited in North America,
India and Australia23 at around 2.1 Gyr (our samples come from South
Dakota and the Bastar Craton, Supplementary Table 1). These stage 3
BIFs display some positive d53Cr values when compared to the range of
magmatic values and compared to stage 1 BIFs (Fig. 2), and roughly
coincide with BIFs characterized by positively fractionated d56Fe
values2 of sedimentary pyrite, interpreted to reflect an increase in
the precipitation of iron sulphides relative to iron oxides in a redox
stratified ocean. These fractionated Cr isotopes follow the GOE by
some 200 to 300 Myr and would be consistent with concomitant elevated levels of atmospheric oxygenation. However, measured Cr fractionations are lower than the elevated fractionations observed in stage
2, a time where consensus would argue for very low levels of atmospheric O2 punctuated by occasional ‘whiffs’ of oxygen19–21. As we shall
see below, there is evidence that atmospheric oxygen declined after the
GOE. Therefore, lower fractionations in late stage 3 could, in part, have
resulted from a return to reduced oxygen levels.
Indeed, the oldest of our stage 4 samples, which come from the
Gunflint Iron Formation (Ontario, Canada), do not show any
positively fractionated d53Cr values (Fig. 2; Supplementary Table 1).
This is further evidence for a decrease in atmospheric oxygen after
the GOE. Supporting evidence for low atmospheric oxygen comes
from the precipitation of iron oxides in the high-energy near-shore
region as observed in the 1.88-Gyr-old Gunflint Iron Formation24. This
required the transport of dissolved Fe21 over a broad continental shelf
to the palaeoshoreline and atmospheric O2 concentrations no greater
io n
Re
Mn2+
idati
on
Ox
Atmospheric O2
du ct
Cr(III)
MnO2
(1)
Cr(VI)
Cr
(V
Soil
I)
~~~~~~~~~~~~~~~~~~~~~~~~
(2) Cr(III)
Fe(II)
Fe(III)
Fe(III), Cr(III) oxides
Cr(VI)
(3)
Water
Fe(II)
± Cr(III)
BIF
BIFs
Groundwaters
Igneous rocks
Adsorption/precipitation
Reduction, Cr(III) –[CrO4]2–
Equilibrium, [CrO4]2– – Cr2O3
Isotopic compositions
Fractionations
–4
Hydrothermal
vents
–2
0
δ53Cr
2
4
6
8
Figure 1 | Schematic of the surface chemistry of chromium. (1): Oxidation
of Cr(III) in soils is catalysed by MnO2 and positively fractioned Cr(VI)10,16
enters the aquatic phase (groundwater, rivers) mainly as HCrO42
complexes, and eventually enters the ocean. (2): Abiotic reduction of Cr(VI)
by upwelling Fe(II)10 is efficient, fast and complete. (3): Subsequent
scavenging of Cr by Fe–Cr oxyhydroxides is a major removal pathway of Cr
into the sedimentary environment. The positively fractionated Cr
(d53Cr < 20.3% to 4.9%; this study) in BIFs and Fe-rich cherts thereby
mirrors the riverine Cr(VI) input. Biotic (bacterial) reduction of Cr(VI) has
recently been reported with Cr isotopic shifts (D53CrCr(III)–Cr(VI)) of up to
24.1% (ref. 9). Adsorption and complexation of Cr(III), and to a lesser
extent Cr(VI), on or with organic and inorganic particles is not accompanied
by Cr isotopic shifts17. Cr(III) input into sea water by hydrothermal vents is
considered small and the Cr isotopic composition of this fraction possibly
reflects the d53Cr values of ,0.15% typical of magmatic high-temperature
reservoirs18. Back-transformation of Cr(III) to Cr(VI) from sediments to sea
water is again only possible when catalysed by MnO2.
5
6
Up to
+4.9‰
0.9
2
3
4
1
0.7
δ53Cr (‰)
0.5
0.3
0.1
–0.1
High-temperature-associated Cr
–0.3
Ferruginous ocean
Sulphidic ocean
–0.5
0
0.5
1
1.5
2
2.5
3
3.5
4
Age (Gyr)
Figure 2 | Graph showing the key aspects of the
Precambrian history of hexavalent chromium in
sea water. d53Cr values (grey filled diamond
symbols) for BIF versus age (22 localities in total;
high values up to 14.9% from Neoproterozoic
Fe-rich cherts plot outside the graph; data in
Supplementary Table 1). Six stages (separated by
dashed vertical lines) are identified and
compared to the ocean deep water
chemistry27.The light-grey shaded fields depict
the first and second GOE, respectively, as defined
by other redox-sensitive tracers. Open diamonds
designate data from the upper Gunflint Iron
Formation (Ontario, Canada) which is
transitional into the overlying Rove Formation
(detail in Fig. 3). The horizontal rectangular field
outlines the d53Cr values of magmatic Cr(III)rich ores and minerals formed under high
temperatures18. Data are reported using the delta
notation relative to the certified National Bureau
of Standards Cr reference standard SRM 979 (see
Methods). Error bars associated with the
individual symbols for d53Cr values are ,0.1%
(2s level; full analytical data are available in
Supplementary Table 1).
251
©2009 Macmillan Publishers Limited. All rights reserved
LETTERS
NATURE | Vol 461 | 10 September 2009
200
1,840 Myr ago
Intense
silification
R-5
R-2
175
Silification
S-25
S-24
175
S-23
150
150
High-temperature Cr
Gunflint
Rove
Formation Formation
than about 0.1% of the present levels25. Furthermore, we speculate that
reduced oxygen levels may have inhibited the weathering flux of
sulphate to the oceans, encouraging the return of marine ferruginous
conditions and BIF deposition. Does this mean that oxygen concentrations were reduced to levels comparable to the early Archaean? We
see no evidence to suggest this. Mass independent sulphur isotope
fractionations persist through the Archaean26, but are not observed
in stage 4 sediments25. We suggest that chromium systematics are very
sensitive to oxygen, but not linearly so, such that factors controlling
MnO2 availability (as it relates to Cr(III) oxidation) might also be
important. More work on these systematics will settle this issue.
Higher 53Cr isotope values are observed during the very final stages
of Gunflint deposition (open diamonds), implying a subsequent
increase in oxygen concentrations (Fig. 3; Supplementary Table 1).
This represents the time immediately before the likely development
of widespread sulphidic oceanic conditions, which are thought to have
Stromatolites
Trough
cross-stratification
Ripples
125
S-21
Flaser bedding
125
Wavy bedding
Parallel lamination
in chemical
sediments
S-20
100
100
Parallel lamination
in siliciclastics
Tuffaceous layers
S-16
75
75
S-13
Archaean gneiss
1,878 ± 1 Myr ago
Siliciclastics content
S-10
50
All siliciclastics
50
S-8
25
S-4
S-2
Abundant
siliciclastics
Some
siliciclastics
25
Rare
siliciclastics
0
–0.3 –0.2 –0.1
M
ud
Si
l
F
M in t
ed e
C ium
oa
rs
e
0
R-12 Sample
location
0.0
0.1
0.2
δ53Cr (‰)
Sand
Figure 3 | Stratigraphy of the Gunflint Formation and its transition into the
Rove Formation29 and sample horizons. Data show an increase in d53Cr
values in the uppermost Gunflint Iron Formation relative to values typical of
minerals and ores associated with magmatic (high-temperature) processes18
(grey filled rectangle in the data log). Error bars correspond to the 2sm of
repeat analyses of respective samples (full analytical data are available in
Supplementary Table 1).
persisted throughout much of the Mesoproterozoic27–29. The traditional explanation for the development of sulphidic conditions calls
for an increase in the flux of sulphate to the oceans (and hence
increased rates of sulphide production by bacterial sulphate reduction)
due to enhanced oxidative weathering of continental sulphide minerals
as a result of the GOE27. It has never been clear, however, why it would
take several hundred million years for the sulphide flux to overwhelm
the hydrothermal Fe(II) flux, thus allowing sulphidic conditions
eventually to develop at ,1.84 Gyr (ref. 29). Our Cr isotope data
implies that atmospheric O2 concentrations fluctuated over this
period, and the onset of sulphidic conditions at ,1.84 Gyr is a consequence of an increased sulphate flux arising owing to a previously
unrecognized rise in atmospheric O2 during the final stages of
Palaeoproterozoic BIF deposition.
Stage 5, between ,1.8 and 0.75 Gyr, comprises the Mesoproterozoic
period in which sulphidic oceans predominated27 and during which
BIF deposition was largely prevented because of the preferential titration of Fe21 by HS2.
Stage 6 comprises the late Neoproterozoic era between ,750 Myr
and the Precambrian–Cambrian boundary at 542 Myr. BIFs deposited
during this stage include the 755–730-Myr-old Rapitan BIF, deposited
in a glaciomarine setting during the early Cryogenian (‘Sturtian’)
glaciation, and BIF- and Fe-bearing cherts of the ,570–550-Myrold Yerbal and Cerro Espuelitas formations (Arroyo del Soldado
Group, Uruguay) which were deposited after the Gaskiers glaciation30.
All of these BIFs record strongly positively fractionated Cr isotopes,
with d53Cr values ranging from 0.9% to 4.9%. These high values
provide independent support for Late Neoproterozoic oxygenation,
which further points to a causal link between this oxygenation and the
emergence of the Ediacara biota3 and bilateral motile animals.
Our Cr isotope record provides new and complementary insights into
the history of Precambrian biospheric oxygenation. The data highlights
fine-scale fluctuations in the oxygenation of the ocean and atmosphere
through time, and we foresee that combining Cr isotope systematics
with information from other redox-sensitive elements, such as C, S, Fe
and Mo, will greatly enhance our understanding of the complex history
of chemical and biological evolution on the early Earth.
METHODS SUMMARY
Individual mesobands of BIF samples were isolated from one-centimetre-thick
slices of hand specimens or drill core pieces and subsequently milled in an agate
mortar. Rock powder aliquots (amounts adjusted to yield 2–5 mg Cr in the final
separate) were spiked with an adequate amount of a 50Cr–54Cr double spike and
digested in HF:HNO3 mixtures in closed teflon vials on a hot plate at 150 uC. The
samples were then taken up in 6 M hydrochloric acid and passed through an
exchange column charged with 6 ml Dowex AG 1 3 12 anion resin to remove Fe.
Oxidation of Cr(III) to Cr(VI) in dilute hydrochloric acid was then achieved by
addition of (NH4)S2O8 as an oxidizing agent on a hot plate at 130 uC. In a second
chromatographic separation, the dilute Cr(VI) solutions were processed over
chromatographic columns charged with Dowex AG 1 3 8 anion resin. Release of
Cr from the anion resin was achieved by reduction to Cr(III) with the help of 2 M
nitric acid and hydrogen peroxide. All Cr isotope measurements were performed
on a IsotopX/GV IsoProbe T thermal ionization mass spectrometer equipped
with eight Faraday collectors that allow simultaneous collection of all four chromium beams (50Cr1, 52Cr1, 53Cr1, 54Cr1) together with 49Ti1, 51V1 and 56Fe1
as monitors for small interferences of these masses on 50Cr and 54Cr. Cr separates
were measured from Re-filaments at 1,000–1,100 uC and loaded with ultraclean
water into a mixture of 3 ml silica gel, 0.5 ml 0.5 M H3BO3 and 0.5 ml of 0.5 M
H3PO4. Every separate was analysed up to six times with minimum 52Cr beam
intensities of 400 mV. Data are reported relative to the certified Cr isotope
standard NIST SRM 979 (see online-only Methods). Within-run precisions of
the sample d53Cr values were 60.08% (2s) or better.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 12 February; accepted 30 June 2009.
1.
2.
Bekker, A. et al. Dating the rise of atmospheric oxygen. Nature 427, 117–120 (2004).
Rouxel, O. J., Bekker, A. & Edwards, K. J. Iron isotope constraints on the Archean
and Paleoproterozoic ocean redox state. Science 307, 1088–1091 (2005).
252
©2009 Macmillan Publishers Limited. All rights reserved
LETTERS
NATURE | Vol 461 | 10 September 2009
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Canfield, D. E., Poulton, S. W. & Narbonne, G. M. Late-Neoproterozoic deepocean oxygenation and the rise of animal life. Science 315, 92–95 (2007).
Fike, D. A. et al. Oxidation of the Ediacaran Ocean. Nature 444, 744–747 (2006).
Bekker, A. & Kaufman, A. Oxidative forcing of global climate change: a
biogeochemical record across the oldest Paleoproterozoic ice age in North
America. Earth Planet. Sci. Lett. 258, 486–499 (2007).
Canfield, D. E. & Teske, A. Late Proterozoic rise in atmospheric oxygen
concentration inferred from phylogenetic and sulphur-isotope studies. Nature
382, 127–132 (1996).
Oze, C., Bird, D. K. & Fendorf, S. Genesis of hexavalent chromium from natural
sources in soil and groundwater. Proc. Natl Acad. Sci. USA 104, 6544–6549
(2007).
Campbell, J. A. & Yeats, P. A. Dissolved chromium in the northwest Atlantic
Ocean. Earth Planet. Sci. Lett. 53, 427–433 (1981).
Sikora, E. R., Johnson, T. M. & Bullen, T. D. Microbial mass-dependent
fractionation of chromium isotopes. Geochim. Cosmochim. Acta 72, 3631–3641
(2008).
Ellis, A. S., Johnson, T. M. & Bullen, T. D. Chromium isotopes and the fate of
hexavalent chromium in the environment. Science 295, 2060–2062 (2002).
Eary, L. E. & Rai, D. Kinetics of chromate reduction by ferrous-ions derived from
hematite and biotite at 25uC. Am. J. Sci. 289, 180–213 (1989).
Fendorf, S. E. Surface-reactions of chromium in soils and waters. Geoderma 67,
55–71 (1995).
Sass, B. M. & Rai, D. Solubility of amorphous chromium(III)-iron(III) hydroxide
solid-solutions. Inorg. Chem. 26, 2228–2232 (1987).
Eary, L. E. & Rai, D. Chromate removal from aqueous wastes by reduction with
ferrous ion. Environ. Sci. Technol. 22, 972–977 (1988).
Schauble, E., Rossman, G. R. & Taylor, H. P. Theoretical estimates of equilibrium
chromium-isotope fractionations. Chem. Geol. 205, 99–114 (2004).
Izbicki, J. A. et al. Chromium, chromium isotopes and selected elements, western
Mojave Desert, USA. Appl. Geochem. 23, 1325–1352 (2008).
Ellis, A. S., Johnson, T. M. & Bullen, T. D. Using chromium stable isotope ratios to
quantify Cr(VI) reduction: lack of sorption effects. Environ. Sci. Technol. 38,
3604–3607 (2004).
Schoenberg, R. et al. The stable Cr isotope inventory of solid Earth reservoirs
determined by double spike MC-ICP-MS. Chem. Geol. 249, 294–306 (2008).
Anbar, A. D. et al. A whiff of oxygen before the Great Oxidation Event? Science 317,
1903–1906 (2007).
Kaufman, A. et al. Late Archean biosheric oxygenation and atmospheric evolution.
Science 317, 1900–1903 (2007).
Wille, M. et al. Evidence for a gradual rise of oxygen between 2.6 and 2.5 Ga from
Mo isotopes and Re-PGE signatures in shales. Geochim. Cosmochim. Acta 71,
2417–2435 (2007).
22. Kamber, B. S. & Whitehouse, M. J. Micro-scale sulphur isotope evidence for
sulphur cycling in the late Archean shallow ocean. Geobiology 5, 5–17 (2007).
23. Isley, A. E. & Abbott, D. H. Plume-related mafic volcanism and the deposition of
banded iron formation. J. Geophys. Res. 104, 15461–15477 (1999).
24. Canfield, D. E. The early history of atmospheric oxygen: homage to Robert A.
Garrels. Annu. Rev. Earth Planet. Sci. 33, 1–36 (2005).
25. Johnston, D. T. et al. Evolution of the oceanic sulfur cycle at the end of the
Paleoproterozoic. Geochim. Cosmochim. Acta 70, 5723–5739 (2006).
26. Farquhar, J. & Wing, B. A. Multiple sulfur isotopes and the evolution of the
atmosphere. Earth Planet. Sci. Lett. 213, 1–13 (2003).
27. Canfield, D. E. A new model for Proterozoic ocean chemistry. Nature 396,
450–453 (1998).
28. Scott, C. et al. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature
452, 456–459 (2008).
29. Poulton, S. W., Fralick, P. W. & Canfield, D. E. The transition to a sulphidic ocean
similar to 1.84 billion years ago. Nature 431, 173–177 (2004).
30. Gaucher, C. et al. Acritarchs of Las Ventanas Formation (Ediacaran, Uruguay):
implications for the timing of coeval rifting and glacial events in western
Gondwana. Gondwana Res. 13, 488–501 (2008).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements This study was funded by the Danish Agency for Science,
Technology and Innovation and by the Danish National Research Foundation
(Danmarks Grundforskningsfond).We thank R. Schoenberg for sharing his Cr
double spike with us and for contributions during many thematic and analytical
discussions. To add to our own collection, critical samples were provided by
C. Klein, A. Polat, P. S. Dahl, S. K. Mondal, H.-J. Hansen and E. F. Duke. Help with the
clean-laboratory chemical separation procedures by T. Larsen is acknowledged.
T. Leeper helped with the mass spectrometry and kept the mass spectrometer in
excellent condition.
Author Contributions R.F. and C.G. collected critical Neoproterozoic samples
during fieldwork in Uruguay in 2006. D.E.C. and S.W.P. provided the important
sediment samples from the Gunflint and Rove formations. Methods development
and thermal ionization mass spectrometer analytical work was conducted by R.F.
The manuscript was produced by significant contributions by R.F., D.E.C., S.W.P.
and C.G. Furthermore, D.E.C. and S.W.P. provided deeper insights into Proterozoic
atmospheric and deep ocean oxygenation models and stimulated discussion of the
early Earth system evolution.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to R.F. ([email protected]).
253
©2009 Macmillan Publishers Limited. All rights reserved
doi:10.1038/nature08266
METHODS
Individual mesobands of BIF samples were isolated from one-centimetre-thick
slices of hand specimens or drill core pieces and subsequently milled in an agate
mortar.
For trace elemental analyses, the rock powders were attacked in HBr, dissolved
with HF and HNO3 during addition of HBO3 (ref. 31), and then dried and redissolved in HNO3. Trace-element concentrations were determined by solution
ICP-MS (inductively coupled plasma mass spectrometry) with a Perkin Elmer
ELAN 6100 DRC spectrometer at the Geological Survey of Denmark and
Greenland (GEUS), using international standards for calibration. For a comparison
of GEUS analytical results on some standards with published values, refer to table 1
in ref. 32.
Rock powder aliquots (amounts adjusted to yield 2–5 mg Cr in the final separate)
were spiked with an adequate amount of a 50Cr–54Cr double spike and digested in
HF:HNO3 mixtures in closed PFA vials on a hot plate at 150 uC. After drying down,
the residues were taken up in aqua regia and reheated to 170 uC for a couple of
hours to destroy fluoride complexes that may have formed during the digestion.
After renewed drying down, the sample was then taken up in 6 M hydrochloric acid
for the Cr extraction.
We used an anion exchange chromatography technique adapted from previously published methods33,34 with few modifications to separate the Cr of natural
samples from the other matrix elements. First, because of the high iron contents
of our BIF, we passed the solutions through a cation exchange column charged
with 6 ml Dowex AG 1 3 12 in 6 M HCl to remove Fe. Sometimes we passed the
samples twice over this column to ensure that Fe was removed quantitatively. In
a second chromatographic separation over 1 ml stem columns charged with
Dowex AG 1 3 8 anion resin, we cleaned the Cr fractions collected from the
Fe-cleanup columns in dilute 0.2 M HCl from other rock matrices. This separation method is based on exchange of chloride ions on the Dowex AG 1 3 8 resin
by the Cr(VI) oxyanions18. After sample digestion and the first cationic exchange
Cr is present in its trivalent (CrIII) form, so oxidation of Cr(III) to Cr(VI) was
achieved using (NH4)S2O8 as oxidizing agent33 on a hot plate at 130 uC. Release
of Cr from the anion resin was achieved by reduction to Cr(III) using 2 M HNO3
and H2O2. The procedure yields for Cr in the above described separation method
varied between 80–90%, and Cr procedure blanks were in the order of 5–10 ng,
which is negligible compared to the amount of Cr separated from the samples
studied herein.
The addition of a 50Cr–54Cr double spike of known isotope composition to a
sample before chemical purification allowed accurate correction of both the
chemical and the instrumental shifts in Cr isotope abundances10,18. With this
method we achieve a 2s external reproducibility of the d53Cr value with 1.5 mg Cr
loads of the NIST SRM 3112a standard on our IsotopX/GV IsoProbe T thermal
ionization mass spectrometer of 60.05% with 52Cr signal intensities of 1 V and
of 60.08% for 52Cr beam intensities of 500 mV. The double-spike correction
returns Cr isotope compositions of samples as the % difference to the isotope
composition of the NIST SRM 3112a Cr standard (which was used for the spike
calibration18, so to maintain inter-laboratory comparability of Cr isotope data,
we recalculate our data of natural samples relative to the certified Cr isotope
standard NIST SRM 979 as follows:
d53Crsample (SRM 979) 5 [(53Cr/52Cr)sample/
(53Cr/52Cr)SRM 979 2 1] 3 1,000.
All Cr isotope measurements were performed on an IsotopX/GV IsoProbe T
thermal ionization mass spectrometer equipped with eight Faraday collectors that
allow simultaneous collection of all four chromium beams (50Cr1, 52Cr1, 53Cr1,
54 1
Cr ) together with 49Ti1, 51V1 and 56Fe1 as monitors for small interferences of
these masses on 50Cr and 54Cr. Cr separates were measured from Re-filaments at
1,000–1,100 uC and loaded with ultraclean water into a mixture of 3 ml silica gel,
0.5 ml 0.5 M H3BO3 and 0.5 ml of 0.5 M H3PO4. Every separate was analysed 1–6
times with minimum 52Cr beam intensities of 400 mV, allowing within-run precision of the d53Cr value of 60.09% or better. To achieve this, we ran the sample
over 120 cycles (grouped into 24 blocks of five cycles each) in static mode, and
integrated over 10 s with 20 s background (baseline) collection at 0.5 AMU on either
side of the peaks. This led to an average analysis time of ,1.5 hours. The final d53Cr
value of a sample was then calculated as the average of the repeated analyses. We
spiked our samples with an aliquot of the double spike used by Schoenberg et al.18
in their study of silicates and oxides of magmatic and metamorphic rocks, and
employed the double-spike correction developed by their group. The external
reproducibility of the NIST SRM 3112a standard over a period of one year, using
the 50Cr/54Cr ratio of Shields35 for mass bias correction, is shown in Supplementary
Fig. 1. Our average 53Cr/52Cr and 50Cr/52Cr ratios for this standard are
0.113452 6 50 p.p.m. (n 5 200, 2s) and 0.0282095 6 151 p.p.m. (n 5 200; 2s),
respectively. These values are indistinguishable from those of the SRM 979 isotopic
standard, which we reproduce at a 53Cr/52Cr ratio of 0.1134502 6 78 p.p.m.
(n 5 100; 2s) and a 50Cr/52Cr ratio of 0.0282089 6 161 p.p.m. (n 5 100; 2s).
This coincidence means we did not have to correct our d53CrNIST SRM 3112a values
to maintain inter-laboratory comparability. The average d53CrNIST SRM 3112a is
20.019 6 0.050% (n 5 32, 2s). The small deviation from the nominal value of
0% is most probably due to a small inaccuracy in the calibration of the Cr isotope
composition of the double spike18.
31. Connelly, J. N. et al. A method for purifying Lu and Hf for analyses by MC-ICP-MS
using TODGA resin. Chem. Geol. 233, 126–136 (2006).
32. Kalsbeek, F. & Frei, R. The Mesoproterozoic Midsommerso dolerites and
associated high-silica intrusions, North Greenland: crustal melting, contamination
and hydrothermal alteration. Contrib. Mineral. Petrol. 152, 89–110 (2006).
33. Ball, J. W. & Bassett, R. L. Ion exchange separation of chromium from natural
water matrix for stable isotope mass spectrometric analysis. Chem. Geol. 168, 1–2
(2000).
34. Frei, R. & Rosing, M. T. Search for traces of the late heavy bombardment on
Earth—results from high precision chromium isotopes. Earth Planet. Sci. Lett. 236,
28–40 (2005).
35. Shields, W. R. et al. Absolute isotopic abundance ratios and atomic weight of a
reference sample of chromium. J. Res. Natl Bureau Standards A70, 193–197 (1966).
©2009 Macmillan Publishers Limited. All rights reserved