See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8936178

Osmium Isotope Heterogeneity in the Constituent Phases of Mid-Ocean Ridge

Basalts

Article  in  Science · February 2004

DOI: 10.1126/science.1090266 · Source: PubMed

CITATIONS READS
63 115

6 authors, including:

Abdelmouhcine Gannoun Kevin W. Burton

Université Clermont Auvergne Durham University
112 PUBLICATIONS   2,252 CITATIONS    237 PUBLICATIONS   7,767 CITATIONS   

SEE PROFILE SEE PROFILE

Louise Thomas Ian Parkinson

The Open University (UK) University of Bristol
51 PUBLICATIONS   1,693 CITATIONS    112 PUBLICATIONS   5,296 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Melt inclusions entrapment mechanisms and their effects on MI composition View project

The Cathole Rock Art Project View project

All content following this page was uploaded by Abdelmouhcine Gannoun on 02 June 2014.

The user has requested enhancement of the downloaded file.

 Osmium Isotope Heterogeneity in the Constituent
Phases of Mid-Ocean Ridge Basalts
Abdelmouhcine Gannoun, et al.
Science 303, 70 (2004);
DOI: 10.1126/science.1090266

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of March 10, 2008 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/cgi/content/full/303/5654/70

Supporting Online Material can be found at:

Downloaded from www.sciencemag.org on March 10, 2008

http://www.sciencemag.org/cgi/content/full/303/5654/70/DC1

This article has been cited by 13 article(s) on the ISI Web of Science.

This article has been cited by 1 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/303/5654/70#otherarticles

This article appears in the following subject collections:

Geochemistry, Geophysics
http://www.sciencemag.org/cgi/collection/geochem_phys

Information about obtaining reprints of this article or about obtaining permission to reproduce
this article in whole or in part can be found at:
http://www.sciencemag.org/about/permissions.dtl

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2004 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 REPORTS
reported earlier for CsHSO4, in which the
membrane thickness was ⬃1.4 mm (2).
A second type of MEA was constructed
with PtRu [50:50 mole percent (mol %)] as
the anode electrocatalyst to demonstrate the
viability of a CsH2PO4 electrolyte in vapor-
fed DMFCs (4). Such fuel cells are anticipat-
ed to provide the advantages of complete
impermeability of the electrolyte to methanol
and elevated operating temperatures. These
features, in turn, translate into higher power
densities, higher CO tolerance, and, ultimate-
ly, lower precious metal catalyst loadings.
We introduced vapor-phase methanol and
water (0.92 :1 molar ratio) into the anode
compartment by using argon as a carrier gas,
while humidified oxygen was fed into the
cathode. The cell was operated at 243°C.
Electrical current (80 mA cm–2) was drawn
curves are again quite similar. Peak power
densities were 37.2 and 34.4 mW cm–2, re-
spectively, for the two measurements. Open-
circuit voltages, measured at 0.897 and 0.865
V before and after stability examination, re-
spectively, cannot be compared to a theoret-
ical (Nernst) potential because, as is typical
for DMFCs, the partial pressures of the prod-
uct gas, CO2, and possible intermediates,
such as CH4, were not controlled or moni-
tored. Nevertheless, the OCVs are greater
than those obtained from polymeric DMFCs
(⬃0.8 V) (9, 10), presumably because of the
higher methanol content used here and the
lower membrane permeability. The power
densities reached in this work are within
about a factor of 5 of the most advanced
DMFCs (100 to 200 mW cm–2) (9, 10).
Despite the apparently poor thermal sta-
demanding humidification requirement
lead to system simplifications.
References and Notes
1. A. I. Baranov, L. A. Shuvalov, N. M. Shchagina, JETP
Lett. 36, 459 (1982).
2. S. M. Haile, D. A. Boysen, C. R. I. Chisholm, R. B. Merle,
Nature 410, 910 (2001).
3. R. B. Merle, C. R. I. Chisholm, D. A. Boysen, S. M. Haile,
Energy Fuels 17, 210 (2003).
4. Materials and methods are available as supporting
material on Science Online.
5. B. M. Nirsha et al., Russ. J. Inorg. Chem. 27, 770 (1980).
6. D. A. Boysen, S. M. Haile, H. Liu, R. A. Secco, Chem.
Mater. 15, 727 (2003).
7. J. Otomo, N. Minagawa, C. Wen, K. Eguchi, H. Taka-
hashi, Solid State Ionics 156, 357 (2003).
8. E. Rapoport, J. P Clark, P. W. Richter, J. Solid State
Chem. 24, 423 (1978).
9. A. K. Shukla, C. L. Jackson, K. Scott, R. K. Raman,
Electrochem. Acta 47, 3401 (2002).
10. S. C. Thomas, X. M. Ren, S. Gottesfeld, P. Zelenay,
Electrochem. Acta 47, 741 (2002).
11. Supported by the NSF, Division of Materials Research,

Downloaded from www.sciencemag.org on March 10, 2008

continuously for 35 hours (without an initial
equilibration step) and polarization curves
were obtained both before and after the mea-
surement (Fig. 4). As in the case of the H2 /O2
cell, the methanol cell showed remarkable
stability, with the voltage decreasing slightly
from an initial value of 0.441 V to a final
value of 0.381 V; the majority of that de-
crease occurred in the first few hours and is
attributed to slight evolution of the electro-
catalyst layer. Moreover, the two polarization
bility of CsH2PO4 in its superprotonic phase,
both H2/O2 and DMFCs based on this mate-
rial show excellent long-term performance
when stabilized with water partial pressures
of ⬃0.30 atm, depending on the temperature
of operation. These “humidity-stabilized”
solid acid fuel cells exhibit higher OCVs
than polymer electrolyte fuel cells and thus
may ultimately yield better overall system
efficiencies. More important, the high-
temperature operation and the much less
and the Office of Naval Research, Office of Electro-
chemical Power, Science, and Technology.
Supporting Online Material
www.sciencemag.org/cgi/content/full/1090920/DC1
Materials and Methods
SOM Text
Figs. S1 to S4
Table S1
References and Notes
28 August 2003; accepted 12 November 2003
Published online 20 November 2003;
10.1126/science.1090920
Include this information when citing this paper.

Osmium Isotope Heterogeneity ments in MORB and the upper mantle should
be the same. It is also assumed that the constit-

in the Constituent Phases of uent phases of the basalts preserve isotope com-
positions that relate directly to the mantle
source. For lithophile elements, such as Sr or
Mid-Ocean Ridge Basalts Nd, parent/daughter ratios in MORB phases are
relatively low such that shifts in the radiogenic
Abdelmouhcine Gannoun,1* Kevin W. Burton,1 isotope composition cannot be produced in less
Louise E. Thomas,1 Ian J. Parkinson,1 Peter van Calsteren,1 than 103 million years (My) (8). Thus, any
Pierre Schiano2 variations in Sr and Nd isotope composition are
commonly attributed to a compositional heter-
Radiogenic isotope variations in mid-ocean ridge basalts (MORB) are commonly ogeneity in the upper mantle (8). However, for
attributed to compositional variations in Earth’s upper mantle. For the rhenium- the Re–Os isotope system silicate phases [such
osmium isotope system, constituent MORB phases are shown to possess ex- as olivine (9) and glass (3)] possess high 187Re/
187
ceptionally high Re/Os (parent/daughter) ratios, consequently radiogenic 187Os Os (parent/daughter) ratios. This raises the
is produced from the decay of 187Re over short periods of time. Thus, in the possibility that radiogenic 187Os can be pro-
absence of precise age constraints, Os isotope variations cannot be unambig- duced in situ from the decay of 187Re over short
uously attributed to their source, although Re–Os isotope data for constituent periods of time [i.e., 100 thousand years (ky) or
minerals can yield crystallization ages, details of equilibration, and initial Os less] (5). Consequently, the Os isotope compo-
isotope values that relate directly to the mantle source. sition of constituent phases in MORB may be
variable depending on their Re/Os ratio and the
The decay of 187Re to 187Os provides an ex- over time, which in turn can be readily traced as time that has elapsed since crystallization. If the
ceptional tracer of recycled crustal materials in recycled material if mixed back into the con- magmatic phases are in equilibrium, then they
Earth’s mantle. This is because oceanic and vective mantle (1–5). MORB are believed to may yield an isochron that will give the age of
continental crust possess high Re/Os ratios, and form by partial melting of Earth’s upper mantle, crystallization, and the initial Os isotope com-
develop radiogenic Os isotope compositions and their concentration ratios of incompatible position defined by that best-fit line corre-
elements or radiogenic isotope compositions sponds to that of the mantle source. However, if
are considered to relate directly to the mantle some of the phases were assimilated from previ-
1
Department of Earth Sciences, The Open University,
Walton Hall, Milton Keynes, MK7 6AA, UK. 2Labora-
source. A fundamental assumption underlying ously solidified basalts or contaminated by seawa-
toire Magmas et Volcans, Université Blaise-Pascal, 5 the use of radiogenic isotopes in such mantle- ter, then they may possess different isotope infor-
Rue Kessler, 63038 Clermont-Ferrand, France. derived basalts is that they are in equilibrium mation to that of their basalt host or other phases.
*To whom correspondence should be addressed. E- with their mantle source (6, 7). Thus, the com- Thus, understanding the nature of the Re–Os iso-
mail: a.gannoun@open.ac.uk position of long-lived isotopes of heavy ele- tope information preserved in MORB requires

70 2 JANUARY 2004 VOL 303 SCIENCE www.sciencemag.org


precise information on the time of igneous crys-
tallization as well as details of Re–Os equilibration
between all magmatic phases.
With the advent of ultra low-blank solvent
extraction chemistry (10), it is now possible to
measure Re and Os in all magmatic phases (9).
Here, we present high-precision Re–Os isotope
data for glass, hyalocrystalline matrix, olivine,
plagioclase, clinopyroxene, spinel, and sulphide
phases separated from two MORB samples
from the FAMOUS region (36°50’N) on the
mid-Atlantic ridge. These results are used to
assess the age information preserved by MORB
glass and coexisting phases, some of the pro-
cesses involved in MORB genesis, and the
187
Os/188Os composition of the MORB source.
Sample ARP1974-011-018 (36.85°N;
33.25°W) is an olivine basalt containing olivine
(Fo90–Fo80), plagioclase (An91–An95), and cli-
corresponding to an age of 565 ⫾ 336 ky and
an initial 187Os/188Os ratio of 0.1265 ⫾ 0.0046
(Fig. 1). The data for clinopyroxene are distinct
from this best-fit line, suggesting either an older
age or a different and more radiogenic source
for this phase. Sample ARP1973-010-003
(36.8372°N; 33.2482°W; 2760-m water depth)
is a porphyritic, picritic basalt with abundant
olivine phenocrysts (Fo91–Fo89; up to 5 mm in
diameter) set in a glassy to hyalocrystalline
matrix. Cr-spinel [Cr/(Cr ⫹ Al) ⫽ 48.01] phe-
nocrysts and sulphide [⬃14 weight percent (wt
%) Ni] blebs (up to 1 mm in diameter) occur as
inclusions in olivine or discrete crystals in the
groundmass. Plagioclase (An80) microlites are
also common (11, 12). Re–Os isotope data for
olivine, plagioclase, glass, and sulphide (Table
1) yield a best-fit line corresponding to an age
of 2.53 ⫾ 0.15 My and an initial 187Os/188Os
REPORTS
that the best-fit lines represent the time of ig-
neous crystallization and the initial Os isotope
composition represents that of the mantle
source. The crystallization ages of 565 ⫾ 336
ky and 2.53 ⫾ 0.15 My, respectively, are much
older than might be expected from age-distance
relations with the ridge axis that suggest ages of
5 to 10 ky (13). These ages are also at odds with
new 230Th-226Ra data for these samples (table
S1). ARP1974-011-018 glass gives a 226Ra/
230
Th activity ratio close to 1, suggesting that
the sample is ⱖ 8 ky old, whereas the 230Th/
232
Th activity ratio is 1.273, suggesting that the
sample is ⱕ 350 ky old, consistent with previ-
ous 230Th bulk rock data (14). However, regres-
sion of the Re-Os data for ARP1974-011-018
does not yield a statistically valid isochron, and
within the given age uncertainty this sample
could be as young as 350 ky. Glass from

Downloaded from www.sciencemag.org on March 10, 2008

nopyroxene (Wo44En15Fs5–Wo40En15Fs9)
phenocrysts (up to 1 to 2 mm in diameter) and
microphenocrysts in a hyalocrystalline matrix,
and, in places, a glassy pillow rim (11, 12).
Re–Os isotope data for matrix, glass, plagio-
clase, and olivine (Table 1) yield a best-fit line
Fig. 1. 187Re–188Os isotope evolution
diagram for coexisting phases from the
olivine– basalt ARP1974-011-018. Oli-
vine, plagioclase, glass, and matrix yield
a best-fit line corresponding to an age
of 565 ⫾ 336 ky (2␴). Clinopyroxene
(not shown) does not lie on this best-fit
line, suggesting either an older age or a
different and more radiogenic source
for this phase. MWSD, mean square
weighted deviation.
ratio of 0.129 ⫾ 0.002 (Fig. 2). Spinel, which is
relatively Os rich (Table 1), possesses a distinct
isotope composition from this best-fit line (Fig.
2) and is probably the phase responsible for the
displacement of the matrix from the same line.
The most straightforward interpretation is
ARP1973-010-003 gives 226Ra/230Th ratio of
1.3, which might at first be taken to indicate that
the sample is less than 8 ky old. However, the
same sample has a 234U/238U ratio of 1.043,
and such elevated values are often taken to
indicate seawater contamination, consistent
with previously published data for this sample
(14), which raises the possibility that Ra has
also been affected by the same seawater con-
tamination. The 230Th/232Th activity ratio is
1.282, which indicates an age of ⱕ 350 ky, but
only if seawater alteration has had no effect on Th.
Alternatively, the data may indicate that few
if any of the constituent phases crystallized in
their present basalt host (i.e., they are xeno-
crysts not phenocrysts). There is evidence for
assimilation of xenocrystic phases in samples
from the FAMOUS region, including those
studied here (11, 15–18). For example, in
ARP1973-010-003 high-Al spinel is considered

Table 1. Re and Os isotope data for FAMOUS samples. Sample weight in mg, except
sulphides are given in ␮g. Concentrations in parts per trillion (10⫺12 g/g) by weight,
except sulphides and spinel are given in parts per million. Errors are 2␴m. Two
suites of procedural blanks were made during this study. Blank A
(low level): Os ⫽ 0.0149 ⫾ 13 pg, 187Os/188Os ⫽ 0.234 ⫾ 17 (n ⫽ 7 blanks),
Re ⫽ 0.83 ⫾ 12 pg (n ⫽ 3); blank B (normal level): Os ⫽ 0.0349 ⫾ 11 pg,
187
Os/188Os ⫽ 0.261 ⫾ 12 (n ⫽ 5), Re ⫽ 2.20 ⫾ 22 pg (n ⫽ 3). Full details of
procedural blanks, standards, and normalizing values are given in the SOM Text.

Sample Weight (mg) 187

Os/188Os 187
Re/188Os [Re] (pg g⫺1) [Os] (pg g⫺1)

ARP1974-011-018
Clinopyroxene* 24.41 0.2497 ⫾ 0.0057 4796 ⫾ 141 1155 1.167
Olivine* 14.02 0.1859 ⫾ 0.0055 5881 ⫾ 85 4112 3.361
Plagioclase* 66.21 0.1900 ⫾ 0.0043 7164 ⫾ 74 1220 0.819
Glass 1† 515.6 0.1400 ⫾ 0.0012 1137 ⫾ 3 1461 6.268
Glass 2† 503.5 0.1386 ⫾ 0.0014 1198 ⫾ 3 1730 7.048
Matrix 1† 506.8 0.1261 ⫾ 0.0004 180 ⫾ 1 706 18.68
Matrix 2† 488.2 0.1266 ⫾ 0.0007 172 ⫾ 1 715 19.80
ARP1973-010-003
Sulphide 1* 103 0.1310 ⫾ 0.0005 21.6 ⫾ 0.3 0.713 0.158
Sulphide 2* 39.6 0.1305 ⫾ 0.0009 22.7 ⫾ 0.2 3.230 0.679
Spinel* 0.51 0.1874 ⫾ 0.0045 514 ⫾ 7 0.126 0.0012
Olivine 1* 54.46 0.1399 ⫾ 0.0023 288 ⫾ 6 687 11.39
Olivine 2* 25.21 0.2411 ⫾ 0.0074 2492 ⫾ 71 1153 2.242
Plagioclase* 18.75 0.1892 ⫾ 0.0028 1456 ⫾ 53 277.9 0.925
Glass 1* 8.56 0.1816 ⫾ 0.0041 1281 ⫾ 35 3518 13.16
Glass 2* 17.62 0.1846 ⫾ 0.0024 1310 ⫾ 20 3163 11.53
Matrix 1† 120.8 0.1359 ⫾ 0.0006 24.0 ⫾ 0.3 1489 295.7
Matrix 2† 331.2 0.1374 ⫾ 0.0005 24.3 ⫾ 0.1 1264 248.6
Matrix 3† 150.6 0.1359 ⫾ 0.0009 23.8 ⫾ 0.3 1151 234.6
*Corrected from blank A. †Corrected from blank B.

www.sciencemag.org SCIENCE VOL 303 2 JANUARY 2004 71

 REPORTS
to be a relict from high-pressure crystallization
(19), consistent with the Os isotope data pre-
sented here, which suggests that spinel is not in
Os isotopic equilibrium with the other phases.
However, if most of the phases lie on the same
best-fit line, then this interpretation demands
that all such minerals are xenocrysts. For the
picritic basalt (ARP1973-010-003), if eruption
occurred about 5 to 10 ky, then the Re–Os data
suggests that original crystallization of the min-
erals occurred about 2.5 My before that. In this
case, the xenocrysts were assimilated from pre-
viously solidified “olivine–plagioclase” basalts
or cumulates through which the present host
basalts have ascended. Furthermore, the Os
isotope composition of the glass, which
also lies on the same best-fit line, must
reflect contamination by dissolution of
those older phases.
It might be argued that the regression lines
are due to contamination by radiogenic Os from
seawater, rather than having some age signifi-
cance. This would require that the contamina-
tion occurred during mineral crystallization and
has affected phases such as olivine and plagio-
clase in a systematic manner; otherwise, it is
difficult to imagine how different phases would
align to yield the correlations observed. How-
ever, the observation that sulphide, of the same
composition, occurs not only as inclusions in
olivine and plagioclase but also as discrete
grains in the matrix, suggests that these phases
grew at approximately the same time but were
not equally affected by contamination. Even if
the phases are xenocrysts or have been affected
by seawater contamination, the initial Os iso-
tope composition defined by the best-fit lines is
still likely to represent that of the mantle source.
The MORB glasses preserve 187Os/188Os
values of ⬇0.14 (ARP1974-011-018) and
⬇0.18 (ARP1973-010-003), respectively,
which could be taken to indicate the presence
of source heterogeneity in the FAMOUS re-
gion. However, the initial 187Os/188Os values
defined by the best-fit regressions are differ-
ent from the measured glass values and in-
distinguishable from each other, indicating
that the two samples may have been derived
from a common mantle source. During man-
tle melting and basalt genesis, the composi-
Fig. 2. 187Re–188Os isotope evolu-
tion diagram for coexisting phases
from the picritic basalt ARP1973-
010-003. Olivine, plagioclase, glass,
and sulphide lie on a best-fit line
corresponding to an age of 2.53 ⫾
0.15 My (2␴). Spinel possesses a
distinct isotope composition from
this best-fit line and is probably the
phase responsible for the displace-
ment of the matrix from the
same line.
72 2 JANUARY 2004 VOL 303 SCIENCE www.sciencemag.org
tion of long-lived isotopes of heavy elements
in both the melt (MORB) and residual solid
(abyssal peridotites) should be the same (6,
7). However, many MORB glasses possess
Os isotope compositions that are more radio-
genic than those of abyssal peridotites (3,
20–22). The radiogenic Os values have been
ascribed to the presence of some other rock
type, such as pyroxenite, in the mantle source
(3), whereas others have attributed the data to
contamination by seawater (4) or in situ de-
cay of 187Re following igneous crystalliza-
tion (5). Our data for the mantle source com-
position are within the range obtained for
abyssal peridotites, and indistinguishable
from estimates for the modern “primitive up-
per mantle” (23) (a hypothetical undifferen-
tiated upper mantle reservoir). Although lim-
ited in global coverage, if such compositions
were to prove typical of the MORB source,
then the apparent contradiction with abyssal
peridotite values may, at least in part, be
reconciled. This would also suggest that the
processes of mantle differentiation through
partial melting, recycling of oceanic crust,
and homogenization through convective mix-
ing have had little effect on the Re/Os ratio
and, hence, Os isotope composition of
Earth’s upper mantle.
It is tempting to suggest that all of the
radiogenic Os isotope compositions observed
in MORB glasses (3) can be attributed to in
situ decay of 187Re after crystallization (5).
However, it is unlikely that suites of MORB
glasses from the Indian and Pacific-Atlantic
Oceans, selected solely on the basis of their
optical purity, should align to yield ages of
ca. 1.06 My and 678 ky, respectively. Fur-
thermore, there is no evidence of a correlation
between Os isotope composition with spread-
ing rate (5) [fig. S1; supporting online mate-
rial (SOM) Text]. Thus, though in situ decay
of 187Re may explain some of the variation
seen in those MORB glasses (3), their sys-
tematic Re–Os variations and radiogenic Os
isotope compositions cannot be explained
solely as a function of eruption age.
Overall, our results demonstrate that ex-
ceptional fractionation of Re/Os occurs be-
tween magmatic phases in MORB. Thus,
unlike other commonly utilized isotope
systems, such as Sr, Nd, or Pb, radiogenic
Os isotope compositions can be produced
on ⬃102 ky time scales. We have also
shown that the Re–Os isotope data for the
constituent phases from MORB can be used
to determine the precise time of crystalli-
zation of magmatic minerals, details of
equilibration between phases, and, in favor-
able circumstances, the actual Os isotope
composition of the mantle source. These
results demonstrate that, in the absence of
such mineral data or precise age con-
straints, the Os isotope data for MORB
glass or whole-rock samples cannot be un-
ambiguously interpreted.
References and Notes
1. C. J. Allègre, J.-M. Luck, Earth Planet. Sci. Lett. 48, 148
Downloaded from www.sciencemag.org on March 10, 2008
(1980).
2. B. J. Pegram, C. J. Allègre, Earth Planet. Sci. Lett. 111,
59 (1992).
3. P. Schiano, J.-L. Birck, C. J. Allègre, Earth Planet. Sci.
Lett. 150, 363 (1997).
4. S. B. Shirey, R. J. Walker, Annu. Rev. Earth Planet. Sci.
26, 423 (1998).
5. E. H. Hauri, Philos. Trans. R. Soc. London 360, 2371
(2002).
6. A. W. Hofmann, S. R. Hart, Earth Planet. Sci. Lett. 38,
44 (1978).
7. A. Zindler, S. R. Hart, Annu. Rev. Earth Planet. Sci. 14,
493 (1986).
8. A. W. Hofmann, Nature 385, 219 (1997).
9. K. W. Burton et al., Earth Planet. Sci. Lett. 198, 63
(2002).
10. J.-L. Birck, M. Roy Barman, F. Capmas, Geostandards
Newsl. 20, 9 (1997).
11. A. P. le Roex, A. J. Erlank, H. D. Needham, Contrib.
Mineral. Petrol. 77, 24 (1981).
12. Y. Su, C. H. Langmuir, PETDB: Petrological Database of
the Ocean Floor, updated 5 February 2002. Available
at http://petdb.ldeo.columbia.edu/petdb/.
13. M. Selo, D. Störzer, C. R. Acad. Sci. Paris Sér. II 289,
1125 (1979).
14. M. Condomines, P. Morand, C. J. Allègre, Earth Planet.
Sci. Lett. 55, 247 (1981).
15. R. Clocchiatti, C. R. Acad. Sci. Paris Sér. II 285, 1155
(1979).
16. N. Shimizu, Phys. Earth Planet. Inter. 107, 183 (1998).
17. V. Kamenetsky, Earth Planet. Sci. Lett. 142, 479
(1996).
18. P. I. Nableck, C. H. Langmuir, Contrib. Mineral. Petrol.
93, 1 (1986).
19. H. Sigurdsson, J.-G. Schilling, Earth Planet. Sci. Lett.
29, 7 (1976).
20. M. Roy-Barman, C. J. Allègre, Geochim. Cosmochim.
Acta 58, 5043 (1994).
21. J. Snow, L. Reisberg, Earth Planet. Sci. Lett. 133, 411
(1995).
22. A. D. Brandon, J. Snow, R. J. Walker, J. W. Morgan,
T. D. Mock, Earth Planet. Sci. Lett. 177, 319 (2000).
23. T. Meisel, R. J. Walker, J. W. Morgan, Nature 383, 517
(1996).
24. Supported by the Natural Environment Research
Concil (NER/A/S/2001/00538) and in part by the
European community’s Human Potential Programme
under contract HPRN-CT-2002-00211 [EUROMELT].
M. Gilmour and J. Dryden are thanked for their sup-
port, and S. Hart and G. Parson are thanked for their
helpful and constructive comments.
Supporting Online Material
www.sciencemag.org/cgi/content/full/303/5654/70/DC1
SOM Text
Tables S1 and S2
Figs. S1 and S2
References
8 August 2003; accepted 20 November 2003
Supporting Online Material

Th-U-Ra data
New Th-U-Ra isotope and elemental data for the glass from both of the samples used in
this study have been obtained. For the plagioclase-phyric basalt ARP1974-011-018 the
226
Ra/230Th activity ratio is close to 1, within the given uncertainty (i.e. at equilibrium),
230
suggesting that the sample is ≥ 8ka old. While, the Th/232Th activity ratio is 1.273
230
suggesting that the sample is ≤ 350 ka. This is consistent with previous Th bulk rock
data for this sample (1). For the picritic basalt ARP1973-010-003 the 226Ra/230Th activity
ratio is 1.3, which at first sight might be taken to indicate that the sample is less than 8 ka
234
old. However, the same sample has a U/238U ratio of 1.043. Such elevated values are
commonly taken to indicate seawater contamination (consistent with previously
published data for this sample (1)) which raises the possibility that Ra has also been
230
affected by the same seawater contamination. The Th/232Th activity ratio is 1.282,
which may indicate an age of ≤ 350 ka, but only if seawater alteration has had no effect
on Th (most likely). Taken together, the Th-U-Ra data indicates that the age of
ARP1974-011-018 lies between 8 and 350 kyr, and that of ARP1973-010-003 is similar
but only if Th has not been affected by seawater.

Table S1. Th-U-Ra data for glass samples

234
Sample U/238 U 238
U/ 232 Th 230
Th/ 232Th 226
Ra/ 230 Th [U] [Th]

ARP1974 1.005±0.002 0.903±0.002 1.273±0.003 1.055±0.010 0.177 0.593

ARP1973 1.043±0.002 1.271±0.005 1.282±0.010 1.300±0.010 0.093 0.223

All errors 2σ. U and Th concentrations are in ppm. Duplicate determinations of the
ATHO Th standard gave (230Th/232Th) = 1.015 ± 0.006 (n = 6) and TML gave
(230Th/232Th) = 1.0883 ± 0.01 (n = 2). Chemical preparation blanks were less than 10 pg
for Th and for U, which are negligible compared with the ≥ 100 ng of sample usually
loaded and the error of U/Th ratios is ≤ 0.5%. Average value of Mt Lassen Ra standard
1065 +/- 9 fg g-1 (n = 6) was in close agreement with the published values of 1063 +/- 10
and 1068 +/- 11 fg (2). Blanks for Ra are below limits of detection.
.

Spreading rate and MORB osmium isotopic composition

187
Re–187Os isotope data for MORB glasses from the Pacific-Atlantic Ocean give
apparent ages of ca. 678 kyr, while the data for Indian Ocean MORB give an apparent
age of 1.06 Myr (3). The older age for the Indian Ocean MORB correlates with a slower
mean-spreading rate, compared to the Atlantic and Pacific MORB. MORB are also
difficult to precisely date because zero-age basalts are erupted in a 500–1000 m wide
zone on the ridge axis (4). Thus, it has been proposed that the general relationship
between MORB age and spreading rate can explain the difference between the Indian and
Pacific-Atlantic Oceans, and also that this suggests that the radiogenic Os isotope values
187
for these samples are due in situ decay of Re following crystallisation (4). However,
spreading rates calculated for each individual sample locality from ref. (3) show no
correlation with Os isotope composition (see Figure S1 below) (Euler vectors used for the
calculation of the spreading rates are based on those given in refs. (5, 6)).

Figure S1: Plot of Os isotopic ratio of MORB glass versus spreading rate. Os data are
from (3), Spreading rate are calculated for each individual sample locality from (3) by
using (5,6).
Analytical method

Silicate minerals and glass were hand-picked and cleaned in dilute HCl, ethanol and
water. Following separation all mineral phases were powdered in agate and dissolved in
HBr + HF in a PFA pressure vial heated at 150 at least, for 48 h. Os and Re separation
were achieved using a solvent extraction technique (7). Re and Os isotopic measurements
(as ReO4- and OsO3-) were made using the ion counting electron multiplier on the
192
Finnigan TRITON. Os isotopic ratios are normalized to Os/188Os=3.08271 (8) and
18
corrected using measured O/16O and 17
O/16O ratios of 0.002047 and 0.00037,
respectively (9). Series of standards with various concentrations were measured. The
187
mean Os/188Os ratio for 14 pg JM standard yields 0.173978 ± 0.000048 (n=24),
0.17421±38 for 1 pg (n=5); 0.17381±91 for 500 fg (n=8) and 0.1742±10 for 100 fg (n=5)
(fig. S2). For comparison, the published values for the large quantity of the same standard
(10-100 ng) on faraday cups are 0.17394±14 (7), 0.17429±25 (10), 0.17410±48 (11),
0.17398±35 (12).

Fig. S2: Low level JM Os standard data obtained as part of this study.
Re-Os procedural blanks
Twelve Osmium blanks were run as part of this study, and their concentration and
isotopic composition (and their uncertainties) are given in Table S2. Two blank
populations were used in this study, the first, using normal reagent volumes, the second,
using reduced reagent volumes for the lower level mineral samples. For the minerals and
2 of the glasses 7 blanks were run that gave an Os concentration of 0.0148±0.0013 pg.
(that is, an uncertainty on the blank of ±8.7%). Whereas, For the higher level samples the
Os concentration of the blank is 0.0348±0.0010 pg. (that is, an uncertainty of ±4%). Such
variation between different reagent batches and/or volumes is typical for Os and
highlights the necessity to correct to the specific blank appropriate to any given samples,
rather than some long term mean integrating different reagents and/or reagent volumes.

Table S2: Re-Os isotope and elemental data for total procedural blanks.

187
[Os] pg Os/188Os

Normal blk#1 0.0347±12 0.262±12

blanks blk#2 0.0349±15 0.263±11
blk#3 0.0345±11 0.261±09
blk#4 0.0356±15 0.261±12
blk#5 0.0346±10 0.256±15

blk#1 0.0147±7 0.240±16

blk#2 0.0149±6 0.230±13
Low blk#3 0.0158±6 0.237±17
level blk#4 0.0151±5 0.231±30
blanks blk#5 0.0135±7 0.243±19
blk#6 0.0154±6 0.227±16
blk#7 0.0147±8 0.231±09

Average blk A low level 0.0149±4 0.234±15

Average blk B normal level 0.0349±10 0.261±11

Re blanks (pg)

Blk#1 Blk#2 Blk#3 Average

Low level 0.78 0.82 0.90 0.83±11
Normal level 2.21 2.30 2.08 2.20±22
 References

1. M. Condomines, P. Morand, C. J. Allègre, Earth Planet. Sci. Lett. 55, 247 (1981).
2. A.M. Volpe, J.A. Olivares, M.T. Murrell, Anal. Chem. 63 913(1991).
3. P. Schiano, J.-L. Birck, C. J. Allègre, Earth Planet. Sci. Lett. 150, 363 (1997).
4. E. H. Hauri, Phil. Trans. R. Soc. Lond. 360, 2371 (2002).
5. C. De Mets, R.G. Gordon, D.F. Argus, S. Stein, Geophys. J. Int. 101, 425 (1990).
6. C. De Mets, R.G. Gordon, D.F. Argus, S. Stein, Geophys. Res. Lett. 21, 2191
(1994).
7. J.-L. Birck, M. Roy Barman, F. Capmas, Geostandards Newsl. 20, 9 (1997).
8. J.M. Luck, K.K. Turekian, Science 222, 613 (1983).
9. A.O. Nier, Phys. Rev. 52, 885 (1937)
10. S.B. Shirey, Can. J. Earth Sci. 34, 489 (1997).
11. E. Hauri, S.R. Hart, Earth Planet. Sci. Lett. 114, 353 (1993).
12. B.F Schaefer, I.J. Parkinson, C. Hawkesworth, Earth Planet. Sci. Lett. 175, 105
(2000).

View publication stats