139149
www.elsevier.comrlocaterchemgeo
a,)
Department of Earth Sciences, Uniersity of Oxford, Parks Road, Oxford, OX1 3PR UK
b
Department of Earth Sciences, Uniersity of Wales, Cardiff, CF1 3YE UK
Received 21 December 1998; accepted 29 April 1999
Abstract
Techniques for the high precision measurement of 65Cur63Cu ratios by multiple-collector plasma-source mass spectrometry has been developed. Two approaches, namely Zn-doping and sample-standard bracketing, have been exploited. By
using the sample-standard bracketing technique, a range of samples including native copper, Cu-carbonate and
Cu-sulphides from terrestrial and marine environments have been analysed. An overall variation in 65Cur63Cu of 22 parts
per 10 4 22 units. is observed. This is more than 30 times the 2 s analytical uncertainty of the technique employed, and
thus demonstrates the great potential for using stable Cu isotopes as tracers in geological and planetary processes. The
variations in 65Cu values observed in this study display some regularity. Those samples involving formation through low
temperature aqueous solutions display large differences in 65Cu values even at a single locality, whereas chalcopyrite
samples hosted in igneous rocks show similar Cu-isotope compositions worldwide. This indicates that the 65Cu variations
arise principally through mass fractionation in low temperature aqueous processes, rather than through source heterogeneity.
In contrast to continental sulphides, chalcopyrites from black smoker sulphide chimneys on the ocean floor show large
variations in 65Cu values. Relative to active high temperature hydrothermal vents, the old inactive vent deposits are
enriched in 63Cu and show smaller variations in 65Cu values. Within in a single active chimney, Cu isotopes become lighter
from bottom to top. This variation pattern is explained tentatively by means of a two-stage-process model, which involves:
1. the preferential leaching of 65Cu by hydrothermal processes, and 2. subsequent isotopic exchange between the early
formed Cu-sulphides and 65Cu-depleted late-stage hydrothermal fluids. This new capability for Cu-isotope measurement is
expected to have a major impact across disciplines ranging from cosmochemistry and geochemistry through biogeochemistry
to biochemistry and alimentology. q 2000 Elsevier Science B.V. All rights reserved.
Keywords: Cu-isotopes; Sulphides; Plasma-source mass spectrometry; Hydrothermal systems; Black smokers; Isotope fractionation
1. Introduction
Natural variations in stable isotope abundances
have over the years provided profound insights into
)
0009-2541r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 5 4 1 9 9 . 0 0 0 7 6 - 5
140
mained unexplored in many elements. The introduction of multiple collector plasma source mass spectrometry Walder and Freedman, 1992; Walder et al.,
1993; Halliday et al., 1998. has changed the situation
dramatically and in principle stable isotope abundances in 10 or more additional elements with masses
less than 100 amu are now accessible. Of particular
interest here are the stable isotopes of the transition
metals. These have the potential to provide new
insights into processes such as occurred in the solar
nebula, during planetary accretion and melting as
well as those that link inorganic and biological
chemistry in low temperature environments.
The aim of the project is to investigate the use of
isotopic measurements in the study of inorganic and
biological geochemistry and cosmochemistry of Fe,
Cu and Zn. It commences here with Cu-isotopes in
terrestrial materials. The first search for natural variations in Cu-isotope abundances was made by Walker
et al. 1958. who claimed a total range of 9 per mil
in 65 Cur63 Cu ratios in Cu minerals, sediments and
organic samples. These results were supported by the
more extensive study of Shields et al. 1965., but in
both cases the analytical precision achieved between
1 and 1.5 per mil was relatively poor by modern day
standards. More recently, some preliminary results
on Cu isotope measurements using plasma-source
mass spectrometry have been reported Marechal et
al., 1998, 1999; ONions et al., 1998..
In this study, the detailed methodology for high
precision measurement on Cu isotopes using multiple-collector plasma-source mass spectrometry is
presented together with the results obtained on a
range of native copper, Cu-carbonate and Cu-sulphide
samples from terrestrial and marine environments.
The observed variations in 65 Cur63 Cu ratios are
some 40 times the analytical precision at 2 s level.
The potential of Cu isotopes as tracers for hydrothermal systems is evaluated and a preliminary model
for the Cu-isotopic variation in hydrothermal sulphides is proposed.
isotopic ratio measurements e.g., Walder and Freedman, 1992.. It has been successfully applied to the
analyses of W, Hf and Nd isotopes where mass
discrimination effects have been corrected for using
a pair of stable and non-radiogenic isotopes e.g.,
Walder et al., 1993; Lee and Halliday, 1995; Halliday et al., 1996., or by the addition of Tl as a dopant
in the case of Pb Walder et al., 1993; Belshaw et al.,
1998; Halliday et al., 1998.. For Cu, which has two
isotopes, 63 Cu and 65 Cu, the natural variation in
65
Cur63 Cu ratio is determined relative to a Cu standard such as NIST 976, and the isotopic variability
expressed in relative term as has been the practice
for O and C isotopes. A further requirement for
highly precise and accurate isotopic measurements is
to ensure that any mass spectral interferences are
either eliminated or negligible, and to ascertain that
no isotope fractionation occurred during sample
preparation and its introduction into the mass spectrometer.
2. Analytical methodology
2.2. Repeatability of
Multiple collector plasma source mass spectrometry is a relatively new technique for high precision
65
65
Cu r63Cu measurements
141
65 Cu s
R sample
R standard
y 1 = 10 000
1.
Fig. 1. Repeatability of 65 Cur63 Cu ratio measurements. a. Repeatability of 65 Cur63 Cu ratio measurements using a Zn-doping
technique. The measurements are expressed as 65 Cu units which
are parts per 10 4 deviation from the average value of Romil
Cu-solution. Mass spectrometric discrimination is corrected by
normalising 68 Znr66 Zn ratio of dopant zinc to a fixed value. The
analyses were carried out over a period of one week which
demonstrate a repeatability better than 0.3 units at 2 s level. b.
Repeatability of 65 Cur63 Cu ratio measurements using the standard-sample-standard approach. The sample used here is
Romil Cu-solution, and the standard is NIST 976. The results
were acquired over a period of two weeks and demonstrate an
external precision better than 0.6 units at 2 s level. The 65 Cu
value of the Romil Cu-solution obtained is 3.43"0.16 2 sm .
relative to the standard material NIST 976. Error bars represent
the internal precision at 2 s level.
142
working conditions used. However, the mass fractionation may vary with different operating conditions and during the course of an analytical session.
Variation of mass fractionation between analyses of
sample and standard is a source of error and overall
results in the repeatability of 65 Cur63 Cu measurements being significantly worse than the precision of
single measurements Fig. 1..
The two approaches described above are both
applicable to geological and biological samples. The
advantage of the standard-sample bracketing
technique is that Cu may be analysed in the presence
of other elements so long as they do not interfere
spectrally or influence the mass discrimination. The
Zn-doping technique however involves fewer measurements and is both faster and more time efficient.
It further offers improved repeatability Fig. 1a and
b. through minimisation of any time-dependence in
the mass discrimination. For the present study of
Cu-minerals, the standard-sample bracketing
method has been chosen. This is because, to avoid
any possibility of chromatographic fractionation of
Cu isotope, samples used in this study are not subjected to chemical purification. This method enables
the variation in Cu-isotopes observed to be confidently ascribed to processes occurring in nature.
143
3. Results
Samples of native copper, carbonate and chalcopyrite from continental and ocean floor deposits
have been analysed. The results of 65 Cur63 Cu analyses are presented in Tables 1 and 2 as 65 Cu units
deviations in parts per 10 4 . relative to the NIST 976
Cu standard. The results of repeat analyses are also
presented Tables 1 and 2., which show an excellent
agreement with the external precision of 0.6 units
defined by the instrument repeatability Fig. 1b..
These results show an overall variation of 22 in
Table 1
Cu-isotope results for samples from continental ore deposits
Sample
Natie copper
OUM22647
OUM15126
OUM15120
OUM00061
OUM15127
OUM00078
Ditto
Average
Malachite [Cu 2 (OH)2 CO3 ]
OUM23585
Azurite [Cu(OH)2 (CO3 )2 ]
OUM1616
OUM11242
OUM6629
OUM1637
Ditto
Average
Chalcopyrite (CuFeS2 )
E147
OUM25139
Acc. Ser. 25526
OUM22443
OUM21981
Acc. Ser. 25295
Ditto
Acc. Ser. 25300
Acc. Ser. 25301
Acc. Ser. 25309
Ditto
Acc. Ser. 25312
Ditto
Average
Locality
65 Cu
y2.00
4.47
5.40
y3.31
4.05
12.50
12.30
3.84
y2.58
All samples listed in this table are from the Oxford University Museum of Natural History.
15.85
14.60
14.71
4.43
4.58
12.42
y1.07
0.65
1.28
y1.46
y1.43
y1.88
y2.27
3.98
y1.84
y2.20
y1.83
y5.26
y6.17
y0.97
144
Table 2
Cu-isotope composition in chalcopyrite from black smoker sulphides deposits
Age ka. a
Sample
Locality
114-2-15
1156-2-3
1154-1-2
Ditto
Average
HR2505
HR2504
HR1805
Ditto
Average
G-4
G-7
G-17
G-22
G-14
Ditto
Average
2625-4-3
Ditto
2625-3-1
Ditto
2625-3-3
Ditto
2625-3-4
Ditto
Average
65 Cu
7.75
3.40
11.47
11.71
7.58
20
20
y4.82
y1.86
6.80
7.79
0.21
9
9
9
9
9
y4.43
y3.67
y3.45
y2.37
y4.55
y4.43
y3.68
3.38
3.09
3.57
3.24
6.46
6.36
10.87
11.30
6.04
Age data are from Lalou et al., 1985. The age for active vent is zero.
145
Fig. 3. Histogram showing the distribution of 65 Cu values in a selection of Cu-minerals from continental ore deposits. Results are
expressed as 65 Cu values relative to NIST 976. a. Native copper samples from North America, Europe and Russia. b. Cu-carbonates
from North America, Europe and Russia. c. Chalcopyrite samples from igneous-hosted ore deposits in North America, Europe and Africa.
Overall, carbonate samples tend to be enriched in 65 Cu, whereas the igneous-hosted chalcopyrites show smaller variation and 63 Cu
enrichment.
146
4. Discussion
4.1. The natural ariation of Cu isotopes and low
temperature mass fractionation
A variation of 22 units in Cu-isotope compositions has been observed in this study. This result
unambiguously demonstrates that natural variations
in 65 Cur63 Cu exist as claimed in the earliest studies
by Walker et al. 1958. and Shields et al. 1965..
These authors reported variations in Cu isotope ratios of 8 and 12 per mil, respectively. However, of
the more than 140 samples analysed, over 90% show
Cu isotopic variations which are within or around
their analytical errors of 1 to 1.5 per mil. Taken
together, these indicate that the Cu isotope variations
in the majority of natural samples are within the
range observed in this study. It is worth pointing out
that the variation of 22 units observed in this
study is some 40 times the analytical uncertainty at
95% confidence. This immediately demonstrates the
enormous potential for using stable Cu isotopes as
tracers in geological and planetary processes.
The variations in 65 Cu values observed in this
study display some interesting regularities. Samples
of native copper, carbonate and black smoker chalcopyrite display large differences in 65 Cu values
even at a single locality, whereas chalcopyrite samples in igneous-hosted continental deposits show remarkably similar Cu-isotope compositions worldwide. It is worth pointing out that those samples with
highly variable 65 Cu values have involved aqueous
solutions at relatively low temperature in their formation. This suggests that isotopic variations principally reflect mass fractionation in response to low
temperature processes rather than the source heterogeneity.
4.2. Cu isotopic ariation in black smoker sulphides
and hydrothermal circulation
wall give 65 Cu values of 6.46 and 10.45, respectively, whereas sample 2625-3-1 which was collected
from the chimney top of the same vent Site 3. has
65 Cu s 3.57. Sample 2526-4-3 collected from the
beehive top of another chimney Site 4., which is
some 30 m apart from Site 3, yields a 65 Cu value of
3.38 Table 2..
Seyfried and Ding, 1993. that sulphide deposits developed at the sites of black smokers in the oceans
derive their Cu predominantly from a mantle source.
Moreover, the Cu concentration in seawater of ca
0.001 to 0.007 mmol kgy1 Wu and Boyle, 1997. is
exceedingly low compared to that of 9.7 to 150
mmol kgy1 Elderfield and Schultz, 1996; Edmonda
et al., 1996. in high temperature hydrothermal fluids.
Although the Cu isotope composition in seawater is
presently unknown, it is unlikely that the difference
in 65 Cu between seawater and the sulphides is
sufficiently large to produce an observable effect
given the limitations imposed by the elemental mass
balance. Furthermore, mid-ocean ridge sides where
the samples were collected are sediment-starved, and
little possibility exists therefore for the incorporation
of significant amounts of Cu come from a sedimentary source. Therefore mixing models between seawater and hydrothermal fluids or mantle source used
to explain S-isotopes in these sulphide deposits
Woodruff and Shank, 1988; Knott, 1995; Duckworth et al., 1995. and Li isotopes in ocean basalts
Chan et al., 1992, 1993. are unlikely to be relevant
to the observed 65 Cu variations. Also, it seems
unlikely that the Cu-isotopic variations in the mantle
will much exceed the variations of 65 Cu in
igneous-hosted sulphides. Therefore, 65 Cu variations recorded in the black smoker sulphides are
thought to predominantly reflect processes operating
in the hydrothermal systems themselves.
It should be recalled that the 65 Cu values from
the black smoker sulphides display the following
features: 1. relative to active high temperature hydrothermal vents, the old inactive sulphide deposits
have lower 65 Cu values and show smaller variations; 2. in a single active chimney, 65 Cu value
decreases from the bottom to top. These observations
may be explained by means of a two-stage model,
which involves: 1. the preferential leaching of 65 Cu
during the hydrothermal processes, and 2. subsequent isotopic exchange between the early formed
Cu-sulphides and 65 Cu-depleted late-stage fluids.
Analogous to the mechanism suggested to explain
Li isotopes Chan et al., 1993., the heavy isotope of
Cu may be preferentially leached from the basalt
source and the Cu-isotope composition of the hydrothermal fluids is then time-dependent. Thus, early
formed sulphides deposited in the lower part of a
147
148
Acknowledgements
We thank M.T. Price for providing the samples
from the Oxford University Museum of Natural History. A.N. Halliday is thanked for his comments on
an early version, and R.J.P. Williams is thanked for
his interest in this project. Ariel Anbar and an anonymous reviewer are thanked for their formal comments on the paper. [MB]
References
Belshaw, N.S., Feedmann, P.A., ONions, R.K., Frank, M., Guo,
Y., 1998. A new variable dispersion double-focusing plasma
mass spectrometer with performance illustrated for Pb isotopes. Int. J. Mass Spectrom. Ion Phys. 181, 5158.
Chan, L.H., Edmond, J.M., Thompson, G., 1993. A lithium
isotope study of hot springs and metabasalts from mid-ocean
ridge hydrothermal systems. J. Geophys. Res. 98, 96539659.
Chan, L.H., Edmond, J.M., Thompson, G., Gillis, K., 1992.
Lithium isotopic composition of submarine basalts: implications for the lithium cycle in the oceans. Earth Planet. Sci.
Lett. 108, 151160.
Deutsch, A., Grieve, R.A.F., 1994. The Sudbury structure
constraints on its genesis from Lithoprobe results. Geophys.
Res. Lett. 21, 963966.
Dietz, R.S., 1964. Sudbury structure as an astrobleme. J. Geol. 72,
412434.
Duckworth, R.C., Knott, R., Fallick, A.E., Rickard, D., Murton,
B.J., van Dover, C., 1995. Mineraloy and sulphur isotope
geochemistry of the Broken Spur sulphides, 298N, Mid-Atlantic
Ridge. In: Parson, L.M., Walker, C.L., Dixon, D.R. Eds..,
Hydrothermal Vents and Processes. Geol. Soc. Spec. Publ. 87,
pp. 175189.
Edmonda, H.N., German, C.R., Green, D.R.H., Huh, Y., Gamo,
T., 1996. Continuation of the hydrothermal fluid chemistry
time series at TAG, and the effects of ODP drilling. Geophys.
Res. Lett. 23, 34873489.
149