Anal. Chem. 2006, 78, 8477-8484
High Precision Measurements of
Non-Mass-Dependent Effects in Nickel Isotopes in
Meteoritic Metal via Multicollector ICPMS
David L. Cook,*,†,‡,§ Meenakshi Wadhwa,†,‡,§ Philip E. Janney,§ Nicolas Dauphas,†,‡,⊥
Robert N. Clayton,†,‡,⊥ and Andrew M. Davis†,‡,⊥
Department of the Geophysical Sciences, The University of Chicago, 5734 South Ellis Avenue, Chicago, Illinois 60637,
Chicago Center for Cosmochemistry, 5640 South Ellis Avenue, Chicago, Illinois 60637, Department of Geology, The Field
Museum, 1400 South Lake Shore Drive, Chicago, Illinois 60605, and Enrico Fermi Institute, 5640 South Ellis Avenue,
Chicago, Illinois 60637.
We measured the Ni isotopic composition of metal from
a variety of meteorite groups to search for variations in
the 60Ni abundance from the decay of the short-lived
nuclide 60Fe (t1/2 ) 1.49 My) and for possible nucleosynthetic effects in the other stable isotopes of Ni. We
developed a high-yield Ni separation procedure based on
a combination of anion and cation exchange chromatography. Nickel isotopes were measured on a single-focusing, multicollector, inductively coupled mass spectrometer (MC-ICPMS). The external precision on the massbias-corrected 60Ni/58Ni ratio ((0.15 E; 2σ) is comparable
to similar studies using double-focusing MC-ICPMS. We
report the first high-precision data for 64Ni, the least
abundant Ni isotope, obtained via MC-ICPMS. The external precision on the mass-bias-corrected 64Ni/58Ni ratio
((1.5 E; 2σ) is better than previous studies using thermal
ionization mass spectrometry. No resolvable excesses
relative to a terrestrial standard in the mass-bias-corrected
60Ni/58Ni ratio were detected in any meteoritic metal
samples. However, resolvable deficits in this ratio were
measured in the metal from several unequilibrated chondrites, implying a 60Fe/56Fe ratio of ∼1 × 10-6 at the time
of Fe/Ni fractionation in chondritic metal. A 60Fe/56Fe
ratio of (4.6 ( 3.3) × 10-7 is inferred at the time of Fe/
Ni fractionation on the parent bodies of magmatic iron
meteorites and pallasites. No clearly resolvable non-massdependent anomalies were detected in the other stable
isotopes of Ni in the samples investigated here, indicating
that the Ni isotopic composition in the early solar system
was homogeneous (at least at the level of precision
reported here) at the time of meteoritic metal formation.
Meteorites provide a record of events and processes that
occurred during the formation and early evolution of the solar
system. Various phases in undifferentiated meteorites (e.g.,
ordinary and carbonaceous chondrites) contain information on
* Towhomcorrespondenceshouldbeaddressed.E-mail: davecook@uchicago.edu.
† The University of Chicago.
‡ Chicago Center for Cosmochemistry.
§ The Field Museum.
⊥ Enrico Fermi Institute.
10.1021/ac061285m CCC: $33.50
Published on Web 11/15/2006
© 2006 American Chemical Society
nebular processes, such as condensation, and on parent body
processes, such as accretion, metamorphism, and aqueous alteration. These primitive meteorites may also provide a record of
the precursor materials (e.g., presolar dust grains) involved in
the formation of the solar nebula. Differentiated meteorites (e.g.,
eucrites, pallasites, and iron meteorites) record diverse parent
body processes, such as silicate differentiation and core formation.
Constraining the timing of these processes is critical for unraveling
early solar system history. The isotopic record left by the former
presence of short-lived, now extinct, radionuclides in meteorites
can provide chronometric data for this early period. Chronometry
based on a number of extinct radionuclides has been applied to
meteoritic materials.1,2 The precision of ages based on short-lived
chronometers is typically better than 1 My, which makes them
ideally suited to studying the timing of early solar system
processes.1,2
Unambiguous evidence of excesses in radiogenic 60Ni from
the decay of 60Fe (t1/2 ) 1.49 My) was first reported in bulk eucrite
samples.3,4 Since then, analyses of various phases in unequilibrated
chondrites have revealed 60Ni excesses in sulfides, oxides, and
silicates.5-8 Recently, 60Ni excesses of up to 1.5 were reported
for Fe-Ni metal from several iron meteorites9 and ordinary
chondrites;10 these were the first reports of excess 60Ni in
meteoritic metal. The metal phase in meteorites may have been
subjected to both nebular and parent body processes, such as
condensation, oxidation, melting, metal-silicate segregation, and
core crystallization.11 To date, only the 107Pd-107Ag 12,13 (t1/2 ) 6.5
(1) McKeegan, K. D.; Davis, A. M. In Treatise on Geochemistry; Davis, A. M.,
Ed.; Elsevier Pergamon: San Diego, CA, 2004; Vol. 1, pp 431-460.
(2) Gilmour, J. D. Space Sci. Rev. 2000, 92, 123-132.
(3) Shukolyukov, A.; Lugmair, G. W. Science 1993, 259, 1138-1142.
(4) Shukolyukov, A.; Lugmair, G. W. Earth Planet Sci. Lett. 1993, 119, 159166.
(5) Tachibana, S.; Huss, G. R. Astrophys. J. 2003, 588, L41-L44.
(6) Guan, Y.; Huss, G. R.; Leshin, L. A.; MacPherson, G. J. Meteorit. Planet.
Sci. 2003, 38, A138.
(7) Mostefaoui, S.; Lugmair, G. W.; Hoppe, P. Astrophys. J. 2005, 211, 271277.
(8) Tachibana, S.; Huss, G. R.; Kita, N. T.; Shimoda, G.; Morishita, Y. Astrophys.
J. 2006, 639, L87-L90.
(9) Moynier, F.; Telouk, P.; Blichert-Toft, J.; Albarède, F. Lunar Planet Sci.
2004, 35, no. 1286.
(10) Moynier, F.; Blichert-Toft; J.; Telouk; P.; Albarède, F. Lunar Planet Sci.
2005, 36, no. 1593.
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006 8477
My) and 182Hf-182W 14-17 (t1/2 ) 8.9 My) chronometers have been
applied toward obtaining high-resolution chronology of metal
samples. The findings of Moynier et al.9,10 suggested that the
60Fe-60Ni chronometer may add a tool for deciphering the history
of meteoritic metal.
The resolution of small time differences and of non-massdependent isotopic anomalies (possibly of nucleosynthetic origin)
using short-lived chronometers requires isotope measurements
of high precision. Nickel isotopes have been measured using
secondary ion mass spectrometry (SIMS),5-8 thermal ionization
mass spectrometry (TIMS),3,4,18-20 and multicollector inductively
coupled plasma mass spectrometry (MC-ICPMS).9,10,21 Analyses
using SIMS can provide high spatial resolution (e.g., 5 µm) but
suffer from relatively poor precision.5-8 Isobaric interferences on
58Ni by 58Fe and on 64Ni by 64Zn in SIMS analyses make it difficult
to measure these Ni isotopes. Thus, the full complement of Ni
isotopes (58Ni, 60Ni, 61Ni, 62Ni, 64Ni) cannot be easily determined
using this technique. Analysis via chemical separation and
purification of Ni followed by TIMS allows for the measurement
of all five Ni isotopes and for the correction of 58Fe and 64Zn
interferences. The typical precision of the isotope ratio measurements by this technique is better than by SIMS: a precision on
60Ni/58Ni, the ratio used for chronology, as good as ( 0.6 (2σ)
has been achieved.3 However, precisions on the other Ni isotope
ratios are on the order of at least several epsilon units.18-20 Like
TIMS, MC-ICPMS also provides the ability to measure all Ni
isotopes and correct for interferences from 58Fe and 64Zn. Additionally, significantly better precision on all Ni isotope ratios may
be possible. Thus far, high-precision Ni isotope ratio measurements have been made only on double-focusing MC-ICPMS
instruments.9,10,21 We report the first highly precise and accurate
measurements of Ni isotopes using a single-focusing MC-ICPMS.
This instrument employs a hexapole collision cell to thermalize
ions and remove interfering polyatomic species. The external
precision of Ni isotope ratio measurements with this technique
is comparable to or better than that of prior work.9,10,21 On the
basis of high precision analyses of Ni isotopes in metal from a
variety of undifferentiated and differentiated meteorites using this
technique, we present implications for the initial abundance of
60Fe and for the degree of Ni isotopic homogeneity in the early
solar system.
EXPERIMENTAL METHODS
Samples and Sample Preparation. Samples were chosen
to represent a wide variety of meteorite classes,22 including
magmatic (IIA, IIB, IID, IIIA, IIIB, IVA, and IVB) and nonmagmatic (IAB and IIICD) iron meteorites, pallasites (main group and
(11) Kelly, W. R.; Larimer, J. W. Geochim. Cosmochim. Acta 1977, 41, 93-111.
(12) Chen, J. H.; Wasserburg, G. J. Geophys. Monogr. 1996, 95, 1-20.
(13) Carlson, C. W.; Hauri, E. H. Geochim. Cosmochim. Acta 2001, 65, 18391848.
(14) Lee, D.-C.; Halliday, A. N. Science 1996, 274, 1876-1879.
(15) Horan, M. F.; Smoliar, M. I.; Walker, R. J. Geochim. Cosmochim. Acta 1998,
62, 545-554.
(16) Lee, D. C. Earth Planet Sci. Lett. 2005, 237, 21-32.
(17) Markowski, A.; Quitté, G; Halliday, A. N.; Kleine, T. Earth Planet Sci. Lett.
2006, 242, 1-15.
(18) Morand, P.; Allègre, C. J. Earth Planet Sci. Lett. 1983, 63, 167-176.
(19) Shimamura, T.; Lugmair, G. W. Earth Planet Sci. Lett. 1983, 63, 177-188.
(20) Birck, J. L.; Lugmair, G. W. Earth Planet Sci. Lett. 1988, 90, 131-143.
(21) Quitté, G; Meier, M.; Latkoczy, C.; Halliday, A. N.; Günther, D. Earth Planet
Sci. Lett. 2006, 242, 16-25.
8478
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
Eagle Station), and chondrites (LL, H, EH, CR, and CBa). Small
pieces (<50 mg) of fresh metal were cut from iron meteorites,
pallasites, and terrestrial josephinite (a naturally occurring FeNi metal) using a slow-speed saw with a diamond wafering blade.
Samples were examined under a binocular microscope, and any
rusty surfaces were mechanically cleaned by polishing with silicon
carbide. Sample sizes for chondrites were determined on the basis
of the metal content and ranged from 12.5 to 104 mg. The
chondrite samples were crushed with a boron carbide mortar and
pestle. These crushed samples were passed through a series of
sieves, the size fractions greater than 75 µm were combined, and
the metal was separated with a hand magnet. The metal separates
were rinsed twice with acetone and allowed to dry. A metal globule
from the chondrite Gujba was sampled and processed in a manner
similar to the iron meteorites.
Sample Digestion. Chondrite metal and two pallasite samples
(Albin and Molong) that contained visible silicate inclusions were
first digested for 24 h on a hot plate at 90 °C in 15-mL Teflon
beakers containing 5 mL of 1.0 M HCl. Next, the solutions were
centrifuged for 30 min to separate any silicate grains, and the
supernatant containing the dissolved metal was removed after
centrifugation and evaporated to dryness. Finally, 3 mL of reverse
aqua regia (2:1 conc HNO3 to conc HCl) was added to these dried
samples, which were then redigested by heating at 120 °C for 24
h. The other two pallasites, all iron meteorites, and the josephinite
samples were digested directly by treatment with 3 mL of reverse
aqua regia for 24 h at 120 °C. Following digestion in reverse aqua
regia, all the samples were evaporated to dryness; equilibrated
for several hours in concentrated HCl; evaporated to dryness
again; and finally, dissolved in 6 M HCl. This solution was split
into two aliquots, one for chemical separation of Ni and another
for Fe/Ni elemental ratio measurements.
Chemical Separation of Nickel. Nickel was separated from
the matrix using a combination of anion and cation exchange
chromatography. Nickel was first separated on an anion exchange
column.23 Disposable Bio-Rad Poly-Prep columns were packed
with 1 mL of AG1-X8 anion resin (200-400 mesh). The columns
were washed and conditioned with the following: 10 mL of H2O,
5 mL of 1 M HNO3, 10 mL of H2O, 10 mL of 0.4 M HCl, and 3 mL
of 6 M HCl. The sample was loaded in 0.2 mL of 6 M HCl, and Ni
was eluted with an additional 6 mL of 6 M HCl in a sequence of
0.5, 0.5, 1.0, 2.0, and 2.0 mL. At this molarity, Fe partitions strongly
onto the resin, whereas Ni is not retained,24 providing efficient
separation of these two elements. The eluted Ni fraction was taken
to dryness and then dissolved in 0.2 mL of 6 M HCl.
Nickel was further purified on a cation-exchange column. A
protocol was developed on the basis of the behavior of Ni in an
acetone-HCl medium.25,26 Disposable Bio-Rad Poly-Prep columns
were packed with 2 mL of AG50W-X4 cation resin (200-400
mesh). The columns were washed and conditioned with the
following: 10 mL of H2O, 10 mL of 4 M HCl, 10 mL of H2O, 6 mL
(22) Krot, A. N.; Keil, K.; Goodrich, C. A.; Scott, E. R. D. In Treatise on
Geochemistry; Davis, A. M., Ed.; Elsevier Pergamon: San Diego, CA, 2004;
Vol. 1, pp 83-128.
(23) Dauphas, N.; Janney, P. E.; Mendybaev, R. A.; Wadhwa, M.; Richter, F. M.;
Davis, A. M.; Hines, R.; Foley, C. N. Anal. Chem. 2004, 76, 5855-5863.
(24) Helfferich, F. G. Ion Exchange; McGraw Hill: New York, 1962.
(25) Korkisch, J.; Ahluwalia, S. S. Talanta 1967, 14, 155-170.
(26) Strelow, F. W. E.; Victor, A. H.; van Zyl, C. R.; Eloff, C. Anal. Chem. 1971,
43, 870-876.
of 30% H2O/70% acetone, and 6 mL of 0.6 M HCl/90% acetone.
The sample was loaded in 2 mL 0.6 M HCl/90% acetone, and the
column was washed with an additional 4 mL of 0.6 M HCl/90%
acetone. Nickel was eluted with 6 mL of 4 M HCl in a sequence
of 0.5, 0.5, 1.0, 2.0, and 2.0 mL. In the above mixture of HCl and
acetone, Ni partitions onto the resin (Kd > 227), whereas the
interfering elements Fe and Zn do not (Kd < 1).25 One advantage
of this method is that it removes both ferrous and ferric Fe,26
whereas only ferric Fe is removed by the anion column procedure.24 Furthermore, several elements present in meteoritic metal
that may not be effectively separated (e.g., Co, Cu), or are not
separated at all (e.g., Mn), on the anion column24 are separated
on the cation column.25 It is important to cap the column once
reagents containing acetone are introduced in order to prevent
acetone evaporation, which could lead to a change in the
partitioning behavior of the elements.25
A third anion exchange column was employed for some
samples to remove any Ti that may be present. Titanium, being a
lithophile element, is not expected to be present in any significant
amounts in metallic samples. However, metal samples containing
inclusions of silicate (i.e., pallasites) or hosted in a silicate matrix
(i.e., chondrites) could include a minor amount of silicate that,
although it would not significantly affect the Ni (which is a major
element in the metal but only a trace element in the silicates),
could contribute some Ti. Interferences on 62Ni and 64Ni by TiO
species can occur. Because it is not possible to easily monitor
and correct for interferences from TiO species during the isotopic
analysis, it is imperative that the sample solutions be free from
Ti. Therefore, the metal from the pallasites Albin and Molong and
all chondrite samples were subjected to a third column exchange
separation.
Disposable Bio-Rad Poly-Prep columns were packed with 1 mL
of AG1-X8 anion resin (200-400 mesh). The columns were
washed and conditioned with the following: 10 mL of H2O, 5 mL
of 1 M HNO3, 10 mL of H2O, 10 mL of 0.4 M HCl, 5 mL of H2O,
and 5 mL of an 0.5 M HF/1 M HCl mixture. The sample was
loaded in 0.2 mL of 0.5 M HF/1 M HCl, and Ni was eluted with
an additional 3 mL of 0.5 M HF/1 M HCl in a sequence of 0.5,
0.5, 1.0, and1.0 mL. Tests showed that under these conditions,
Ni was quantitatively recovered, and Ti was fully retained by the
resin. The total procedural blank (≈3 ng) is insignificant compared
to the amount of Ni in the samples.
Iron/Ni Elemental Ratio Measurements. The Fe/Ni ratios
were determined in sample solutions using a Varian ICPMS
instrument equipped with a quadrupole mass analyzer at the Field
Museum. An aliquot of digested but chemically unprocessed
sample was dried and then dissolved in 3% HNO3. A known
amount of a manganese concentration standard was added to the
sample solution and served as an internal standard. The isotopes
55Mn, 57Fe, and 60Ni were measured, and Fe and Ni concentrations
were calculated using calibration curves obtained with external
standards. Each aliquot was measured three times, and the
uncertainty represents the 2σ standard deviation of all measurements.
Nickel Isotopic Measurements. Nickel isotopic measurements were performed at the Isotope Geochemistry Laboratory
of the Field Museum on a Micromass (now GV Instruments)
IsoProbe MC-ICPMS. This instrument has nine Faraday collectors;
Table 1. The Ni Isotopic Compositions of Chemically
Processed Aliquots of the Ni Isotopic Standard SRM
986 and of Terrestrial Josephinite
sample
60 ( 2σ
61 ( 2σ
64 ( 2σ
n
SRM no. 1
SRM no. 2
SRM no. 3
Josephinite no. 1
Josephinite no. 2
2-Column Chemistry
-0.05 ( 0.27
0.19 ( 1.12
-0.06 ( 0.16 -0.67 ( 2.21
-0.04 ( 0.40
0.05 ( 1.74
-0.15 ( 0.11
0.58 ( 0.72
0.16 ( 0.29
0.27 ( 0.72
2.3 ( 4.4
1.4 ( 1.7
0.4 ( 1.6
0.4 ( 2.0
2.6 ( 2.0
5
5
5
9
5
SRM no. 1
SRM no. 2
SRM no. 3
Josephinite no. 1
Josephinite no. 2
3-Column Chemistry
-0.01 ( 0.43 -0.64 ( 1.16
-0.01 ( 0.20 -0.60 ( 0.87
0.12 ( 0.14
0.17 ( 1.89
-0.08 ( 0.06
0.71 ( 0.48
-0.12 ( 0.06 -0.05 ( 0.37
0.0 ( 3.2
1.3 ( 2.0
0.0 ( 2.6
0.6 ( 0.7
0.2 ( 0.6
5
5
5
14
13
thus, all Ni isotopes can be measured simultaneously, and 57Fe
and 66Zn can be monitored and used to correct for isobaric
interferences on 58Ni from 58Fe (0.28 atom % of total Fe) and on
64Ni from 64Zn (49.18 atom % of total Zn). The sample solution (1
ppm in 3 wt % HNO3) was introduced through a PFA Teflon
nebulizer (100 µL/min) in a Cetac Aridus desolvating system using
a PFA spray chamber heated to 95 °C. Argon and N2 were
introduced into the desolvating system at approximately 3 L min-1
and 35 µL min-1, respectively. Argon was introduced into the
collision cell at 1.8 mL min-1 as a thermalizing collision gas. The
instrument was optimized to obtain a signal of g6.0 V on 58Ni
when running a 1 ppm Ni solution. This generates a signal of
g100 mV on 64Ni, the least abundant Ni isotope (0.925%).27
Sampler and skimmer cones made of Ni were used because they
were found to provide better signal stability than Al cones. Despite
the use of Ni cones, the background signal is negligibly small
(1-3 mV on 58Ni) compared to the sample signal. Samples were
measured via the standard-sample bracketing technique using
the NIST SRM 986 as the Ni isotope standard. SRM 986 is the
only commercially available Ni standard with a certified isotopic
composition.27 Samples were corrected for mass bias using an
exponential law and 62Ni/58Ni ≡ 0.053388.27 The analytical protocol
consisted of alternating between standard and sample solutions,
with each being measured for 200 s (admittance delay 2 min).
Each 200-s measurement (consisting of 20 cycles of 10-s integrations) was preceded by 4 min of washout and an on-peak blank
measurement consisting of a 60-s integration measurement while
aspirating a clean 3 wt % HNO3 solution. Each reported datum
comprises the mean of a minimum of five repeat measurements.
Precision and Accuracy. All isotope ratio data (Tables 1 and
2) are reported in units, given as i ) [(Rsample - Rstandard)/Rstandard]
× 104, where R is the mass-bias-corrected iNi/58Ni ratio (i ) 60,
61, or 64). Internal precisions for individual samples represent
the standard error of the mean (2σm). The external precision was
determined by repeated analyses of an Aesar Ni concentration
standard over a 24-month period. The mean values of the isotopic
ratios from each analysis of the Aesar Ni solution were calculated
from five repeat measurements. These mean values for all ratios
are identical to SRM 986 within uncertainty. Figure 1 shows these
data for the mass-bias-corrected 60Ni/58Ni ratio in epsilon units
(27) Gramlich, J. W.; Machlan, L. A.; Barnes, I. L.; Paulsen, P. J. J. Res. Natl.
Bur. Stand. (US) 1989, 94, 347-356.
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
8479
Table 2. The
56Fe/58Ni
sample
Ratios and Ni Isotopic Compositions of Meteoritic Metal
group
56Fe/58Ni
( 2σ
60 ( 2σ
61 ( 2σ
64 ( 2σ
n
Renazzo
Gujba
Semarkona
Bishunpur
Forest Vale
Indarch
CR
CBa
LL 3.0
LL 3.1
H4
EH4
21.9 ( 0.3
22.0 ( 0.2
14.4 ( 0.2
18.4 ( 0.1
16.4 ( 0.3
18.3 ( 0.5
-0.23 ( 0.10
-0.23 ( 0.08
-0.25 ( 0.10
0.01 ( 0.25
0.07 ( 0.30
-0.13 ( 0.20
0.31 ( 0.54
-0.33 ( 0.33
0.98 ( 0.23
0.32 ( 1.26
-0.05 ( 1.86
0.47 ( 1.29
0.1 ( 0.8
-0.2 ( 0.7
-0.3 ( 0.7
-0.7 ( 1.7
0.0 ( 2.0
0.3 ( 4.0
14
13
10
5
5
5
Eagle Station
Albin
Brenham
Molong
PES
PMG
PMG
PMG
7.16 ( 0.09
14.3 ( 0.2
10.3 ( 0.1
11.4 ( 0.1
-0.12 ( 0.12
0.04 ( 0.17
-0.02 ( 0.26
-0.06 ( 0.13
0.39 ( 0.62
-0.53 ( 0.70
0.23 ( 1.22
0.17 ( 0.76
2.0 ( 1.6
0.8 ( 1.3
1.8 ( 4.1
0.0 ( 1.9
5
9
5
5
Coahuila
Santa Luzia
Carbo
Bella Roca
Casas Grandes
Henbury
Gibeon
Yanhuitlan
Cape of Good Hope
Hoba
Tlacotepec
IIAB
IIAB
IID
IIIAB
IIIAB
IIIAB
IVA
IVA
IVB
IVB
IVB
23.1 ( 0.1
19.9 ( 0.3
12.6 ( 0.2
12.4 ( 0.2
17.2 ( 0.3
17.5 ( 0.2
15.9 ( 0.2
16.9 ( 0.3
7.27 ( 0.10
7.01 ( 0.07
7.17 ( 0.03
0.00 ( 0.14
0.05 ( 0.10
-0.08 ( 0.25
0.03 ( 0.17
0.02 ( 0.09
-0.01 ( 0.12
-0.13 ( 0.19
0.00 ( 0.36
-0.13 ( 0.39
-0.21 ( 0.07
-0.18 ( 0.20
0.41 ( 1.45
0.06 ( 0.55
-0.07 ( 1.06
0.26 ( 1.23
0.00 ( 0.36
-0.04 ( 0.35
-0.02 ( 1.86
0.82 ( 0.95
0.11 ( 2.30
-0.21 ( 0.44
-0.15 ( 0.76
1.0 ( 3.4
1.5 ( 1.0
-0.3 ( 2.1
-1.0 ( 0.7
1.4 ( 1.1
-0.4 ( 0.6
2.2 ( 1.7
-0.1 ( 2.9
1.9 ( 1.2
-0.1 ( 0.7
0.3 ( 0.9
5
10
5
8
14
14
5
5
5
13
9
Canyon Diablo
Toluca
Dayton
Mundrabilla
IAB
IAB
IIICD
IIICD
18.6 ( 0.3
19.7 ( 0.3
7.08 ( 0.05
16.2 ( 0.4
0.06 ( 0.07
-0.06 ( 0.09
0.08 ( 0.21
-0.12 ( 0.08
0.14 ( 0.58
-0.02 ( 0.34
-0.13 ( 0.75
-0.01 ( 0.47
1.9 ( 0.8
0.3 ( 0.7
2.3 ( 3.2
0.2 ( 0.5
19
15
5
15
Figure 1. 60 values for repeated analyses of an Aesar Ni solution
over the course of a 24-month period. Each datum represents the
mean of five repeat measurements performed during a single analysis
session. The individual error bars are 2σm errors, based on the five
repeat measurements for each datum. The external precision is the
standard deviation (2σ) based on all of the data plotted here and is
shown by the two dashed lines ((0.15 ).
(60); the external precision (2σ) is (0.15 . The external precisions
for the 61 and 64 values are (0.85 and (1.5 , respectively
(Figures S-1 and S-2).
Gramlich et al.28 measured terrestrial sulfides and metals and
showed that the mass-bias-corrected Ni isotopic compositions of
terrestrial samples do not deviate from those of SRM 986. Three
aliquots of SRM 986 were processed using our Ni separation
chemistry, both with and without the final cleanup column for Ti.
Two samples of terrestrial josephinite were also processed using
both procedures. As expected, the mean values for all ratios are
identical to SRM 986 within uncertainty (Table 1). Figure 2 shows
these data for the mass-bias-corrected 60Ni/58Ni ratios in epsilon
(28) Gramlich, J. W.; Beary, E. S.; Machlan, L. A.; Barnes, I. L. J. Res. Natl. Bur.
Stand. (US) 1989, 94, 357-362.
8480 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
Figure 2. 60 values for SRM 986 (circles) and terrestrial josephinite
(triangles) processed through Ni separation chemistry; filled symbols
included the Ti cleanup column. Plotted errors are 2σm; the external
precision (2σ) is shown by the two dashed lines ((0.15 ). Smaller
error bars on some josephinite samples reflect a larger number of
repeat measurements (Table 1).
units (60). These data confirm that our Ni separation chemistry
for metallic samples does not introduce any analytical artifacts
and demonstrate our ability to precisely and accurately measure
the Ni isotopic compositions of such samples.
Effects of Fe and Zn. As previously noted, Fe and Zn isotopes
interfere with Ni isotopes, and these interferences must be
corrected for when present. To test our ability to correct for these
interferences, we doped aliquots of SRM 986 with varying
concentrations of Fe and Zn (four aliquots each). Samples were
corrected for Fe and Zn interferences using 58Fe/57Fe ≡ 0.1330
and 64Zn/66Zn ≡ 1.7698. Because all Ni isotopes are normalized
to 58Ni, an interference from 58Fe has the potential to affect all of
the measured Ni ratios. Figure 3 shows the interference-corrected
60 values for the SRM 986 aliquots doped with Fe as well as an
undoped aliquot. These data show that the Fe interference
Figure 3. 60 vs the Fe/Ni elemental ratio. Data are from analyses
of four aliquots of SRM 986 doped with varying amounts of Fe and
one undoped aliquot (Fe/Ni ≈ 10-4). Plotted errors are 2σm; the
external precision (2σ) is shown by the two dashed lines ((0.15 ).
Figure 5. 60 vs the ratio of the sample-to-standard signal intensity.
Data are from analyses of five aliquots of SRM 986 diluted to various
concentrations and one undiluted aliquot (ratio ) 1). Plotted errors
are 2σm; the external precision (2σ) is shown by the two dashed lines
((0.15 ).
the Aesar solutions differed from the standard by 10% or less. All
measured sample solutions had signal intensities within 15% of
the standard signal intensity, and 95% of those were within 10%.
These discrepancies in standard-sample concentration matching
are well within the range investigated and do not affect our
measurements.
Figure 4. 64 vs the Zn/Ni elemental ratio. Data are from analyses
of four aliquots of SRM 986 doped with varying amounts of Zn and
one undoped aliquot (Zn/Ni ≈ 10-5). Note the scale break on the y
axis. Plotted errors are 2σm; the external precision (2σ) is shown by
the two dashed lines ((1.5 ).
correction is effective, even at high Fe concentrations (i.e., Fe/
Ni ) 0.1). Additionally, no resolvable effects were observed on
61 or 64 due to the presence of Fe. Figure 4 shows the
interference-corrected 64 values for the SRM 986 aliquots doped
with Zn as well as an undoped aliquot. These data show that the
Zn interference correction is not effective for Zn/Ni g 0.01 and
illustrate the importance of separating Ni from Zn during chemical
processing. For all the samples processed through Ni separation
chemistry, the Fe/Ni and Zn/Ni ratios (<3.7 × 10-2 and <4.5 ×
10-4, respectively) were substantially below the values for which
the interference corrections become ineffective. Therefore, our
chemical separation procedure for metallic samples provides Ni
with the purity required for accurate isotope measurements.
Effects of Signal Intensity. To determine whether differences
between the Ni signal intensities for standard and sample solutions
can affect our measurements and the level to which samplestandard concentration matching would be required, we analyzed
five aliquots of SRM 986 with varying Ni concentrations plus one
undiluted aliquot (i.e., 1 ppm) bracketed with a 1 ppm SRM 986
solution. No resolvable effects were observed for any of the Ni
isotope ratios, even when the sample signal differed from the
standard signal by 45%. Figure 5 shows 60 values from these
measurements. All measured Aesar solutions (Figure 1) had signal
intensities within 26% of the standard signal intensity, and 95% of
RESULTS AND DISCUSSION
Table 2 lists the Ni isotopic compositions, given in epsilon
units, of the meteoritic metal samples analyzed in this study and
the number of repeat measurements (n) made for each sample.
Also listed are the measured 56Fe/58Ni ratios for each sample. In
the following discussion, we will only consider those deviations
in ,60 61, and 64 values to be “resolvable” if they are outside of
either the 3σm error or the 2σ external precision, whichever is
larger.
Possible Nucleosynthetic Anomalies in Ni Isotopes in
Meteoritic Metal. Isotopic anomalies of nucleosynthetic origin
may be preserved in meteoritic metal. Anomalies in the isotopic
composition of Mo have been reported in several iron meteorites.29,30 Additionally, some irons also contain Ru isotope anomalies.31 In the case of the iron peak elements, anomalies in the
neutron-rich isotopes (e.g., 48Ca, 50Ti, 54Cr, 62Ni, 64Ni, 66Zn) have
been identified in CAIs20,32-36 but have thus far been noted only
for 54Cr in bulk samples of undifferentiated and differentiated
meteorites.37,38 Confirmation of the presence of such effects,
(29) Dauphas, N.; Marty, B.; Reisberg, L. Astrophys. J. 2002, 565, 640-644.
(30) Chen, J. H.; Papanastassiou, D. A.; Wasserburg, G. J.; Ngo, H. H. Lunar
Planet Sci. 2004, 35, no. 1431.
(31) Chen, J. H.; Papanastassiou, D. A.; Wasserburg, G. J. Lunar Planet Sci.
2003, 34, no. 1789.
(32) Lee, T.; Papanastassiou, D. A.; Wasserburg, G. J. Astrophys. J. 1978, 220,
L21-L25.
(33) Jungck, M. H. A.; Shimamura, T.; Lugmair, G. W. Geochim. Cosmochim.
Acta. 1984, 48, 2651-2658.
(34) Niederer, F. R.; Papanastassiou, D. A.; Wasserburg, G. J. Astrophys. J. 1980,
240, L73-L77.
(35) Niemeyer, S.; Lugmair, G. W. Earth Planet Sci. Lett. 1981, 53, 211-225.
(36) Loss, R. D.; Lugmair, G. W.; Davis, A. M.; MacPherson, G. J. Astrophys. J.
1994, 436, L193-L196.
(37) Trinquier, A.; Birck, J. L.; Allègre, C. J. Lunar Planet Sci. 2005, 36, no.
1259.
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
8481
particularly in meteorites from differentiated parent bodies, has
important implications for the efficiency and spatial scale of mixing
and homogenization in the solar nebula.
Our analyses show no clearly detectable non-mass-dependent
anomalies in the 61Ni/58Ni and 64Ni/58Ni ratios in the samples
analyzed here (Table 2; Figures S-3 and S-4). Possible exceptions
may be metal from the Semarkona chondrite and the Canyon
Diablo iron meteorite, which have slightly positive 61 and 64
values, respectively. However, in these two cases, the anomalies
are only marginally outside of the 2σ external precision and are
not considered unambiguous. If there had been systematic
anomalies in the neutron-rich isotopes of Ni (62Ni and 64Ni), our
normalization to the 62Ni/58Ni ratio would have resulted in
systematic anomalies in the ,60 61, and 64 values, which we do
not observe. A preliminary report by Bizzarro et al.39 suggests
the presence of anomalies in 62Ni in iron meteorites; however,
the reported effects are smaller than our analytical precision.
Therefore, at the current level of our analytical precision, we are
unable to unambiguously resolve any anomalies of nucleosynthetic
origin in the Ni isotopic compositions of meteoritic metal samples.
If, as is suggested by the data for CAIs,20 Ni isotopic heterogeneity
existed in the solar nebula, our data indicate that Ni was
homogenized (at least to the level of the analytical precision
demonstrated here) sometime after the formation of these CAIs
but before the formation of meteoritic metal in the parent bodies
of both undifferentiated (i.e., chondrites) and differentiated (i.e.,
iron meteorites and pallasites) meteorites. Alternatively, the
homogeneous isotopic composition of meteoritic metal relative
to CAIs could represent spatial rather than temporal differences
in the solar nebula, and the 60Fe-60Ni chronometer does not
necessarily constrain the timing of a homogenization event, if it,
in fact, took place. Furthermore, the formation of meteoritic metal
may have acted to homogenize the precursor material to the
degree that anomalies were not preserved at the parent-body scale
sampled by the metal. In the following discussions, it will be
assumed that any resolvable 60 deviations in meteoritic metal are
solely the result of the decay of 60Fe. This does not preclude the
possibility of the presence of nucleosynthetic anomalies in Ni
isotopes that may be unresolved with our current analytical
precision.
Effects in 60Ni/58Ni Ratios from 60Fe Decay in Meteoritic
Metal: The Expectation. Figure 6 shows the predicted values
of 60 in meteoritic metal if one assumes that the Fe/Ni fractionation event that resulted in the formation of this metal occurred
within the first ∼6 My of solar system history. Different symbols
depict metals with different 56Fe/58Ni ratios, and the curves
represent the values of 60 that would be measured today in
samples that formed from a chondritic reservoir at the time
specified on the x axis, where 0 My marks the beginning of the
solar system. For these calculations, the solar system initial 60Fe/56Fe ratio was assumed to be 1.0 × 10-6,7-8 which corresponds
to an initial 60 ) -0.63. Moving from 0 to 6 My along the x axis,
the 60Fe/56Fe ratio decreases, and the 60 value increases in the
reservoir from which the metal forms. If the solar system initial
60Fe/56Fe ratio was as high as 4.4 × 10-6,40 the calculated 60
trajectories in Figure 6 would be shifted toward more negative
(38) Shukolyukov, A.; Lugmair, G. W.; Bogdanovski, O. Lunar Planet. Sci. 2003,
34, no. 1279.
(39) Bizzarro, M.; Ulfbeck, D.; Thrane, K. Lunar Planet. Sci. 2006, 37, no. 2020.
8482 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
Figure 6. Predicted values of 60 in meteoritic metal that formed in
the early solar system. Formation from a chondritic reservoir with a
solar system initial 60Fe/56Fe ) 1.0 × 10-6 and 60 ) -0.63 is
assumed. The time of formation is given on the x axis, where a time
of 0 My corresponds to the beginning of the solar system. The 60
values expected for metals with the following 56Fe/58Ni ratios are
shown by the various curves: 5 (squares), 10 (circles), 15 (triangles),
and 20 (diamonds). The external precision (2σ) for our 60 measurements is shown by the two dashed lines ((0.15 ).
values. As shown in Figure 6, deficits in 60 are expected in early
formed objects with subchondritic 56Fe/58Ni ratios. Metals in iron
meteorites,41 pallasites,42 and chondrites43,44 typically have Ni
concentrations g5.5 wt %, with the maximum 56Fe/58Ni ratio in
these samples being nearly identical to the chondritic value
(24.4).45 Therefore, Figure 6 demonstrates that if 60Fe was live in
the early solar system, early formed meteoritic metal with a
subchondritic 56Fe/58Ni ratio is expected to record deficits in 60
unless it underwent subsequent equilibration. Meteoritic metal
is not expected to show excesses in 60, regardless of its time of
formation. As discussed in more detail in the following sections,
our data and those of Quitté et al.21 are fully consistent with this
expectation.
Effects in 60Ni/58Ni Ratios from 60Fe Decay in Meteoritic
Metal: The Observation. Figure 7 shows the 60 values measured in meteoritic metal. As is evident from Table 2 and Figure
7, no resolvable excesses in 60 were detected. In particular, no
such excesses were observed in the metal phase of Casas
Grandes, Canyon Diablo, and Forest Vale, contrary to a previous
report.9 Furthermore, an 60 value of ≈ +1 had been previously
measured in Toluca metal;46 therefore, two separate aliquots from
a single digestion were obtained from Lyon for interlaboratory
comparison. One aliquot had been chemically processed in Lyon
for Ni separation, and the other was unprocessed; the latter was
processed at the Field Museum, and both aliquots were analyzed.
Neither aliquot showed excesses in 60, and the datum for Toluca
in Figure 7 represents the mean of all the measurements made
on both aliquots. An independent study of the Ni isotopic
(40) Quitté, G.; Latkoczy, C.; Halliday, A. N.; Schönbächler, M; Günther D. Lunar
Planet Sci. 2005, 36, no. 1827.
(41) Scott, E. R. D. Geochim. Cosmochim. Acta 1972, 36, 1205-1236.
(42) Wasson, J. T.; Choi, B.-G. Geochim. Cosmochim. Acta 2003, 67, 3079-3096.
(43) Rambaldi, E. Earth Planet Sci. Lett. 1976, 31, 224-238.
(44) Rambaldi, E. R. Earth Planet Sci. Lett. 1977, 36, 347-358.
(45) Anders, E.; Grevesse, N. Geochim. Cosmochim. Acta 1989, 53, 197-214.
(46) Moynier, F. Personal communication; 2005, Ecole Normale Supérieure de
Lyon.
Figure 7. 60 for metal from nonmagmatic irons (half-filled circles),
magmatic irons (open circles), pallasites (triangles), and chondrites
(half-filled squares). Plotted errors are 2σm; the external precision (2σ)
is shown by the two dashed lines ((0.15 ).
composition of the metal from 33 iron meteorites (including
Canyon Diablo and Toluca) found no resolvable non-massdependent effects in the 60Ni/58Ni ratio in these samples.21
Bizzarro et al.39 recently reported that the metal from magmatic
and nonmagmatic iron meteorites is characterized by deficits in
60 ranging from -0.11 ( 0.07 to -0.28 ( 0.11 (indistinguishable
within their analytical uncertainties). Our analyses of iron meteorites yield 60 values that are mostly zero within 2σm uncertainties
(some as small as (0.06 ), and there is no clear indication of a
uniform deficit in 60 of ≈-0.2. Nevertheless, it is possible that
some may record slightly negative 60 values that are not
resolvable given our current external precision on 60 of (0.15.
In addition, there is one iron meteorite sample (Hoba) that does
appear to have a resolvable deficit in 60 (Table 2; Figure 7).
No resolvable effects in 60 were detected in the metal from
pallasites. However, clearly resolvable deficits in 60 were measured in the metal from the least equilibrated chondrites belonging
to the LL3.0 (Semarkona), CR (Renazzo), and CBa (Gujba) classes,
whereas the metal from the more equilibrated chondrites belonging to the LL3.1 (Bishunpur), H4 (Forest Vale), and EH4 (Indarch)
classes are characterized by normal (i.e., terrestrial) 60 values.
Resolvable E60 Deficits in Primitive Chondrite Metals:
Implications for the Initial 60Fe/56Fe in the Early Solar
System. As discussed above, resolvable deficits in 60 were
measured in the metal from the least equilibrated chondrites
studied (Semarkona, Renazzo, and Gujba). These 60 deficits
indicate that the fractionation events that established the Fe/Ni
ratios in these metals occurred while 60Fe was extant in the early
solar system. Assuming a chondritic initial source reservoir (i.e.,
characterized by a normal 60 value and a 56Fe/58Ni ratio of 24.445),
the 60Fe/56Fe ratios at the time of these fractionation events may
be estimated on the basis of a single-stage model calculation using
the measured 60 deficits (considered with the 3σm error or the
2σ external precision, whichever is larger) and the Fe/Ni ratios
in these metals. Such a calculation yields 60Fe/56Fe ratios of (1.0
( 0.6) × 10-6, (3.4 ( 2.8) × 10-6, and (3.5 ( 3.0) × 10-6 for metal
samples from Semarkona, Renazzo, and Gujba, respectively.
Within uncertainties, these values are indistinguishable and yield
a weighted mean of (1.2 ( 0.6) × 10-6 for the 60Fe/56Fe ratio at
the time of the Fe/Ni fractionation event for metal in these
Figure 8. A plot of 60 vs 56Fe/58Ni for metal from magmatic irons
and pallasites. The plotted errors are either the 2σm errors or the 2σ
external precision ((0.15 ), whichever is larger. The chondritic point
(square) is shown for reference only. The slope of the best-fit line
corresponds to an 56Fe/60Fe ratio of (4.6 ( 3.3) × 10-7. Symbols are
the same as in Figure 6. Note that the error bars on the 56Fe/58Ni
ratios are smaller than the symbols.
primitive chondrite samples. Assuming that the 60 deficits in metal
from Semarkona, Renazzo, and Gujba are solely from the decay
of 60Fe, this value must be treated strictly as a lower limit on the
initial 60Fe/56Fe ratio for the solar system. The components,
including metal grains, in the least equilibrated primitive chondrites are among the most pristine materials in the early solar
system and record nebular processes.22,47 Although this is almost
certainly the case for Semarkona and Renazzo, it may be argued
that the Fe/Ni ratio in the Gujba metal was established ∼5 My
or so after the formation of the refractory calcium aluminum-rich
inclusions (CAIs, believed to be among the first solids to form in
the solar nebula),48 during an impact event that formed the silicate
chondrules in this meteorite.49 However, the fact that the measured 60 deficits (and Fe/Ni ratios) in Gujba metal are similar to
those in Renazzo metal may be an indicator that the Fe/Ni ratio
in the former was inherited from precursor material whose Fe/
Ni ratio was established earlier in the solar nebula. If, indeed,
the Fe/Ni ratios in these metals were established early in the
history of the solar nebula (i.e., almost contemporaneously with
CAI formation), the value of (1.2 ( 0.6) × 10-6 is likely to be
close to the solar system initial 60Fe/56Fe ratio. This value falls in
the range of those proposed for the solar system initial 60Fe/56Fe
by other recent studies.7-8 Moreover, as also suggested by these
studies, this inferred initial 60Fe/56Fe ratio of ∼1 × 10-6 provides
support for the hypothesis that this radionuclide was injected into
the solar nebula from a stellar source (most likely a Type II
supernova) and is likely to have served as an important heat
source for planetesimal differentiation.
Nickel Isotope Systematics in Magmatic Iron Meteorites
and Pallasites: Implications for Timing of Fe/Ni Fractionation on Differentiated Parent Bodies. Figure 8 shows a plot
of 60 vs 56Fe/58Ni for metal from differentiated parent bodies (i.e.,
magmatic iron meteorites and pallasites). The slope of the best(47) Krot, A. N.; Meibom, A.; Weisberg, M. K.; Keil, K. Meteorit. Planet. Sci.
2002, 37, 1451-1490.
(48) Amelin, Y.; Krot, A. N.; Hutcheon, I. D.; Ulyanov, A. A. Science 2002, 297,
1678-1683.
(49) Krot, A. N.; Amelin, Y.; Cassen, P.; Meibom, A. Nature 2005, 436, 989992.
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
8483
fit line through these data yields an initial 60Fe/56Fe ratio of (4.6
( 3.3) × 10-7 and does not support the previously reported10 value
of (3.0 ( 0.2) × 10-6 that is based on metal from iron meteorites,
pallasites, chondrites, and mesosiderites.
There are three main, sequential processes that affected the
compositions of planetary cores: condensation and accretion of
the parent body in the nebula, metal-silicate separation on the
parent body during core formation, and crystallization of the
metallic core.41 The chemical differences between planetary cores
(i.e., represented by different groups of the magmatic irons) are
attributed to a combination of condensation, accretion, and core
formation processes, whereas variations among different samples
from the same group are attributed to core crystallization. All of
these processes likely contributed to the full range in Fe/Ni ratios
shown in Figure 8. However, the bulk Fe/Ni compositions of
planetary cores were probably established during metal-silicate
separation on their parent bodies. During this process, Fe may
partition into silicates, sulfides, metal, or all three, whereas Ni
partitions strongly into the metal phase.41 The final Fe/Ni ratio
of a particular sample within a magmatic iron meteorite group
will depend on its position in the crystallization sequence, since
the solid metal becomes progressively enriched in Ni as crystallization proceeds. Thus, the Fe/Ni compositions of magmatic iron
meteorites and pallasites predominantly result from parent body,
rather than nebular, processes. Furthermore, the total variation
in Fe/Ni ratios within many magmatic groups is ∼20%,41 whereas
the Fe/Ni ratios shown in Figure 8 vary by a factor of 3.3. As
such, most of this variation is controlled by differences among
groups, rather than within groups. Therefore, the overall variation
in Fe/Ni shown in Figure 8 is likely controlled primarily by the
Fe/Ni fractionation during core formation on different parent
bodies. The slope of the Fe-Ni isochron shown in Figure 8 is
meaningful only if the fractionation events that established most
of the variation in Fe/Ni ratios occurred approximately contem-
8484
Analytical Chemistry, Vol. 78, No. 24, December 15, 2006
poraneously (at least within the precision of the 60Fe/60Ni
chronometer). If this is so, and if one assumes a solar system
initial 60Fe/56Fe ratio of ∼1 × 10-6 (see discussion in the previous
section), the Fe/Ni fractionation recorded in metal from the
+2.7
magmatic iron and pallasite meteorites occurred at 1.7-1.2
My
after CAI formation.
Previous studies of 182Hf-182W systematics in the iron meteorites indicate that metal-silicate segregation on the parent bodies
of the various magmatic iron meteorite groups occurred within a
narrow time interval, at most within a few million years of CAI
formation.15-17 The 60Fe-60Ni systematics in the magmatic iron
and pallasite meteorites studied here (including the timing of the
major Fe/Ni fractionation events implied by the isochron in Figure
8) are consistent with these 182Hf-182W systematics.
ACKNOWLEDGMENT
We thank Laure Dussubieux for assistance with the quadrupole
ICPMS analyses, F. Moynier and F. Albarède (Ecole Normale
Supérieure de Lyon) for providing aliquots from their solutions
of Toluca, the Smithsonian Institution for providing the sample
of Semarkona (USNM 1805), and helpful comments by two
anonymous reviewers. This work was supported in part by the
National Aeronautics and Space Administration through grants
to MW, RNC, and AMD.
SUPPORTING INFORMATION AVAILABLE
Figures S-1-S-4 (two pages) are available as Supporting
Information. This material is available free of charge via the
Internet at http://pubs.acs.org.
Received for review July 16, 2006. Accepted September
28, 2006.
AC061285M
S-1
Authors and Affliations
David L. Cook1,2,3, Meenakshi Wadhwa1,2,3, Philip E. Janney3, Nicolas Dauphas1,2,4, Robert N.
Clayton1,2,4, and Andrew M. Davis1,2,4
1
Department of the Geophysical Sciences, The University of Chicago, 5734 S. Ellis Ave.,
Chicago, IL, 60637. 2Chicago Center for Cosmochemistry, 5640 S. Ellis Ave., Chicago, IL,
60637. 3Department of Geology, The Field Museum, 1400 S. Lake Shore Dr., Chicago, IL,
60605. 4Enrico Fermi Institute, 5640 S. Ellis Ave., Chicago, IL 60637.
Title
High Precision Measurements of Non-Mass Dependent Effects in Nickel Isotopes in Meteoritic
Metal via Multi-Collector ICPMS
Contents
Figure S-1.
Figure S-2.
Figure S-3.
Figure S-4.
4
3
2
1
ε
61
0
-1
-2
-3
-4
0
5
10 15 20 25 30 35 40
analysis number
Figure S-1. ε61 values for repeated analyses of an Aesar Ni solution over the course of a 24month period. Each datum represents the mean of 5 repeat measurements performed during a
single analysis session. The individual error bars are 2σm errors, based on the 5 repeat
measurements for each datum. The external precision is the standard deviation (2σ) based on all
of the data plotted here, and is shown by the two dashed lines (± 0.85 ε).
6
4
2
ε
64
0
-2
-4
-6
0
5
10 15 20 25 30 35 40
analysis number
Figure S-2. ε64 values for repeated analyses of an Aesar Ni solution over the course of a 24month period. Each datum represents the mean of 5 repeat measurements performed during a
single analysis session. The individual error bars are 2σm errors, based on the 5 repeat
measurements for each datum. The external precision is the standard deviation (2σ) based on all
of the data plotted here, and is shown by the two dashed lines (± 1.5 ε).
-4
-3
-2
-1
0
1
2
3
4
-4
-3
-2
-1
0
1
61
2
3
4
ε
Figure S-3. ε61 for meteoritic metal. Symbols and samples same as Figure 7. Plotted errors are
2σm; the external precision (2σ) is shown by the two dashed lines (± 0.85 ε).
-6
-4
-2
0
2
4
6
-6
-4
-2
0
64
2
4
6
ε
Figure S-4. ε64 for meteoritic metal. Symbols and samples same as Figure 7. Plotted errors are
2σm; the external precision (2σ) is shown by the two dashed lines (± 1.5 ε).