Unlocking the Zinc Isotope Systematics of Iron Meteorites

Supplementary Text –

Unlocking the Zinc Isotope Systematics of Iron Meteorites

L.J. Bridgestock1, H. Williams2,3, M. Rehkämper1,4 *, F. Larner1,2, M.D. Giscard1,4, S. Hammond5 , B. Coles1, R. Andreasen1, B.J. Wood2, K.J. Theis6, C. Smith4, G.K. Benedix7, M. Schönbächler8

1Department of Earth Science and Engineering, Imperial College London, SW7 2AZ, UK

2Department of Earth Sciences, University of Oxford, OX1 3AN, UK

3Department of Earth Sciences, Durham University, DH1 3LE, UK

4Natural History Museum, London, SW7 5BD, UK

5Centre for Earth, Planetary, Space & Astronomical Research, The Open University, Milton Keynes MK7 6BJ, UK

6Dept. School of Earth, Atmospheric and Environmental Sciences, University of Manchester, M13 9PL, UK

7Department of Applied Geology, Curtin University, Bentley, Western Australia, 6845, Australia

8Institut für Geochemie und Petrologie, ETH Zürich, Clausiusstrasse 25, 8092 Zürich, Switzerland

*Corresponding author: ; +44 20 7594 6391

Detailed information on analytical techniques

General materials and procedures

Following physical separation, the samples were handled in Class 10 laminar flow hoods within Class 1000 clean rooms either at the MAGIC Laboratories, Imperial College London (leaching, digestion of meteorite samples and preparation of all samples for isotopic analyses), or in the geochemical laboratories of the Department of Earth Sciences, Durham University (digestion of samples from experiments).

At the MAGIC Labs, the mineral acids 6 M HCl and 15.4 M HNO3 were purified from AnalaR grade acids by sub-boiling distillation, whilst high-purity (distilled-quality) 8 M HBr and 28 M HF were purchased from VWR (UK). Dilutions were made using 18.2 MΩ cm grade H2O and aqua regia was prepared as a 3 + 1 mixture of 6 M HCl + 15.4 M HNO3.

Leaching and digestion of samples

To remove surface contamination, all meteorite samples were sequentially leached at room temperature with (i) 0.5 M HBr (1 hour, in ultrasonic bath), (ii) 6 M HCl (1 hour, in ultrasonic bath) and (iii) aqua regia (30 min) with intermittent rinses in 18 MΩ cm H2O. The samples were then washed in ethanol, dried and weighed. Approximately 20 to 30% of the original meteorite metal was removed by leaching. Following leaching, the samples were digested in aqua regia at 80°C on a hotplate.

The digestion of the sulfide nodules was incomplete and insoluble residues were removed by centrifugation. Troilite, and schreibersite for TS3, were observed to dissolve in this first digestion step. The insoluble residue of TS3, comprising silicates and chromite, was weighed after drying and then dissolved in a second step to yield sample TS3-D2. This involved (i) treatment with a 3 + 1 mixture of 28 M HF + 15.4 M HNO3 at 80 °C on a hotplate for 1 week, (ii) evaporation to dryness and digestion in 6 M HCl at 80 °C on a hotplate for 2 days, and (iii) evaporation to dryness and digestion in aqua regia in a Parr bomb at 120°C for 2 days. Finally, all samples were dissolved in 6 M HCl for aliquoting.

The separated silicate glass and metal fractions from the high-pressure experiments were also cleaned prior to digestion, by treating with ethanol in an ultrasonic bath (20 min; to remove residual organics) followed by 18 MΩ cm H2O (20 min). The samples were then digested using an 8 + 1 mixture of 28 M HF + 15.4 M HNO3 (silicates) or aqua regia (metal) at 130°C and finally dissolved in 6 M HCl for aliquoting.

Aliquoting of sample solutions

Small (typically ~5%) aliquots of the 6 M HCl sample solutions were used for preliminary Zn concentration measurements (to enable optimal application of the Zn double spike), whilst larger aliquots were employed for subsequent double spike Zn isotope analyses. For the IIAB and IIIAB irons, the latter measurements used up all available sample material, whilst solution remained for IAB irons. Aliquots of the remaining IAB iron meteorite solution were employed to determine the concentrations of Cr and other metals by ICP-MS at the Open University (see supplementary Table S3.

Double spiking for Zn isotope analyses

A preliminary determination of the Zn concentrations was carried out using an abbreviated version of the protocol that was applied for the isotope analyses. Appropriate aliquots of the sample solutions were then mixed with a 64Zn – 67Zn double spike, characterized by 64Zn/67Zn ≈ 2.5 (see Arnold et al., 2010), and refluxed for at least 12 hours on a hotplate to ensure spike-spike sample equilibration.

For most samples, the spike addition was calculated to provide a molar ratio of spike Zn to natural Zn (ZnSp/ZnNat) of about 1, to optimize the precision and accuracy of the analyses (Arnold et al., 2010). Test measurements that were carried out for spiked Zn standard solutions with ZnSp/ZnNat ratios of about 0.4 to 1.7 returned δ66Zn values that were indistinguishable, within uncertainty (±0.05 to ±0.10‰; 2sd), to data obtained for solutions with ZnSp/ZnNat = 1. Based on this, all samples were analyzed relative to ‘matching standards’ that featured similar ZnSp/ZnNat ratios and similar Zn concentrations. These matching standard solutions were prepared as mixtures of the MAGIC ‘London Zn’ in-house standard solution (Arnold et al., 2010) with the Zn double spike.

Standard applications of the double spike protocol for stable isotope analyses (e.g., Siebert et al., 2001) assume that that isotopic variations other mass dependent isotope effects are absent. In principle, this requirement can be problematic for meteorites because numerous elements feature isotopic anomalies of radiogenic, nucleosynthetic or cosmogenic origin in such samples (Rehkämper et al., 2012). However, previous isotopic studies, which did not apply a double spike for mass bias correction, found no evidence for non-mass dependent Zn isotope variations between terrestrial materials, lunar rocks and soils, and various meteorites, including chondrites, achondrites, pallasites as well as IAB complex and IIIAB irons (Luck et al., 2005; Herzog et al., 2009; Moynier et al., 2007; Moynier et al., 2011; Ghidan and Loss, 2011; Barrat et al., 2012; Paniello et al., 2012). The only exception is the unconfirmed finding of Ghidan and Loss (2011), who reported a non-mass dependent Zn isotope effect of about 10‰ for the 68Zn/64Zn ratio of the ungrouped iron Redfields.

Separation of Zn for isotope analyses

Following addition of the double spike, the 6 M HCl sample solutions were evaporated to dryness, and redissolved in 1 M HCl for separation of Zn from the sample matrix by anion exchange chromatography. The separation was carried out following a procedure modified from Arnold et al. (2010), using small Teflon columns filled with 200 µl of AG MP1 100-200 mesh resin. After loading of the 1 M HCl sample solutions, matrix elements were rinsed from the resin with 1 M HCl (8 ml) and Zn was subsequently eluted with 0.01 M HCl (6 ml). The Zn fractions were evaporated to dryness, re-dissolved in 1 M HCl (1 ml) and processed through the same anion exchange chemistry for a second time. The dried residues of the Zn fractions were then twice taken up in 15.4 M HNO3 (100 µl) and evaporated to dryness, to convert to nitrate form and oxidize residual organics.

Double spike Zn isotope analyses by MC-ICP-MS

The isotopic analyses were carried out with a Nu Plasma HR MC-ICP-MS instrument at the MAGIC Laboratories using protocols modified from Arnold et al. (2010). Sample introduction utilized either an Aridus (CETAC Technologies) or a DSN-100 (Nu Instruments) desolvation system fitted with nominal 100 µl/min glass nebulizers.

The analyses involved simultaneous collection of the ion beams of 64Zn+, 66Zn+, 67Zn+, 68Zn+ and 62Ni+, 137Ba2+ (for interference correction) using Faraday cups equipped with 1011 Ω resistors. Data collection was performed in 3 blocks of 20 x 5 s integrations, with a 15 s electronic background measurement preceding each block. All samples and standards were prepared for analysis by dissolution in an appropriate volume of 0.1 M HNO3, and nitric acid wash solutions were of the same molarity. The runs typically employed solutions with Zn concentrations of 40 to 60 ng/ml in 0.1 M HNO3, and instrumental sensitivities of better than 110 V (µg/ml)-1 were generally attained for Zn.

The instrumental mass bias encountered during the Zn isotope analyses was corrected using the double spike technique with methods adapted from Arnold et al. (2010). For this, the “raw” measured isotope ratios of 64Zn/67Zn, 66Zn/67Zn and 68Zn/67Zn were processed offline with a spreadsheet-based implementation of a data reduction technique originally developed by Siebert et al., (2001). Corrections for spectral interferences from 64Ni+ and Ba2+ ions (132Ba2+, 134Ba2+ and 136Ba2+) were also subjected to mass bias correction. The required corrections, however, were very small in all cases and typically at levels of 64Ni+/64Zn+ and 134Ba2+/67Zn+ of about 1 x 10-5 or less.

The Zn isotope compositions of the samples were determined relative to previously prepared matching standard solutions (see above) and differences in the 66Zn/64Zn ratio (R) between samples and standards are reported as δ66Zn:

δ66Zn = [(RSample/RStandard) – 1] x 1,000

In the manuscript, all δ66Zn values are reported relative to the widely adopted (but no longer available) JMC Lyon Zn reference solution (Marèchal et al., 1999). To this end, the results were corrected for an isotopic offset of δ66Zn = +0.12 ± 0.04‰ for London Zn relative to Lyon Zn (Larner and Rehkämper, 2012; Arnold et al., 2010). The reported Zn concentrations of the samples were determined by isotope dilution using the 67Zn/68Zn ratios measured in the double spike runs.

The Zn concentration and isotope data reported in Tables 1, 2 (main manuscript) and the associated uncertainties take into account the laboratory blank and the precision of the mass spectrometric data. Blank corrections and the calculation of uncertainties are further detailed below and in supplementary Table S1.

Blank corrections for Zn

Four analyses yielded a total procedural Zn blank of 5 ± 2 ng (total range). The Zn concentrations and δ66Zn values of all meteorite samples were corrected for this blank contribution. The latter correction assumed a ‘normal’ terrestrial isotope composition of δ66Zn = 0.25 ± 0.25‰ for the contamination (Cloquet et al, 2008; Larner and Rehkämper, 2012).

The silicate and metal samples from the partitioning experiments featured sufficient Zn (>150 µg) such that blank corrections were negligible and hence not applied to the Zn isotope and concentration data.

Uncertainties of the Zn isotope and concentration data

Uncertainties for data of meteorite samples

For the Zn concentrations, the total uncertainty u-Zn was calculated by determining the effect on the calculated Zn abundances that is introduced by (i) the uncertainty of the mass spectrometric runs (u-ms) and (ii) the uncertainty of the blank mass (u-blm), such that:

u-Zn2 = u-ms2 + u-blm2

For the Zn isotope data, the total uncertainty (u-δ66Zn) was obtained by determining the effect on the δ66Zn results, which is introduced by (i) the precision of the mass spectrometric analyses (u-ms), (ii) the uncertainty of the blank mass (u-blm), and (iii) the uncertainty in the isotope composition of the Zn blank (u-bli), whereby:

u-δ66Zn2 = u-ms2 + u-blm2 + u-bli2

The detailed uncertainty calculations are shown in supplementary Table S1. The calculations utilize a total procedural blank mass and isotope composition of 5 ± 2 ng and δ66Zn = 0.25 ± 0.25‰, respectively.

The typical within-run (internal) precision of a single δ66Zn measurement, which utilized ~25 ng of natural Zn, was about ±0.05‰ (2se, standard error of the mean). In comparison, the typical reproducibility of δ66Zn values determined from replicate analyses of a London Zn – Zn double spike mixture (each consuming ~25 ng natural Zn) was ±0.05‰ to ±0.20‰ and better than ±0.15‰ in most cases (2sd; Table S1). The larger within-session reproducibility was generally taken as the mass spectrometric uncertainty for the δ66Zn values. Exceptions are samples TS1 and TS2, which were inadvertently overspiked to ZnSp/ZnNat ≈ 7 to 11 and analyzed at low 66Zn+, 68Zn+ ion bean intensities. The Zn isotope analyses of these samples were carried out relative to London Zn – Zn double spike mixtures with similar ZnSp/ZnNat ratios and the within-run precision of the measurements was applied to characterize the mass spectrometric uncertainty of the δ66Zn data, as this was larger than the within-session reproducibility. For the Zn concentration data, the mass spectrometric uncertainties were calculated by propagating the within-run precision (±2sd) of the 67Zn/68Zn ratios determined for the spike-sample mixtures through the isotope dilution equations.

The quoted mass spectrometric uncertainties are justified by results obtained for replicate analyses of (i) Toluca 1, conducted within the same measurement session and (ii) Canyon Diablo 1 and 2, carried out on different days with different sample introduction systems. In all cases, the analyses yielded δ66Zn values (and Zn concentrations) that are identical within the reproducibility (Table 1, main manuscript).

The total uncertainties of the δ66Zn data are dominated by the mass spectrometric reproducibility for the majority of the samples (Table S1). The blank corrections were small or insignificant in this case because the samples featured more than 200 ng of natural Zn, equivalent to >96% natural sample Zn. Uncertainties stemming from the δ66Zn blank corrections are only important for the Zn-poor Toluca nodules TS1, TS2 and the Zn-depleted IIAB irons (Table S1), which featured mass fractions of indigenous natural Zn that were as low as ~60% (Table 1, main manuscript). The uncertainties of the δ66Zn blank corrections are, furthermore, dominated by the blank mass component (u-blm) for all meteorites but the IIIAB irons (Table S1). As a consequence, the total uncertainties calculated for δ66Zn would not change significantly, even if the isotope composition of the blank is assigned a larger uncertainty of ±0.50‰ (Table S1).