The I-Xe Chronometer and Its Constraints on the Accretion and Evolution Ofplanetesimals

The I-Xe Chronometer and its Constraints on the Accretion and Evolution ofPlanetesimals.

J. D. Gilmour and S. A. Crowther

School of Earth, Atmospheric and Environmental Science, University of Manchester, Manchester M13 9PL, United Kingdom.Tel 44 161 275 3681Fax 44 161 306 9361Email

Keywords: extinct radioisotopes, iodine-xenon, chondrites, early solar system, chronology.

Abstract

We report a -1.6±2.6 Myr (1 σ error) I-Xe age of the ungrouped achondrite NWA 7325 relative to the Shallowater standard. We re-evaluate the calibration of the relative I-Xe dating system against the absolute Pb-Pb chronometer in the light of this and other recently reported analyses, and taking into account revisions to the Pb-Pb system, deriving a new absolute age for the Shallowater standard of 4562.7±0.3 Ma. With this calibration, the oldest chondrule I-Xe ages overlap the oldest Pb-Pb chondrule ages and the Pb-Pb ages of Calcium Aluminium-rich inclusions.

Literature datafor large aliquots of equilibrated ordinary chondrites suggest iodine loss during metamorphic processing and show some evidence that bulk 129Xe*/I ratios decreasewith increasing petrologic type. However, the range of ratios at each petrologic type suggests that thermal evolution was affected by changes of thermal insulation with time, perhaps by impact processing of the parent planetesimals. Literature I-Xe ages for chondrules from Bjurböle (L/LL4) and pyroxene from Richardton (H5) suggest closure shortly after the peak of metamorphism, consistent with a high closure temperature in mafic minerals. The extended range of ages reported for chondrules from the LL3.4 chondrite Chainpur is interpreted as a product of collisional processing of material near the surface of the parent body, and may record a decline in the rate of collisions in the asteroid belt over the first 100 Myr of solar system history.

Introduction

Monoisotopic anomalies of 129Xe were first detected in primitive meteorites in 1960 (Reynolds, 1960). They were proposed to be the decay product of short-lived 129I.This was the first evidence of a short-lived radioisotope in the early solar system, and represented the birth of a new field: dating early solar system events based on the decay of extinct, short-lived radioisotopes.

The origin of the anomalies was confirmed by a correlation between excesses of 129Xe and 128Xe that had been produced from stable 127I by artificial neutron irradiation in a reactor (Jeffrey and Reynolds, 1961). The I-Xe dating scheme emerged in close to its current form in Fish and Goles(1962). It was based on sample-to-sample differences in the 129I/127I ratios on closure to xenon loss as measured by the 129Xe*/128Xe* (the ratio of radiogenic 129Xe to 128Xe produced by artificial neutron irradiation).

The I-Xe dating scheme was initially applied in step heating studies of whole rock samples of chondritic meteorites, with results (in the form of isochrones over several high-temperature releases) that were found puzzling. With improved technology it became possible to establish closure ages of individual components of chondritic meteorites and, eventually, to calibrate the chronometer against the Pb-Pb scheme. However, it remains under-utilised. In this paper we review the development of the scheme and the status of the calibration against the Pb-Pb timescale, summarise the constraints it imposes on the evolution of chondritic planetesimals, and suggest how it might be further developed.

Methods

At its simplest, the I-Xe system involves subjecting aliquots of a set of samples to the same neutron fluence, typically ~1019 n cm-2. Atoms of 127I in the samples capture a neutron to form 128I, which then decays to 128Xe with a half life of 25minutes. After decay of 128I is complete, each aliquot is analysed by step pyrolysis – it is heated to each of a series of increasing temperatures, and the gas evolved at each temperature is measured by noble gas mass spectrometry. Neutron irradiation also produces131Xe from 130Ba and 130Te, and 131,132,134,136Xe (4 isotopes) from neutron induced fission of 235U; these reactions provide information on the chemistry of sites releasing xenon at each temperature step (e.g. Claydon et al., 2015). After corrections for fission or spallation components, if present, a correlation is sought between 129Xe and 128Xe by examining ratios among these and an isotope not produced from iodine, most often 132Xe. From such correlations, the 129Xe*/128Xe* ratio can be calculated. Correlations tend to be observed in high temperature releases; lower temperature steps often contain iodine without its full complement of 129Xe*, perhaps indicating terrestrial contamination. This can be particularly apparent in analyses of finds.

Since the conversion factor between 127I and 128Xe depends only on the neutron fluence, the 129Xe*/128Xe* ratios are related to 129Xe*/127I ratios by a constant. Thus for sample 1 and sample 2

(1)

Where Δt12 is the interval between closure of the I-Xe system in sample 1 and in sample 2, λ is the decay constant of 129I,and the subscripts to the parentheses indicate that the ratios are those measured in the two different samples. In this work it has been assumed that the half life of 129I is exactly 16.1 Myrbased on a reported half life of 16.1±0.07 Myr (Chechev and Sergeev, 2004). The error on the half life introduces a systematic error to all I-Xe ages, and so should not be included when reporting the age of a particular sample. The decay constant is then0.0431 Myr-1 to three significant figures. Other half lives are used in the literature and where necessary a conversion has been applied.

In extracting the 129Xe*/128Xe*, most commonly a correlation is sought between 129Xe/132Xe and 128Xe/132Xe over a series of consecutive steps. In this case the desired 129Xe*/128Xe* ratio is the gradient of the correlation line. Alternatively, a correlation is sought between 132Xe/129Xe and 128Xe/129Xe or 128Xe*/129Xe. 128Xe* for any release can be calculated over an assumed underlying trapped composition, such as Q-Xe (Busemann et al., 2000), using the normalising isotope. (In practice corrections for spallation or fission products may need to be made first.) In this case the desired ratio is the reciprocal of the value of 128Xe/129Xe (or, equivalently, 128Xe*/129Xe)where 132Xe/129Xe = 0on the correlation line; this is informally referred to as the “inverse isochron” approach, though both correlation lines are isochrones. An illustration of the two approaches applied to a common dataset can be found in the review by Gilmour et al. (2006). In either case, the correlation line should be extracted with an appropriate algorithmthat takes into account the covariance between the input ratios for each data point introduced by the common denominator: a York fit (York, 1969) or a maximum likelihood method (Titterington and Halliday, 1979) such as minimization of the sum of the squares of the Mahalanobis distances of the data from the line(Gilmour, 2015). If all the uncertainties were normally distributed, the 129Xe*/128Xe* ratios and associated errors derived by the two different approaches would agree within the rounding error of the calculation since they are based on the same underlying data. When this is not the case, it is necessary to consider which approach best reports the uncertainty of the measurement

Differences between the two approaches can arise when there are releases that contribute to an isochron thathave 132Xe contributions within error of the blank; that is to say, when there is no compelling evidence of a contribution to 132Xe from the sample. In these circumstances the conventional isochron approach is invalid. Fitting algorithms are predicated on a normal distribution of the uncertainties around the central value. This is only approximately true for isotope ratios, which are the ratios of quantities that are themselves Poisson distributions approximated by normal distributions. The approximation that isotope ratios are normally distributed breaks down when there is a significant probability that the denominator isotope was not present (Gilmour, 2015). It should also be noted that error bars based on propagated errors do not accurately represent confidence intervals in ratios where the denominator is within error of zero.

Equation (1) allows the determinations of relative closure intervals for a set of samples that each had aliquots included in the same irradiation. In order to compare closure intervals across different irradiations, it is necessary to monitor the conversion efficiency from 127I to 128Xe in each irradiation. This is achieved by including one or more aliquots of a standard. Ideally, a standard should have a consistent, reproducible 129Xe*/I ratio that can be determined with high precision from an isochron; this requires a high I/Xe ratio and formation in the first ~10 Myr of solar system history. It should also be available in sufficient quantities to allow consumption through its use in many irradiations. The first standard widely used as an irradiation monitor was whole rock from the L/LL4 chondrite Bjurböle (Hohenberg and Kennedy, 1981). This has now been replaced by aliquots of enstatite from the anomalous aubrite Shallowater, which appears to be more reproducible than Bjurböle when small (< 1 mg) quantities are used (Gilmour et al., 2006). For this reason, closure of the system in Shallowater enstatite is adopted as the zero of the I-Xe relative chronometer; I-Xe ages relative to Shallowater enstatite are the end product of any analytical campaign. Brazzle et al (1999) measured the closure age of Bjurböle whole rock as 0.47±0.15 Myr relative to closure of Shallowater (positive values indicate earlier closure throughout, so Bjurböle closed earlier than Shallowater), allowing literature ages determined relative to Bjurböle to be directly compared to literature ages determined relative to Shallowater.

Calibrating the I-Xe System.

The procedure described in the previous section is sufficient to allow an extensive database of relative closure ages for early solar system materials in the I-Xe system. This is supplemented by two separate and unrelated forms of absolute calibration: determination of the actual 129Xe*/I ratio in a sample today, which is equivalent to the 129I/127I ratio on closure; determination of absolute closure ages by calibration against the Pb-Pb chronometer.

The 129I/127I ratio of the early solar system.

Determining absolute 129Xe/127I ratios requires independent knowledge of the conversion efficiency of 127I to 128Xe. This is best achieved by irradiating a known amount of iodine and measuring the amount of 128Xe produced. The iodine in samples used for I-Xe dating is typically present in trace concentrations (ppm – ppb); independent (i.e. not relying on an irradiation)determination of such iodine concentrations is challenging. For this reason, absolute calibration has been made by inclusion of potassium iodide in the irradiation so that the iodine content can be determined from the mass of KI irradiated. Macroscopic quantities of KI are required so that accurate mass measurements can be used to determine iodine content. This introduces two complications (Hohenberg and Kennedy, 1981; Hohenberg et al., 2000). Approximately half the conversion of 127I to 128Xe in a typical irradiation proceeds via resonant absorption of an epithermal neutron. In a sample of potassium iodide, the amountof iodine is high enough for self-shieldingto be important – neutrons at the resonant energies are depleted through interaction with iodine in the sample, effectively reducing the flux in parts of the KI aliquot below that experienced by the aliquots of other samples. Secondly, irradiation of such KI samples produces amounts of 128Xe vastly in excess of those that can be measured in a noble gas mass spectrometer, requiring dilution procedures before analysis, and the dilution process is also capable of introducing systematic errors. By addressing these issues, Hohenberg and Kennedy (1981) determined that the absolute initial iodine ratio of Bjurböle whole rock was(1.095 ± 0.029) x 10-4. Taking into account the age of Bjurböle relative to Shallowater and the half life of129I, the 129I/127I ratio for Shallowater is (1.07 ± 0.03) × 10-4 (Brazzle et al., 1999), where the error is dominated by the uncertainty in the 129I/127I ratio of Bjurböle.

Converting I-Xe ages to absolute ages.

Converting relative I-Xe ages into absolute ages is an entirely separate problem, which requires the determination of both an I-Xe age and an absolute age in at least one sample whereit is reasonable to think that they date the same event. As for other short-lived radioisotopes, the absolute chronometer based on lead isotopes produced by decay of 235U and 238U is the only one with precision comparable to the I-Xe system.

No Pb-Pb age has been determined for Shallowater enstatite, but its equivalent age in the Pb-Pb system can be determined from samples that have I-Xe ages determined relative to Shallowater and Pb-Pb ages. The first samples to yield ages in both systems were phosphate grains from Acapulco (Nichols et al., 1994) and ordinary chondrites (Brazzle et al., 1999). Gilmour et al. (2006) supplemented these data with other candidates from the literature and, by comparing Pb-Pb ages with I-Xe ages across a range of candidate samples, proposed an absolute age for the Shallowater standard. Gilmour et al. (2009) revisited this in the light of

1

Table 1. Xenon isotope data for analysis of two aliquots of the ungrouped achondrite NWA7325. All errors are 1 σ

Sample Mass (mg) / Laser Current (A) / Atoms 129Xe (× 10-13 cc STP g-1) / 124Xe/129Xe / 126Xe/129Xe / 128Xe/129Xe / 130Xe/129Xe / 131Xe/129Xe / 132Xe/129Xe / 134Xe/129Xe / 136Xe/129Xe
0.17 / 11.0 / n.d. / 0.1 (2) / n.d. / n.d. / 0.9 (8) / 0.6 (9) / 0.8 (8) / n.d. / n.d.
11.5 / n.d. / 0.0 (6) / n.d. / n.d. / n.d. / 3 (7) / 4 (7) / 0 (2) / n.d.
12.0 / n.d. / 0.1 (3) / 0.1 (2) / n.d. / 0.4 (7) / n.d. / 0.6 (8) / 0.6 (8) / 0.2 (5)
12.5 / 11 (1) / 0.00 (3) / n.d. / 100 (10) / n.d. / 11 (1) / 0.02 (7) / 0.01 (6) / 0.01 (5)
12.7 / 3.1 (9) / 0.03 (7) / 0.03 (7) / 80 (20) / n.d. / 50 (10) / 0.5 (3) / n.d. / n.d.
13.0 / 2.7 (8) / 0.6 (9) / 0.2 (1) / 90 (30) / n.d. / 120 (40) / 1.7 (6) / n.d. / 0.2 (2)
13.5 / 30 (2) / n.d. / 0.004 (8) / 15 (1) / 0.03 (2) / 32 (2) / 1.03 (9) / 0.29 (4) / 0.21 (3)
14.0 / 69 (3) / n.d. / 0.001 (3) / 4.6 (2) / 0.12 (2) / 9.4 (4) / 0.89 (5) / 0.56 (4) / 0.49 (3)
14.5 / 118 (2) / 0.005 (1) / 0.000 (2) / 1.93 (4) / 0.101 (6) / 1.72 (3) / 0.83 (2) / 0.53 (1) / 0.44 (1)
15.0 / 15 (1) / 0.00 (1) / 0.02 (2) / 0.79 (9) / 0.17 (5) / 0.9 (1) / 0.75 (9) / 0.40 (7) / 0.36 (6)
15.5 / 7 (1) / n.d. / n.d. / 1.0 (2) / 0.08 (8) / 0.9 (2) / 0.6 (1) / 0.3 (1) / 0.4 (1)
16.5 / 39 (1) / 0.005 (5) / n.d. / 0.70 (4) / 0.11 (2) / 0.85 (5) / 0.72 (4) / 0.41 (3) / 0.33 (3)
17.5 / 34 (1) / 0.0001 (6) / 0.001 (6) / 0.99 (6) / 0.06 (2) / 0.75 (5) / 0.46 (4) / 0.24 (3) / 0.23 (2)
18.5 / 0.8 (7) / n.d. / n.d. / 2 (2) / n.d. / 2 (2) / 1 (1) / 0.1 (6) / 0.0 (5)
20.5 / 37 (2) / 0.004 (5) / n.d. / 1.26 (7) / 0.06 (2) / 0.77 (5) / 0.31 (3) / 0.13 (2) / 0.11 (2)
20.5 / 10 (1) / 0.02 (2) / n.d. / 1.5 (2) / 0.11 (6) / 0.7 (1) / 0.12 (7) / 0.17 (7) / 0.14 (6)
Total / 376 (5)
1.69 / 11.0 / n.d. / 0.1 (5) / n.d. / n.d. / n.d. / n.d. / 2 (3) / 1 (2) / n.d.
11.5 / n.d. / n.d. / 0.0 (1) / n.d. / n.d. / 0.3 (5) / 0.5 (4) / 0.3 (3) / 0.4 (3)
12.0 / 1.4 (1) / n.d. / 0.001(2) / 67 (6) / n.d. / 12 (1) / 0.27 (7) / 0.07 (5) / 0.08 (4)
12.2 / 3.5 (3) / n.d. / n.d. / 56 (5) / n.d. / 24 (2) / 0.41 (7) / 0.09 (3) / 0.05 (3)
12.3 / 2.7 (3) / 0.03 (2) / 0.03 (2) / 56 (5) / n.d. / 48 (5) / 0.7 (1) / 0.06 (4) / 0.00 (3)
12.4 / 1.9 (2) / n.d. / 0.00 (2) / 63 (8) / n.d. / 139 (17) / 1.6 (3) / n.d. / 0.00 (4)
12.5 / 1.6 (2) / 0.02 (3) / n.d. / 53 (7) / n.d. / 165 (22) / 1.5 (3) / 0.04 (6) / 0.04 (5)
12.6 / 1.0 (2) / 0.03 (5) / 0.06 (4) / 60 (10) / n.d. / 228 (37) / 2.2 (4) / 0.03 (9) / 0.2 (1)
12.7 / 1.1 (2) / n.d. / n.d. / 60 (10) / 0.4 (1) / 219 (35) / 2.4 (4) / 0.09 (8) / n.d.
12.8 / 4.7 (4) / 0.02 (1) / 0.004 (7) / 23 (2) / 0.08 (3) / 57 (4) / 1.2 (2) / 0.22 (4) / 0.21 (4)
12.9 / 0.9 (2) / n.d. / 0.04 (4) / 60 (10) / 0.1 (1) / 130 (26) / 1.8 (4) / n.d. / 0.1 (1)
13.0 / 1.3 (2) / n.d. / 0.03 (3) / 52 (8) / n.d. / 118 (19) / 2.1 (4) / 0.3 (1) / 0.18 (9)
13.1 / 1.5 (2) / 0.02 (3) / 0.00 (2) / 32 (4) / n.d. / 103 (14) / 1.8 (3) / 0.19 (8) / 0.16 (7)
13.2 / 1.8 (2) / n.d. / 0.00 (1) / 24 (2) / n.d. / 63 (6) / 1.3 (2) / 0.25 (6) / 0.19 (5)
13.3 / 2.8 (2) / 0.02 (1) / 0.05 (3) / 16 (1) / 0.13 (4) / 39 (3) / 1.1 (1) / 0.26 (5) / 0.22 (4)
13.4 / 3.0 (2) / 0.01 (1) / n.d. / 16 (1) / 0.12 (3) / 28 (2) / 0.95 (8) / 0.30 (4) / 0.25 (4)
13.5 / 2.9 (1) / n.d. / 0.002(7) / 12.4 (6) / 0.09 (2) / 16.2 (8) / 0.97 (6) / 0.36 (3) / 0.30 (3)
13.6 / 2.9 (2) / 0.01 (1) / n.d. / 9.1 (6) / 0.14 (3) / 16 (1) / 0.83 (8) / 0.34 (4) / 0.29 (4)
13.7 / 3.3 (2) / 0.007 (7) / n.d. / 6.8 (3) / 0.10 (2) / 13.0 (6) / 0.82 (6) / 0.32 (3) / 0.28 (3)
13.8 / 3.4 (1) / 0.003 (6) / 0.007 (6) / 5.1 (2) / 0.10 (2) / 7.9 (3) / 0.86 (5) / 0.38 (3) / 0.31 (3)
14.0 / 8.8 (2) / 0.015 (7) / 0.005 (4) / 5.8 (3) / 0.11 (2) / 24 (1) / 0.97 (7) / 0.35 (4) / 0.33 (3)
14.1 / 10.5 (4) / 0.002 (5) / 0.003 (4) / 6.3 (3) / 0.03 (1) / 18.8 (8) / 0.81 (5) / 0.37 (3) / 0.32 (3)
14.2 / 26.1 (6) / 0.002 (2) / 0.003 (2) / 6.8 (2) / 0.081 (8) / 10.9 (3) / 0.76 (3) / 0.37 (2) / 0.32 (2)
14.3 / 13.3 (5) / 0.003 (4) / 0.001 (3) / 6.3 (3) / 0.08 (1) / 10.1 (4) / 0.87 (5) / 0.46 (3) / 0.41 (3)
14.4 / 11.7 (4) / 0.002 (3) / 0.001 (2) / 4.6 (2) / 0.10 (1) / 6.9 (2) / 0.81 (4) / 0.43 (2) / 0.35 (2)
14.5 / 11.2 (4) / 0.003 (3) / 0.003 (3) / 4.4 (2) / 0.10 (1) / 6.2 (2) / 0.79 (4) / 0.44 (3) / 0.36 (2)
14.6 / 11.7 (3) / 0.003 (3) / 0.002 (2) / 4.3 (1) / 0.089 (9) / 5.9 (2) / 0.79 (3) / 0.45 (2) / 0.40 (2)
14.7 / 12.5 (3) / n.d. / 0.003 (2) / 3.9 (1) / 0.086 (8) / 4.3 (1) / 0.84 (3) / 0.47 (2) / 0.40 (2)
14.8 / 9.9 (2) / 0.001 (2) / 0.000 (2) / 3.65 (9) / 0.094 (7) / 4.25 (9) / 0.79 (2) / 0.47 (2) / 0.42 (1)
14.9 / 15.1 (2) / 0.002 (1) / 0.000 (1) / 2.82 (6) / 0.101 (6) / 2.83(5) / 0.83 (2) / 0.58 (1) / 0.51 (1)
15.0 / 8.0 (2) / 0.003 (2) / 0.001 (2) / 2.68 (8) / 0.073 (9) / 2.70 (8) / 0.79 (3) / 0.53 (2) / 0.47 (2)
15.2 / 6.9 (2) / 0.001 (3) / 0.001 (3) / 2.49 (8) / 0.09 (1) / 2.37 (7) / 0.76 (3) / 0.45 (2) / 0.39 (2)
15.4 / 4.5 (2) / n.d. / 0.002 (4) / 2.9 (1) / 0.07 (1) / 3.2 (1) / 0.70 (4) / 0.44 (3) / 0.37 (3)
15.6 / 3.7 (2) / n.d. / 0.005 (6) / 2.5 (1) / 0.06 (2) / 2.3 (1) / 0.74 (5) / 0.45 (4) / 0.40 (3)
15.8 / 4.4 (2) / 0.002 (4) / n.d. / 3.0 (1) / 0.09 (2) / 2.7 (1) / 0.74 (4) / 0.48 (3) / 0.44 (3)
16.0 / 12.4 (2) / 0.002 (2) / 0.002 (2) / 3.07 (7) / 0.094 (6) / 3.80 (7) / 0.92 (2) / 0.65 (2) / 0.56 (1)
16.2 / 28.6 (3) / 0.004 (1) / 0.0005 (7) / 3.22 (5) / 0.114 (4) / 3.54 (4) / 0.96 (1) / 0.66 (1) / 0.575 (9)
16.4 / 7.1 (2) / 0.003 (3) / 0.001 (3) / 2.62 (9) / 0.10 (1) / 2.55 (9) / 0.81 (3) / 0.50 (3) / 0.41 (2)
16.6 / 7.0 (2) / n.d. / 0.001 (3) / 2.42 (8) / 0.11 (1) / 1.77 (6) / 0.75 (3) / 0.41 (2) / 0.35 (2)
16.8 / 14.5 (3) / 0.002 (2) / 0.000 (1) / 2.51 (6) / 0.098 (7) / 1.77 (4) / 0.73 (2) / 0.38 (1) / 0.33 (1)
17.0 / 10.1 (2) / 0.003 (2) / 0.000 (2) / 2.31 (6) / 0.094 (8) / 1.43 (4) / 0.69 (2) / 0.38 (2) / 0.36 (1)
17.2 / 23.4 (3) / 0.003 (1) / 0.001 (1) / 2.35 (5) / 0.093 (5) / 1.75 (3) / 0.63 (1) / 0.31 (1) / 0.265 (9)
17.4 / 6.6 (2) / 0.005 (3) / n.d. / 1.86 (6) / 0.09 (1) / 1.30 (5) / 0.64 (3) / 0.29 (2) / 0.23 (1)
17.6 / 3.7 (2) / 0.004 (6) / n.d. / 1.68 (9) / 0.09 (2) / 1.09 (7) / 0.59 (4) / 0.25 (3) / 0.23 (3)
18.0 / 1.7 (1) / 0.02 (1) / n.d. / 2.1 (2) / 0.16 (4) / 1.7 (2) / 0.74 (8) / 0.29 (5) / 0.28 (5)
18.5 / 14.8 (2) / 0.001 (1) / 0.003 (2) / 0.96 (2) / 0.100 (6) / 0.73 (2) / 0.64 (1) / 0.251 (9) / 0.200 (7)
19.0 / 4.8 (2) / 0.002 (4) / n.d. / 1.31 (7) / 0.12 (2) / 0.89 (5) / 0.67 (4) / 0.24 (2) / 0.22 (2)
19.5 / 2.0 (1) / 0.00(1) / 0.00 (1) / 2.3 (2) / 0.05 (3) / 1.2 (1) / 0.65 (6) / 0.31 (4) / 0.22 (3)
20.5 / 3.5 (1) / 0.001 (7) / 0.001 (5) / 1.00 (6) / 0.09 (2) / 0.84 (5) / 0.69 (4) / 0.27 (3) / 0.22 (2)
20.5 / 15.6 (2) / 0.002 (1) / 0.001 (1) / 1.01 (3) / 0.096 (6) / 0.94 (2) / 0.67 (2) / 0.27 (1) / 0.208 (8)
20.5 / 0.7 (1) / 0.01 (4) / 0.02 (3) / 1.1 (2) / 0.17 (9) / 1.0 (2) / 0.6 (1) / 0.4 (1) / 0.33 (9)
20.5 / 1.2 (1) / 0.02 (2) / n.d. / 1.0 (1) / 0.17 (5) / 1.0 (1) / 0.8 (1) / 0.37 (7) / 0.24 (5)
Total / 358 (2)

Table 2. Candidate samples for calibration of the I-Xe system (extended from Pravdivtseva et al., 2016).

Sample / I-Xe Age / Pb-Pb Age, / M2b / M2c
(Myr) / (Ma)a
Earliest Chondrule / 4.3 / ± 0.6 / Swindle et al. (1991b) / 4567.3 / ± 0.2 / Connelly et al. (2012) / 0.5 / 0.0
Richardton Px / 1.1 / ± 2.0 / Pravdivtseva et al. (1998) / 4560.4 / ± 0.6 / Amelin (2001) / 2.2 / 3.0
Richardton Chondrule 2 / -0.1 / ± 0.1 / Gilmour et al. (2006) / 4561.4 / ± 0.8 / Gilmour et al. (2006) / 1.3 / 3.3
Richardton Chondrule 6 / -4.1 / ± 0.6 / Gilmour et al. (2006) / 4557.2 / ± 1.4 / Gilmour et al. (2006) / 0.4 / 1.1
Acapulco Fspar / -3.8 / ± 1.5 / Brazzle et al. (1999) / 4556.1 / ± 1.0 / Göpel et al. (1994) / 1.7 / 2.7
Ste Marguerite / 0.7 / ± 0.4 / Brazzle et al. (1999) / 4561.8 / ± 0.3 / Göpel et al. (1994) / 7.2 / ---
Kernouve Phosphate / 43.3 / ± 6.0 / Brazzle et al. (1999) / 4521.6 / ± 0.7 / Göpel et al. (1994) / 0.3 / 0.2
HH237 chondrule / -0.3 / ± 0.2 / Pravdivtseva et al. (2016) / 4562.5 / ± 0.2 / Bollard et al. (2015) / 2.0 / 0.6
Ibitira chip / -7.0 / ± 1.0 / Claydon (2014) / 4557.0 / ± 0.3 / Iizuka et al. (2014) / 2.7 / 1.2
NWA7325 chip / -1.6 / ± 2.3 / This work / 4563.4 / ± 1.3 / Koefoed et al. (2016) / 0.9 / 0.6

aRecalculated where necessary using the uranium isotope date of Goldmann et al. (2015).

bSum of squares of the Mahalanobis distances relative to the best fit line to all data.

cSum of squares of the Mahalanobis distance relative to the best fit line excluding Ste Marguerite.

1

new data and, using likelihood to exclude outliers, proposed an absolute age for the Shallowater standard of 4562.3 ± 0.4Ma.

Since this publication, a correction has been made to the Pb-Pb chronometer. It was recognised that decay of 247Cmrequired a revision of the 235U/238U ratio used to calculate ages(Brennecka et al., 2010). Further work has shown significant variation in the 235U/238U ratio of chondrites, including those on which the I-Xe calibration was based (Goldmann, 2015). Pravdivtseva et al. (2016) incorporated the effects of these changes and, including a new datum for a chondrule from the CB chondrite Hammadah al Hamra (HH) 237, proposed an absolute Pb-Pb age for the Shallowater standard of 4562.4 ± 0.2 Ma. This calibration can be considered in the light of new I-Xe and Pb-Pb data.

The earliestPb-Pb chondrule ages (4567.3±0.2, 1 σ error) are now contemporaneous with CAI formation (Connelly et al., 2012) and, as originally proposed by Gilmour et al. (2006), we consider them comparable with the earliest chondrule I-Xe chondrule age (Swindle, 1991b). Two chips (1.6 mg and 4.6 mg) of the anomalous eucrite Ibitira yielded a“whole rock” I-Xe age of -7±1 Myr (Claydon, 2012), which can be compared to a Pb-Pb age of 4556.8 ± 0.6 (1 σ error, Iizuka et al., 2012).

Finally, the ungrouped achondrite NWA7325 (Irving et al., 2013; Weber et al., 2016) has been reported to have a Pb-Pb age of4563.4±1.3 (1 σ error,Koefoedet al. 2016). In Fig 1, and in Table 1, we present I-Xe data for aliquots of this sample, which yields an I-Xe age of -1.6±2.3.

In Table 2, which extends Table 2 of Pravdivtseva et al. (2016), we summarise the data potentially available for calibration of the I-Xe system as originally proposed by Gilmour (2006). In this approach: absolute Pb-Pb ages are plotted against I-Xe ages relative to Shallowater; a line is fit through the data; the intercept of the line corresponds to the absolute age of Shallowater, and the gradient of the line tests whether ages are varying concordantly. This approach tests the assumptions of contemporaneous closure across a range of samples rather than relying on an argument based on the petrologic history of a single sample. In light of the larger dataset now available, we do not include I-Xe ages that were derived via the Mn-Cr chronometer in the original publication (Gilmour et al., 2006).

When using the dataset of Pravidivtseva et al. (2016) we derive an age for the Shallowater standard of 4562.3±0.2 Ma (consistentwith their 4562.2±0.2 Ma within a rounding error), a gradient of 1.01±0.12 (their 1.01±0.11) and MSWD = 2.2. For the complete dataset of our Table 2 we obtain a Shallowater age of 4562.4 ± 0.2 Ma, a gradient of 1.02 ± 0.09, and an MSWD = 2.4. In each case, the MSWD indicates more scatter than would be expected based on reported measurement errors, indicating a significant breakdown in the assumption that the two systems closed at the same time. This in turn suggests that the derived uncertainties should be increased by a factor of at least 2.

Closer inspection reveals that the Ste Marguerite datum has a squared Mahalanobis distance of 7.2 from the best fit line (Table 2) – the next highest value is for Richardton pyroxene (2.7). The Ste Marguerite datum relies on associatingthe Pb-Pb age of phosphate with the I-Xe age of feldspar, and there is evidence that the I-Xe and Pb-Pb systems have different closure characteristics in these two minerals (Crowther et al., 2009). We therefore exclude this data point from the fit, leading to an MSWD of 1.5 for the fit to the remaining points. Rather than proceed with further outlier rejection, we scale uncertainty for the intercept and gradient by the MSWD, leading to a best age for the Shallowater standard of 4562.7±0.3 (1 σ). The associated gradient is 1.02±0.09. The covariance between the uncertainties on gradient and intercept, which should be considered if using this calibration to derive equivalent Pb-Pb ages, is 1.2x10-3. This fit and the data discussed are displayed in Fig 2. Further sequential rejection of the outlier with the greatest squared Mahalanobis distancewouldlead to a Shallowater age of 4562.8±0.2 (MSWD = 1.13), then 4562.9±0.2 (MSWD = 0.74). As discussed by Pravdivtseva et al. (2016), the proximity of the gradient of the free fit correlation line to unity indicates consistency among the half lives of the radioisotopes to within the tolerance of the fit.

In the revised Pb-Pb chronology the age of CAI formation is 4567.3±0.2 (Connelly et al., 2012), which corresponds to an initial iodine ratio 129I/127I ≈ 1.4 x 10-4. It is interesting to compare this to an average galactic background at the time of solar system formation, a calculation first done by Wasserburg et al., (1960). 129Xe is predominantly an r-process isotope, and thus has been produced via 129I. Since the age of the galaxy at solar system formation was around 9 Gyr, approximately one part in 104 of the current 129Xe budget was produced over each million years of galactic history before solar system formation. The 129Xe/I ratio of average solar system material is ≈1(Anders and Grevesse, 1989), so the rate of change of this ratio at the time of solar system formation was ≈ 10-4 Myr-1. This is related to the 129I abundance at this time by the decay constant, so we have λ(129I/127I) = 10-4, where λ is the decay constant of 129I as above. This leads to a galactic average 129I/127I ratio at solar system formation of 0.0023. To reach the initial solar system value would require a free decay period of around 70 Myr. The star formation rate has declined with time over the lifetime of the galaxy, so the supernova rate and the rate of production of r-process isotopes will also have declined, so the production rate of 129I in the recent history of material from which the solar system was formed was lower, and so this interval is an overestimate. Nonetheless, the calculation demonstrates that there is no requirement for a significant input of 129I shortly before solar system formation, and thus no grounds to expect heterogeneity of the 129I/127I ratio across the solar system. Homogeneous distribution of the parent isotope is a requirement before variations in its relative abundance from sample to sample can be interpreted chronologically.