238U/235U Variations in Meteorites: Extant 247Cm and Implications for Pb-Pb Dating

See allHide authors and affiliations

Science  22 Jan 2010:
Vol. 327, Issue 5964, pp. 449-451
DOI: 10.1126/science.1180871

This article has a correction. Please see:


The 238U/235U isotope ratio has long been considered invariant in meteoritic materials (equal to 137.88). This assumption is a cornerstone of the high-precision lead-lead dates that define the absolute age of the solar system. Calcium-aluminum–rich inclusions (CAIs) of the Allende meteorite display variable 238U/235U ratios, ranging between 137.409 ± 0.039 and 137.885 ± 0.009. This range implies substantial uncertainties in the ages that were previously determined by lead-lead dating of CAIs, which may be overestimated by several million years. The correlation of uranium isotope ratios with proxies for curium/uranium (that is, thorium/uranium and neodymium/uranium) provides strong evidence that the observed variations of 238U/235U in CAIs were produced by the decay of extant curium-247 to uranium-235 in the early solar system, with an initial 247Cm/235U ratio of approximately 1.1 × 10−4 to 2.4 × 10−4.

Meteorites can provide a wealth of information about the formation and evolution of the solar system. In chondrite meteorites, calcium-aluminum–rich inclusions (CAIs) represent the first solids to condense from the cooling protoplanetary disk during the birth of the solar system (1); therefore, the ages of CAIs are generally considered to date the solar system’s origin (24). High-precision Pb-Pb dating studies, which rely on a known ratio of parent U isotopes, assume that the 238U/235U ratio is invariant in meteoritic material (equal to 137.88) (5). Uranium isotope variations in meteorites may be produced by many mechanisms, including the decay of extant 247Cm to 235U, nucleosynthetic anomalies in U isotopes, or fractionation of U isotopes during chemical reactions, as recently observed on Earth (6, 7). Any or all of these mechanisms may play some role in 238U/235U variability in early solar system materials; however, the existence and effect of 247Cm on the 238U/235U ratio can be studied using geochemical proxies for Cm.

247Cm is only created in certain types of supernovae during r-process nucleosynthesis. It decays to 235U with a half-life of 15.6 million years (My) (813). If 247Cm was present during the formation of the solar system, it would be detected by variations of 238U/235U in ancient meteoritic materials in which the original solar system Cm/U ratio may have been substantially fractionated by processes associated with the formation of the meteoritic materials. The CAIs in chondritic meteorites are likely to be such materials, because many of them experienced elemental fractionation during condensation and evaporation processes that were involved in their formation and because Cm is more refractory than U (14).

Quantification of the abundance of extant 247Cm has the potential to provide new constraints on the origin of short-lived radionuclides in the early solar system. If the 247Cm in the early solar system was predominantly inherited from galactic chemical evolution (13), then it should be possible for us to determine the time interval of free decay (Δ) between the last r-process nucleosynthetic event and the formation of the solar system (5, 11, 15, 16). Supposed claims of large variations in the 238U/235U ratio that were caused by the decay of 247Cm (8, 9) were refuted in subsequent studies (5, 10, 11, 17). Here we present high-precision 238U/235U ratios obtained from 13 CAIs of the Allende meteorite to quantify the amount of 247Cm present in the early solar system and to determine the extent of potential offsets in the calculated Pb-Pb ages of early solar system materials (18).

The 238U/235U ratios of the two bulk meteorites (Allende and Murchison) are 137.818 ± 0.012 and 137.862 ± 0.042, respectively (Fig. 1). The 13 CAIs show a large range of U isotope compositions, with 238U/235U ratios varying from 137.409 ± 0.039 to 137.885 ± 0.009. All but two CAIs differ outside uncertainties from the standard value, and five CAIs have significantly lower 238U/235U values than that of bulk Allende.

Fig. 1

238U/235U isotope values for the samples of this study. The box represents the measured value and analytical precision of replicate analyses of 20– to 100–parts per billion solutions of the SRM950a standard. Error bars are calculated as 2 times the standard deviation (2SD) of multiple runs of each sample, when possible. In samples with extremely limited uranium, for which fewer than three runs were possible, the reported errors are conservatively represented by the long-term reproducibilities (2SD) based on multiple runs of SRM950a measured over the course of this study at the same concentration as the sample.

If 247Cm decay is the primary mechanism for 238U/235U variability, then materials with a high initial Cm/U value would contain a higher relative amount of 235U than those with lower initial Cm/U values. However, because Cm has no long-lived stable isotope, the initial Cm/U ratio of a sample cannot be directly determined. Because Th and Nd have similar geochemical behavior to Cm, Th/U and Nd/U ratios can serve as proxies for the initial Cm/U ratio in the sample (9, 11). Our sample set spans a large range of Th/U and Nd/U, and both these ratios correlate with the U isotopic composition (Fig. 2).

Fig. 2

(A) 232Th/238U and (B) 144Nd/238U ratios plotted versus 235U/238U ratios, the reciprocal values of our measured 238U/235U ratios. The gray dashed lines represent the 2SD errors on the best-fit line (solid black). Errors on the y-axis data are ±2SD; x-axis error bars are ±5% of the determined value of the elemental ratio.

Because of the higher volatility of U, thermodynamic calculations suggest that substantial fractionation of Cm (and other geochemically similar elements such as Th and Nd) from U is possible in the early solar nebula (19). Large variations in the Th/U and Nd/U ratios seen in our CAI data set (table S1) support this claim. A special group of CAIs, called group II CAIs, are distinguished by a unique abundance pattern of the rare earth elements (REEs). Group II CAIs are highly depleted in the most refractory (that is, heavy REEs, except Tm and Yb) and the most volatile (that is, Eu and Yb) REEs, yet the moderately refractory light REEs (including Nd) are only present in chondritic relative abundances (20). This REE pattern, which is characteristic of group II CAIs, suggests a complex condensation history involving fractional condensation (21, 22). The four CAIs of this study that have the highest Nd/U and Th/U ratios (as well as the lowest 238U/235U ratios) are all classified as group II CAIs by their REE patterns (Fig. 3). Because of the lower condensation temperature of U relative to Nd and Th (23), the fractional condensation history that resulted in the characteristic group II REE pattern in these objects is likely to have produced the relatively high Nd/U and Th/U ratios.

Fig. 3

REE patterns of four group II CAIs analyzed in this study, normalized to CI chondrites. All other CAI samples studied here (except 3531-D, for which the REE abundances were not measured) display flat REE patterns, indicating chondritic relative abundances of these elements (light gray lines).

The correlation of both Th/U and Nd/U with U isotope ratios in the CAIs indicates that the 238U/235U variations do not arise from nucleosynthetic anomalies or U isotope fractionation, neither of which easily give rise to such a trend, and instead provide evidence for the presence of extant 247Cm in the early solar system. Under this interpretation, deviations from the best-fit lines in Fig. 2 could be caused by heterogeneity of 238U/235U in the solar nebula, Th and Nd acting as imperfect proxies for Cm, or 238U/235U fractionation following Allende CAI formation, possibly from variable redox during secondary alteration processes (7).

In contrast to our findings, a recent study did not detect deviations in the 238U/235U ratio among a variety of bulk meteorite samples, including Allende and Murchison (11). Given the reported precision of the study’s U isotope analysis, the 144Nd/238U ratios should have been sufficient to reveal detectable variations in 238U/235U from 247Cm decay. Although the 238U/235U value of bulk Murchison samples agrees within error with our observed values, those for bulk Allende differ well outside of reported errors. The reason for this disagreement is unclear at this time.

The initial 247Cm/235U ratio in the early solar system can be estimated by using the slopes of the best-fit lines in Fig. 2 (11). Using Th and Nd as proxies for Cm, we estimate the initial solar system 247Cm/235U ratio to be 2.4 × 10−4 ± 0.6 × 10−4 and 1.1 × 10−4 ± 0.2 × 10−4, respectively. The difference between the estimates may be due to slight differences in the geochemical behavior of Th and Nd or possibly because of uncertainties in the assumed solar system Nd/U or Th/U ratios. Nevertheless, these values are, on average, higher than the upper limit derived previously using analyses of the U isotope compositions of bulk chondritic meteorites (11). Our estimates are, however, in agreement with the upper limit of ~4 × 10−3 that was determined previously based on analyses of CAIs (12). If 247Cm is inherited from galactic chemical evolution, the range of initial solar system 247Cm/235U ratios estimated here translates to Δ ~ 110 to 140 My. This value is similar to, but more precise than, previous estimates of Δ based on the inferred initial solar system abundances of other r-process–only radionuclides such as 244Pu and 129I, but does not match the significantly shorter estimate of Δ (~30 My) derived from the initial abundance of 182Hf (16). However, because 182Hf was overabundant in the early solar system compared with its expected abundance from galactic chemical evolution, it may have been injected into the presolar molecular cloud or the solar nebula by a nearby supernova event [for example, (13)].

Our findings also have implications for precise dating of early events in the history of the solar system. The Pb-Pb age equation (Eq. 1) has been used for decades to calculate the absolute ages of both meteoritic and terrestrial materials (24). This equation assumes that 238U/235U is invariant at any given time, and that the present-day value is 137.88.P206b*P206b*=U235eλ235t1U238eλ238t1=1137.88eλ235t1eλ238t1(1)Here, λ is the decay constant for the specific isotope and t is the age. Any deviation from this assumed 238U/235U would cause miscalculation in the determined Pb-Pb age of a sample. A difference of up to 3.5 per mil (‰) implies that a correction of up to –5 My would be required if the Pb-Pb ages of these CAIs were obtained using the previously assumed 238U/235U value (Fig. 4).

Fig. 4

Age adjustment required for samples found not to have a 238U/235U value of 137.88, as assumed in the Pb-Pb age equation (Eq. 1). The shaded region represents the range of U isotope compositions reported in this study, and the asterisks represent the specific 238U/235U ratios measured in these samples.

Because 238U/235U variations in solar system materials are not restricted to CAIs, this requirement may extend to high-precision Pb-Pb dating of other materials as well. It is possible, however, that the 238U/235U values of bulk chondrites are controlled to a substantial degree by CAIs, which may be heterogeneously distributed at the scale at which these analyses were made.

The Pb-Pb dating technique is the only absolute dating technique able to resolve age differences of <1 My in materials formed in the early solar system. Whereas the full range of 238U/235U ratios reported here would result in an overestimation of the ages of these CAIs by up to 5 My, the largest excesses (>3.5‰) in 235U occur in the group II CAIs that appear to have experienced the largest Cm/U fractionation. For non–group II CAIs, the age overestimation is ≤1 My. The apparent discrepancies between absolute Pb-Pb ages and relative (for example, 26Al-26Mg, 53Mn-53Cr, and 182Hf-182W) ages (2, 4, 25, 26) may therefore place limits on the uncertainty of the age of the solar system.

Supporting Online Material

Materials and Methods

Fig. S1

Tables S1 and S2


  • Present address: Institut für Geology und Mineralogie, Universität zu Köln, Cologne, Germany.

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank the Center for Meteorite Studies at Arizona State University and the Senckenberg Museum in Frankfurt for providing the samples. We thank the W. M. Keck Laboratory for Environmental Biogeochemistry and R. Hines for technical assistance. We are grateful to H. Palme for helpful discussions and to R. Williams and L. Borg for assistance with the double spike. We also thank G. Wasserburg, S. Galer, and an anonymous reviewer for thoughtful comments that greatly improved the manuscript. This work was partially supported by NASA Origins of Solar Systems grant NNX07AF49G to M.W., as well as NASA Astrobiology Institute grant NNA09DA79A and NASA Exobiology Program grant NNX07AU15G to A.D.A.

Stay Connected to Science

Navigate This Article