Lead Isotopic Ages of Chondrules and Calcium-Aluminum-Rich Inclusions

See allHide authors and affiliations

Science  06 Sep 2002:
Vol. 297, Issue 5587, pp. 1678-1683
DOI: 10.1126/science.1073950


The lead-lead isochron age of chondrules in the CR chondrite Acfer 059 is 4564.7 ± 0.6 million years ago (Ma), whereas the lead isotopic age of calcium-aluminum–rich inclusions (CAIs) in the CV chondrite Efremovka is 4567.2 ± 0.6 Ma. This gives an interval of 2.5 ± 1.2 million years (My) between formation of the CV CAIs and the CR chondrules and indicates that CAI- and chondrule-forming events lasted for at least 1.3 My. This time interval is consistent with a 2- to 3-My age difference between CR CAIs and chondrules inferred from the differences in their initial26Al/27Al ratios and supports the chronological significance of the 26Al-26Mg systematics.

Chondritic meteorites (chondrites) consist of three major components: refractory CAIs, less refractory ferromagnesian silicate spherules called chondrules, and a fine-grained matrix. It is generally believed that CAIs and chondrules formed in the solar nebula (a disk of dust and gas surrounding the proto-Sun) by high-temperature processes that included condensation, evaporation, and, for all chondrules and many CAIs, subsequent melting during multiple brief heating episodes (1–3). The mechanisms involved in chondrule formation are uncertain: shock waves, lightning discharges, and X-wind (jet flow) are currently being considered (2, 4–8). The existing estimates for the timing of CAI and chondrule formation are either controversial or insufficiently precise. Thus, the total duration of CAI and chondrule formation, which can provide important constraints on their origin, remains obscure.

CAIs and chondrules formed with different initial contents of the short-lived radionuclide 26Al [half-life (t 1/2) ∼ 0.73 Ma]. Most CAIs show large excesses of its decay product, 26Mg*, corresponding to an initial 26Al/27Al ratio [(26Al/27Al)I] of ∼4 × 10−5 to 5 × 10−5 (9). Chondrules, in contrast, show only small or undetectable26Mg*, implying (26Al/27Al)I ≤ 1.5 × 10−5 (10–12). This difference may indicate that CAIs formed at least 1 to 2 My earlier than chondrules (9–13). This chronological interpretation is based on the assumption that 26Al had a stellar origin (such as from a supernova, asymptotic giant branch star, or Wolf-Rayet star) and was injected and homogenized in the solar nebula over a time scale that was short compared to its half-life (14). A stellar origin for 26Al is consistent with correlated abundances of 26Al and41Ca (t 1/2 ∼ 0.1 Ma) in CAIs from the CV (Vigarano-like) and CM (Murchison-like) carbonaceous chondrites (15). However, the observed heterogeneity in26Al distribution among CAIs from the CH (ALH85085-like) carbonaceous chondrites (16), and the possibility that the 26Al-26Mg systematics in chondrules and CAIs has been disturbed by late-stage asteroidal processing (17, 18), casts doubt on its chronological significance.

The alternative nonchronological interpretation of26Al/26Mg systematics involves a local origin of 26Al by energetic particle irradiation near the proto-Sun, resulting in radial heterogeneity of 26Al distribution (4, 19). According to this interpretation, chondrules formed contemporaneously with CAIs, but farther away from the Sun (4). Although the detection in CAIs of the short-lived radionuclides 10Be (t 1/2 ∼ 1.5 Ma) and 7Be (t 1/2 ∼ 52 days), which can be produced only by nuclear spallation reactions (20, 21), may support the irradiation origin of 41Ca and the short-lived radionuclide 53Mn (t 1/2 ∼ 3.7 Ma), this mechanism still remains problematic for 26Al (20,22). The uncertainty of origin (stellar versus irradiation) of53Mn and 41Ca hampers the chronological significance of the 53Mn-53Cr and41Ca-41K systematics. The possible use of the53Mn-53Cr systematics for distinguishing formation ages of CAIs and chondrules is additionally complicated by the unknown distribution (homogeneous versus heterogeneous) of53Mn, the unknown initial 53Mn/55Mn ratio of the solar system, and possible mobilization of Mn during thermal metamorphism (23–25).

In contrast to the 26Al-26Mg and53Mn-53Cr systematics, which can provide only relative ages, the 207Pb/206Pb chronometer can provide absolute formation ages of CAIs and chondrules (26). The errors of207Pb/206Pb dates can be as low as 0.5 to 1.5 My; thus, the 207Pb/206Pb chronometer may be suitable for resolving a potential 2- to 3-My age difference between CAIs and chondrules. The current best model Pb-Pb age of CAIs from the Allende CV chondrite of 4566 ± 2 Ma (27,28) is not sufficiently precise, however. Here, we report more precise Pb-Pb internal isochron (29) ages for two CAIs from the CV chondrite Efremovka (E60 and E49) and a similarly precise Pb isotopic age for chondrules from the CR (Renazzo-like) carbonaceous chondrite Acfer 059. For the CAI E60, we also report the26Al-26Mg systematics.

CR chondrites are among the most primitive meteorites: They escaped thermal metamorphism and suffered only aqueous alteration at temperatures below 100° to 150°C (30–33). The degree of aqueous alteration varies among CR chondrites; e.g., Al Rais and Renazzo contain abundant phyllosilicates and carbonates in their CAIs, chondrules, and matrices, whereas chondrules and CAIs in other CR chondrites, including Acfer 059, are virtually phyllosilicate-free (32–36). The CR CAIs have uniformly 16O-enriched compositions (35) and the canonical (26Al/27Al)Iratio of 4 × 10−5 to 5 × 10−5(16, 36), suggesting that they preserved their primary isotopic characteristics undisturbed.

The vast majority of chondrules in CR chondrites are large (∼0.7 to 1 mm in apparent diameter), FeNi-metal–rich, sulfide-free, and volatile-poor. Chondrule olivine and pyroxenes are Cr2O3-rich (0.5 to 1 weight %) and FeO-poor (Fa/Fs1-2). FeNi metal has a large range in Ni contents (4 to 14 weight %) with a solar Co/Ni ratio (32, 33). Many chondrules are surrounded by silica-bearing igneous rims (37). On the basis of these characteristics, it has been inferred that CR chondrules were formed at high ambient nebular temperatures and escaped remelting at low ambient temperatures (33,37). In contrast to the CAIs, chondrules from CR chondrites contain only small or undetectable 26Mg*. Preliminary Al-Mg results for chondrules from CR and CV chondrites suggest the range of (26Al/27Al)Ifrom ∼1 × 10−5 to <3 × 10−6(12).

The CV chondrites are a diverse group of meteorites that experienced various degrees of aqueous and/or Fe-alkali metasomatic alteration and mild (<500°C) thermal metamorphism (38). The Efremovka meteorite is one of the least altered and metamorphosed CV chondrites, which, however, experienced relatively strong shock metamorphism (39). Two Efremovka CAIs studied for Mg and Pb isotopes are a forsterite-bearing Type B inclusion E60 and a compact Type A inclusion E49. E60 is a spherical (∼15 mm in apparent diameter) inclusion composed of Al-Ti-diopside (8 to 18 weight % Al2O3, 0.4 to 3.3 weight % TiO2), melilite (Åk22-86), anorthite, Mg-spinel, and Ca-rich forsterite (40). E49 is an ellipsoidal (∼3.2 mm by 4.5 mm) inclusion composed of melilite (Åk16-36) and Mg-spinel. Both CAIs are surrounded by spinel-melilite-Al-diopside Wark-Lovering rims and contain very minor secondary nepheline.

Acid-washed and untreated Acfer 059 chondrules, as well as acid leachates, and the Acfer 059 matrix were analyzed for U and Pb concentrations and Pb isotopic compositions (41–43). The concentration ranges for U [2.5 to 35 parts per billion (ppb)] and Pb (6 to 44 ppb) in acid-washed chondrules (Table 1) are similar to previously reported values in ordinary chondrite chondrules (44–46). The range of measured206Pb/204Pb ratios covers almost two orders of magnitude from 23.3 to 2198 (Table 1). The matrix contains >700 ppb of Pb with rather unradiogenic Pb isotopic composition (206Pb/204Pb = 11.09), which is within the range of values for the Allende matrix (47). Contamination of chondrule material with matrix would severely reduce the radiogenic character of Pb and compromise the age determinations. The low 206Pb/204Pb ratios and elevated Pb concentrations in acid leachates of chondrules, together with highly radiogenic Pb and low Pb concentration in most washed chondrules, demonstrate the efficiency of the acid washing (Table 1).

Table 1

Pb isotope data and model dates.

View this table:

Multiple fractions from the Efremovka CAIs E60 and E49 were analyzed with similar techniques. The concentration range of U (11 to 36 ppb) in the acid-washed fractions from the CAIs is the same as in the Acfer 059 chondrules (Table 1). The Pb concentrations (22 to 97 ppb) are slightly higher than in the chondrules. Acid-washed CAI fractions have consistently radiogenic Pb with measured206Pb/204Pb ratios between 203 and 1263. Acid leachates contain elevated concentrations of common Pb, but only small amounts of U.

Model 207Pb/206Pb dates for the Acfer 059 chondrules are calculated by using primordial Pb (48). The dates generally increase for more radiogenic Pb isotopic compositions, approaching the value of ∼4564 Ma (Table 1). The variation of the dates with the measured 206Pb/204Pb suggests that the common Pb isotopic composition in the chondrules differs from that of primordial Pb. Although the bias in the model dates, related to inaccurate assumption of initial common Pb, decreases with increasing 206Pb/204Pb, this bias can be eliminated only by calculating isochron dates instead of model dates (49). Because of the possibility of partial dissolution and preferential leaching of U or Pb during intensive acid washing, we consider only Pb-Pb dates (table S1).

The Pb-Pb isochron and “errorchron” (50) dates (Table 2) show a decrease in the dispersion of the data, expressed by mean square weighted deviation (MSWD) values when analyses having high common Pb contents are removed. This demonstrates not only that the common Pb in the chondrules differs from the primordial Pb, but also that two or more components of common Pb (e.g., primordial Pb and modern terrestrial common Pb introduced by weathering) are present, and that these components are unevenly distributed among the chondrules. In order to obtain a highly precise and accurate date, we applied isochron regression to the data with the lowest common Pb content, for which the variations in the initial Pb isotopic compositions are insignificant compared with the analytical errors. The six most radiogenic Pb isotope data points define an isochron with an age of 4564.7 ± 0.6 Ma, MSWD = 0.5 (Table 2, Fig. 1). The “errorchrons” shown in Table 2 give dates that agree with this best date, but are less precise due to excess scatter, caused mainly by the common Pb isotopic variations.

Figure 1

Pb-Pb isochrons for the six most radiogenic Pb isotopic analyses of acid-washed chondrules from the CR chondrite Acfer 059 (solid line), and for acid-washed fractions from the Efremovka CAIs (dashed lines). 207Pb/206Pb ratios are not corrected for initial common Pb. Error ellipses are 2σ. Isochron age errors are 95% confidence intervals.

Table 2

Summary of Pb-Pb isochron regressions. Isochron and errorchron regressions show a decrease in the dispersion of the data, expressed by MSWD values, when analyses having high206Pb/204Pb are removed. Regression for the washed chondrule data with 206Pb/204Pb > 300 (line 5) shows relatively small residual scatter (MSWD = 2.9). For the data points with 206Pb/204Pb > 395 (line 6), no residual scatter is observed (MSWD < 1). This data set provides sufficient spread in206Pb/204Pb ratios for precisely constraining the slope and intersection of the isochron and yields the most precise age of 4564.7 ± 0.6 Ma. Further elimination of relatively less radiogenic points, e.g., to 206Pb/204Pb > 500 (line 7), decreases scatter to an even lower value, but the uncertainty of the isochron slope and intercept increases because of the smaller spread in 206Pb/204Pb. This illustrates a fundamental limitation of the isochron model—the controversial requirement of sufficient spread of data points while maintaining the uniformity of initial Pb composition.

View this table:

The Pb-Pb isochrons for the Efremovka CAIs (Table 2, Fig. 1) show no excess scatter and give consistent dates of 4567.17 ± 0.70 Ma (MSWD = 0.88) for all six analyzed acid-washed fractions from the CAI E49, and 4567.4 ± 1.1 Ma (MSWD = 1.09) for all 12 fractions from the CAI E60. The weighted average of these two dates gives the best estimate for the timing of the CAI formation of 4567.2 ± 0.6 Ma. This age agrees, within error, with the previous Pb-Pb age determinations for the CAIs from the CV chondrite Allende (27, 28, 51, 52). However, the age of the Efremovka CAIs obtained here is much more precise and is clearly resolved from the age of the Acfer 059 chondrules.

The 27Al/24Mg ratios and Mg isotopic compositions of spinel, pyroxene, and anorthite from the CAI E60 (Table 3), determined by in situ ion microprobe analysis (53), define a line with the slope of (4.63 ± 0.44) × 10−5 on the Al-Mg evolution diagram (Fig. 2), which is indistinguishable from the canonical (26Al/27Al)I value (9).

Figure 2

Al-Mg evolution diagram for the Efremovka CAI E60. Error crosses are 2σ. Isochron initial26Al/27Al error is a 95% confidence interval.

Table 3

Al-Mg isotopic data for the CAI E60.

View this table:

Before applying the Pb-Pb dates to constrain the interval of CAI and chondrule formation, we need to check whether these dates are primary ages, or if they were affected by secondary processes. The possibility that the Pb isotopic system was reset by diffusion during thermal metamorphism can be evaluated by using the Pb diffusion parameters (54) and the closure temperature estimates previously applied to equilibrated ordinary chondrites (46). The closure temperature calculations show that the diffusion of Pb in pyroxene (presumably the main carrier of U) at 150°C (the peak temperature reached during aqueous alteration of CR chondrites) is slow and can be ruled out as a cause of Pb isotopic resetting. Mineralogical and isotopic evidence for the pristine nature of the Acfer 059 chondrules (34) and CAIs (35, 36) suggests that resetting due to aqueous alteration is also very unlikely. The date of 4564.7 ± 0.6 Ma should therefore correspond to the timing of chondrule formation. The pristine mineralogy of the Efremovka CAIs E60 and E49 and 26Al-26Mg systematics of E60 suggest that the data of 4567.2 ± 0.6 Ma corresponds to the timing of CAI formation.

The data presented here have several implications. Small errors of chondrule Pb-Pb isochron regression and the lack of excess scatter among the most radiogenic data points suggest that the chondrules formed within the short interval, probably less than the isochron error limits of 1.2 My. The available 26Al-26Mg data for the chondrules from CR chondrites (12) are not sufficiently precise to verify this inference.

Combining the age of the Acfer 059 chondrules with the age of the CV CAIs gives an interval of 2.5 ± 1.2 My between formation of the CV CAIs and CR chondrules, which indicates that CAI- and chondrule-forming events in the solar nebula continued for at least 1.3 My (55). This time interval is consistent with a 2- to 3-My age difference between the CR CAIs and chondrules inferred from the reported differences in their (26Al/27Al)I (12,37). Together, these observations support the chronological significance of 26Al-26Mg systematics (9) and are inconsistent with a local origin of 26Al by energetic particle irradiation (19). The inferred 2- to 3-My age difference between CAIs and chondrules in CR chondrites (12, 36), which has yet to be confirmed by Pb-Pb dating of CAIs in CR chondrites, is inconsistent with the contemporaneous formation of CAIs and chondrules inferred in the X-wind model (4). The obtained estimate of the CAI-chondrule–formation interval of 2.5 ± 1.2 My is within the range of the X-wind (4), jet flow (5, 6), and shock-wave models (7, 8) of chondrule formation. However, the generation of strong chondrule-forming shocks for a period as long as 1 to 3 My could be a problem for the shock-wave model associated with the gravitational instability of the protoplanetary disk (8), which could have occurred only very early in the proto-Sun's evolution (56).

Supporting Online Material

Table S1

  • * Present address: Geological Survey of Canada, 601 Booth Street, Ottawa, ON K1A 0E8, Canada.

  • To whom correspondence should be addressed. E-mail: yamelin{at}


View Abstract

Navigate This Article