Evidence from Rb-Sr Systematics for a 4.25 Ga disturbance in Al3S4, an Allende CAI. N.E. Marks1, L.E. Borg1, I.D. Hutcheon1, R.N. Clayton2, T.K. Mayeda2. 1Institute of Geophysics & Planetary Physics, Lawrence Livermore National Laboratory, 7000 East Ave. L-231 Livermore CA 94550 (marks23@llnl.gov), 2Enrico Fermi Institute, The University of Chicago, 5640 S Ellis Ave, Chicago, IL 60637, USA.

Introduction: Calcium-Aluminum-rich inclusions

(CAIs) are the oldest solids formed in the protoplane-tary disk that surrounded the young sun and conse-quently contain many clues to the nebular time period of the solar system. The decay of radiogenic 87Rb to stable 87Sr with a half-life of 49.4 Ga provides the ba-sis for the long-lived Rb-Sr geochronometer. This geochronometer has been applied to CAIs to make initial Sr isotopic ratio determinations, which have significance both as a reference point for Sr isotope evolution in terrestrial materials and as a basis for model age determinations used for estimating the rela-tive times of condensation of solar system bodies.

Although a number of studies [1-3] have demon-strated the extreme primitive (unradiogenic) 87Sr/86Sr ratio of some Allende CAIs, they have not defined whole rock or internal mineral isochron ages for the inclusions owing to apparent disturbance of the Rb-Sr systematics. Not even sequential leaching of bulk CAIs has been successful in producing an Rb-Sr iso-chron. This study reports the results of the application of modern mass spectrometry techniques to minerals separated from a type B CAI from Allende. These data facilitate further application of this long-lived chro-nometer to early solar system process and provide new insight into the timing of secondary events on the Al-lende parent body.

Figure 1. Photomicrograph of Al3S4, a coarse grained type B inclusion from Allende meteorite. The outer 1.5 mm border consists of radially oriented melilite crystals, with increasing spinel content inwards. The interior of the inclusion is primarily crystalline clinopyroxene, melilite, and anorthite with spinel inclusions. FOV is 22 mm X 17 mm. From [4].

Methods: Allende type B CAI Al3S4 was sepa-

rated into constituent phases by a combination of mag-

netic and heavy liquids separation and hand-picking at the University of Chicago [4]. Figure 1 is a photomi-crograph of Al3S4 from which the mineral separates were derived [4]. The five separates studied include pyroxene, pyroxene + spinel, melilite, and density separates above and below 3.26 g/cm3. The five min-eral separates were weighed and loaded into Teflon bombs and digested in concentrated HF+HNO3 at LLNL. Residual material undigested in the bombs was then treated with aqua regia and underwent repeated dry downs. Following dissolution, 95% aliquots were spiked with 87Rb and 84Sr in order to determine accu-rate concentrations of Rb and Sr in the individual min-eral fractions. The use of large mineral fractions mini-mized the relative contribution of analysis blank to the ultra-low level Rb analyses associated with CAIs. The remaining 5% aliquots were used for Sr isotope composition measurements. This procedure effec-tively lowered analytical uncertainty on the Sr isotope measurements associated with uncertainties in Sr spike composition by negating the need for spike deconvolu-tion. Rb-Sr and Sm-Nd were separated by standard cation ion-exchange chromatography in HCl. Blanks are 46 pg for Rb and 25 pg for Sr. Mass spectrometry was performed on a ThermoFisher Triton TIMS at LLNL. The running average measured 87Sr/86Sr for NBS-987 was 0.710249±8 (2 stdev; n=64). Data from [1-3] were adjusted based on our measured value of NBS-987 for comparison.

Rb-Sr Systematics of Al3S4: Results of Rb-Sr isotopic analyses for the Al3S4 CAI are presented in an isochron plot in Figure 2. A linear regression fitted through all five points yields an apparent age of 4246 ± 110 Ma and an initial 87Sr/86Sr of 0.6989419 ± 0.0000076. Despite the relatively large uncertainty associated with the Rb-Sr age, the regression repre-sents a very good fit of the data as indicated by a MSWD of 1.3. Thus, the uncertainty in the age does not reflect a lack of linearity, rather it reflects our in-ability to measure Sr to better than ~10 ppm (2 sigma) and the limited spread in 87Rb/86Sr ratios resulting from very low Rb abundances in the CAI minerals. Ratios of 84Sr/86Sr were also measured on the unspiked runs and are very close to standard values, suggesting the isotopic composition of Sr was not anomalous (i.e. contained variable r and s-process components), as recently reported for some Allende CAIs [5].

9141.pdfFormation of the First Solids in the Solar System (2011)

Figure 2. Rb-Sr isochron plot for Allende CAI Al3S4 yielding an age of 4246 + 110 Ma. Error bars are 2σ .

Discussion: The apparent age calculated from the linear regression is much younger than Pb-Pb and Al-Mg ages 4567+0.93 Ma measured in Allende CAIs, e.g. [6,7]. The high degree of linearity of the isochron, however, implies that the Rb-Sr data record a physical event. There are two possible scenarios that can ac-count for the young age: 1) the age represents a mixing line with recent terrestrial contamination, or 2) the age represents radiogenic ingrowth of Sr following isotopic disturbance at ~4246 Ma. In both cases, radiogenic growth of Sr in the CAIs must be minimal in order to preserve the linearity of the regression. These scenar-ios are explored below.

The first possibility is that the regression represents recent terrestrial contamination of the CAI mineral fractions. The array of data is linear because individ-ual mineral fractions are characterized by 87Rb/86Sr ratios near zero and, consequently, have Sr isotopic compositions near the solar system initial value. In this case very small but variable amounts of a high Rb/Sr terrestrial contaminant would shift individual mineral fractions along the apparent isochron, and the regression would not reflect an age. This scenario re-quires a contaminant that fortuitously lies on the 4.25 Ga isochron. This possibility is further complicated by the observation that the mineral separates each lie along the isochron plot in order of igneous partitioning (i.e. melilite closest to the origin, next pyroxene, and the densest, spinel-containing mineral separates fur-thest from the origin). Thus, the proportion of con-taminant in each mineral fraction must increase pro-portionally from melilite to pyroxene to spinel. Note that there is essentially no correlation between the po-sition of a mineral fraction on the isochron and its Sr concentration, implying a complicated mixing proc-esses to preserve the linearity of the regression.

A second possibility is that the age reflects radio-genic decay of Rb (ingrowth of 87Sr) added to the CAI

minerals after they formed. In order to maintain line-arity, the minerals must have little or no Rb. In this case, the Rb-Sr age reflects the timing of alteration. In the simplest case, this mixing would result in a two-stage evolution, starting by shifting the minerals to higher Rb/Sr ratios, followed by subsequent ingrowth of radiogenic 87Sr. The 4246 Ma isochron would then represent the earliest possible timing of this Rb addi-tion.

One difficulty with this interpretation is how to preserve the “old” ages of CAIs recorded by other iso-topic systems, while resetting the Rb-Sr system. This scenario requires the Rb contamination to be located in a physical reservoir in the CAIs that is distinct from the reservoir that preserves the Al-Mg and Pb-Pb chro-nometers, e.g. [1,2,6,7]. One mechanism that could account for this distribution is to concentrate Rb in a secondary alteration product preferentially sited in interstitial spaces between minerals. Alternatively, Rb diffused into the outer few microns of grain boundaries would satisfy the requirements of this model. In these scenarios, the 26Al-26Mg chronology would be pre-served if the minerals were not re-equilibrated. Like-wise, Pb concentrations (and isotopic systematics) of bulk minerals need to be minimally effected by this alteration. If Pb contamination was concentrated in easily mobilized reservoirs, along grain boundaries for example, it might be efficiently removed in the wash-ing and leaching steps associated with Pb-Pb chrono-metry.

Ca-Al rich inclusions are among the earliest solar system condensates, and Tatsumoto et al. [2] suggested that Rb/Sr ratios increased at the time of agglomeration of CAIs with their volatile-rich matrix. Subsequent addition of Rb through interaction with chondrite ma-trix and alteration could explain why other authors have observed young (3.8-4.2 Ga) model ages for Al-lende CAIs, e.g. [2,3].

References: [1] Podosek, F. A. et al. (1991) GCA, 55,

1083–1110. [2] Tatsumoto, M. et al. (1976) GCA, 40, 617-634. [3] Gray, C. M. et al. (1973) Icarus, 20, 213-239. [4] Clayton, R. N. et al. (1977) EPSL, 34, 209-43. [5] Bren-necka, G. A. et al. (2010) LPSC 42, #1302. [6] Connelly, J. N. et al. (2008) Astrophys. J. 675, 121-124. [7] Jacobsen, B. J. et al. (2008) EPSL. 272, 353-364. This work performed under the auspices of the U.S. DOE by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

melilite  

d>3.26  

pyroxene  d<3.26  

pyroxene  +  spinel  

0.69895  

0.69900  

0.69905  

0.69910  

0.69915  

0.69920  

0.69925  

0.69930  

0.69935  

0.69940  

0.69945  

0.000   0.002   0.004   0.006   0.008  

87Sr/86Sr  

87Rb/86Sr  

Allende  CAI  Al3S4  

Al3S4  Solution  on  5  points  Age  =  4246  ±  110  Ma  Initial  87/86=0.6989419  ±  0.0000076  MSWD  =  1.3,  Probability  =  0.27  

9141.pdfFormation of the First Solids in the Solar System (2011)