Hydrothermal processing on planetesimals in the early solar system produced new mineral phases, including those generated by the transformation of anhydrous silicates into their hydrated counterparts. Carbonaceous chondrites represent tangible remnants of such alteration products. Lithium isotopes are known to be responsive to aqueous alteration, yet previously recognized variability within whole rock samples from the same meteorite appears to complicate the use of these isotopes as indicators of processing by water. We demonstrate a new way to use lithium isotopes that reflects aqueous alteration in carbonaceous chondrites. Temperature appears to exert a control on the production of acetic acid-soluble phases, such as carbonates and poorly crystalline Fe-oxyhydroxides. Temperature and degree of water-rock interaction determines the amount of lithium isotope fractionation expressed as the difference between whole rock and acetic acid-leachable fractions. Using these features, the type 1 chondrite Orgueil (δ7Li(whole rock) = 4.3‰; Δ7Li(acetic-whole) = 1.2‰) can be distinguished from the type 2 chondrites Murchison (δ7Li(whole rock) = 3.8; Δ7Li(acetic-whole) = 8.8‰) and carbonate-poor Tagish Lake (δ7Li(whole rock) = 4.3; Δ7Li(acetic-whole) = 9.4‰). This initial study suggests that lithium isotopes have the potential to reveal the role of liquid water in the early solar system.
About 85% of the meteorites observed to fall on the Earth are chondrites (Bischoff and Geiger 1995). The most primitive chondrites are termed “carbonaceous” and are rocks with an overall elemental composition similar to the Sun if the loss of the most volatile elements is ignored.
Carbonaceous chondrites record relatively low-temperature alteration processes that occurred on planetesimals (Krot et al. 2006). The action of heat and water led to mineral transformations on the asteroidal meteorite parent body. Recently published work indicates that a lack of fluid flow meant that alteration can be heterogeneous and reliant on local starting materials (Howard et al. 2011). Different chondrite groups reflect varying types and levels of secondary processing that are dependent on primary composition, local temperatures, water/rock ratios, redox conditions, fluid compositions, and combinations thereof (Krot et al. 2006). Based on variations in their mineralogical and textural composition, they are assigned to a petrologic type (Van Schmus and Wood 1967). It is the products of secondary processing in chondrites that are implicated in the delivery of water and organic matter as a late veneer to the early Earth (Newsom and Jones 1990).
Lithium isotopes (7Li, 6Li) provide a geochemical proxy that has the potential to reveal the role of liquid water on asteroidal meteorite parent bodies in the early solar system. Source regions for lithium have been tabulated and the characteristic isotope ratios indicated (Sephton et al. 2004). Within chondrites, although lithium exists in various components, the majority resides in the matrix (Maruyama et al. 2009). The high vibrational frequency of water relative to mineral phases dominates lithium isotope fractionation (Huh et al. 2001). Consequently, as mineral-water interactions proceed, 6Li is preferentially retained in the solid phase while 7Li passes into solution. Corroborating data from the Murchison meteorite reveal relatively high 7Li contents in the aqueously generated phyllosilicate-rich matrix (δ7Li = 12.8‰) and acetic acid-soluble phases such as carbonates and poorly crystalline Fe-oxyhydroxides (δ7Li = 6.0‰) relative to unaltered chondrules (δ7Li = −1.2‰) (Sephton et al. 2004). Preaccretionary origins have been postulated for both phyllosilicates and carbonates (e.g., Metzler et al. 1992) but the progressive increase in 7Li with putative degrees of exposure to liquid water suggests a parent body origin for the majority of these materials in chondrites (Sephton et al. 2004). The lithium isotope approach has also been shown to be useful for revealing the aqueous history of enigmatic clasts in carbonaceous chondrites (Sephton et al. 2006). One outstanding feature of lithium isotopic signatures is that they are not affected by subsequent heating and dehydration of the aqueously altered minerals suggesting potential for a robust proxy that resists corruption by subsequent thermal events (McDonough et al. 2003; Magna et al. 2011). Whole-rock lithium isotope ratios for carbonaceous chondrites show some variability. Intermeteorite differences have been attributed to varying amounts of 7Li-rich mineral phases resulting from parent-body aqueous alteration (McDonough et al. 2003). Yet, in a recent study, multiple chips from the same meteorites displayed a range of lithium isotope ratios indicating significant intrameteorite isotopic heterogeneity (up to δ7Li 1.5‰) and no overall relationship could be recognized between lithium isotopic composition and petrologic type within the carbonaceous and ordinary chondrites (Pogge von Strandmann et al. 2011). Seitz et al. (2012), however, suggest that differences in the lithium isotope composition of bulk carbonaceous and ordinary chondrites may be attributed to mixing of chondrules, calcium aluminum inclusions, and matrix in different proportions. The efficacy of lithium isotope ratios as indicators of meteorite alteration is currently uncertain. In this article, we assess new means of using lithium isotopes as indicators for parent body processing. For the first time, lithium abundances and isotope ratios have been obtained from whole rock and leachates in a number of carbonaceous chondrites. The findings may help to reveal the influence of aqueous processing on lithium isotope systematics in the early solar system.
The meteorite samples Orgueil (fall, 1864) and Murchison (fall, 1969) were obtained from the Natural History Museum, UK. Pristine Tagish Lake (fall, 2000) was provided by Dr. Mike Zolensky, NASA JSC. As falls, the meteorites have had no prolonged contact with liquid water since arrival on the Earth, precluding extensive terrestrial weathering effects on their lithium isotope composition. Despite the relative lack of terrestrial weathering in falls it must be noted that some alteration is possible during storage in scientific collections (e.g., Velbel and Palmer 2011). Samples were analyzed as whole rocks and leachates. Whole rock samples (approximately 5 mg) were finely ground using an agate pestle and mortar and dissolved in 1 mL of 15 N thermally distilled (TD) HNO3 and 4 mL TD HF, on a hotplate (130 °C, 24 h). This mixture was evaporated to incipient dryness, and 2 mL of 15 N TD HNO3 was added and refluxed (130 °C, 24 h), dried down, and then redissolved in 100 μL of 0.2 N TD HCl. Prior to leaching, loosely bound salts were removed from a finely ground approximately 30 mg subsample by ultrasonication with 1 mL of TD H2O (30 min) and centrifugation (×3). Lithium released from acetic-acid-soluble phases (carbonates, and poorly crystalline Fe-oxyhydroxides: Chan and Hein 2007; Wimpenny et al. 2010) was then obtained using 2 mL of 1 N TD acetic acid heated on a hotplate at 50 °C for 24 h, followed by centrifugation and washing in TD H2O (×3). The leachate and the wash solutions were combined before evaporation to dryness and dissolution in 100 μL 0.2 N TD HCl. Lithium was then separated from the sample matrix on a cation-exchange column (James and Palmer 2000).
Lithium isotope ratios were measured using a Nu Instruments multiple collector inductively coupled plasma mass spectrometer at The Open University, UK. Individual values represent a mean of 20 background-corrected ratios. δ7Li values for the whole rock data were obtained by measurements on a number of subsamples. δ7Li values for the acetic acid leachates were obtained by measurements on more than one occasion using the same subsample. Analyses were bracketed by measurement of the L-SVEC standard (Flesch et al. 1973). Internal precision on 7Li/6Li measurements is <±0.15‰ (2σ) for a 10 ppb solution and is <±0.65‰ (2σ) for a 2 ppb solution. External precision is ±0.66‰ (2σ), based on multiple (n = 15) analyses of the GSJ JB-2 basalt standard reference material. Data are expressed as δ7Li values (in ‰) relative to the L-SVEC standard as follows:
Results and discussion
Whole Rock Values
Orgueil is a type 1 chondrite (CI), which has experienced extensive aqueous alteration causing most olivines and pyroxenes to be altered to hydrous phases (Du Fresne and Anders 1962; Zolensky et al. 2008). Murchison is a type 2 chondrite (CM), in which some anhydrous minerals remain unaltered, reflecting aqueous alteration at lower water to rock ratios and/or different temperatures (Clayton and Mayeda 1984, 1999; Zolensky et al. 1993, 2008; Leshin et al. 1997, 2001; Guo and Eiler 2007). Tagish Lake has characteristics similar to both type 1 CI and type 2 CM chondrites and is designated an ungrouped type 2 (Zolensky et al. 2002). Despite sharing a petrologic type with Murchison, Tagish Lake is distinct with features including relatively high concentrations of carbon, presolar grains, and carbonate, all of which suggest that it is derived from a different parent body with discrete starting materials (Hiroi et al. 2001).
Whole rock lithium isotope data for Orgueil and Murchison are presented in Table 1 and appear consistent with previous work (James and Palmer 2000; Sephton et al. 2004; Pogge von Strandmann et al. 2011) which indicates that Orgueil is slightly enriched in 7Li compared to Murchison. These data are consistent with previously published results that reveal a trend of increasing 7Li/6Li ratios with increasing parent-body aqueous alteration (McDonough et al. 2003). The trend has been interpreted as reflecting greater amounts of phyllosilicates resulting from interaction of 7Li-rich waters with anhydrous minerals on the meteorite parent body (McDonough et al. 2003). Later work identified significant variability within individual meteorites using whole rock carbonaceous, ordinary, and enstatite chondrites (Pogge von Strandmann et al. 2011), casting doubt on the veracity of such interpretations. However, more recent work suggests that the intrameteorite differences are more likely to result from mixing of chondrules, calcium aluminum inclusions, and matrix in different proportions (Seitz et al. 2012).
Table 1. Lithium concentration and lithium isotope data for the Orgueil, Tagish Lake, and Murchison carbonaceous chondrites
Tagish Lake whole rock data (Table 1; Fig. 1) display variations between carbonate-rich and carbonate-poor portions of the meteorite; the carbonate-poor sample is enriched in 7Li. Initially, these data appear counterintuitive as carbonates will precipitate from water with high δ7Li (e.g., Marriott et al. 2004). It might be expected therefore that the presence of carbonate indicates more extensive aqueous alteration, more opportunities for water-rock interaction, and a greater enrichment in 7Li. However, if published data for carbonate carbon in Murchison and Orgueil are used as examples, the less aqueously processed type 2 Murchison contains greater amounts of carbonate (mean = 255 ppm, n = 10) than the more altered type 1 Orgueil (mean = 136 ppm, n = 2) although sample heterogeneity is an issue (Grady et al. 1988). Hence, carbonate concentration does not correlate simply with degree of parent-body aqueous alteration.
If whole rock δ7Li values are used as indicators of aqueous processing (McDonough et al. 2003) it can be inferred that distinct alteration processes, as reflected in carbonate-poor and carbonate-rich portions of Tagish Lake, were occurring on the same parent body in close proximity. Small scale heterogeneity in the preterrestrial alteration history of Tagish Lake would be consistent with previously published reports on mineralogical (Zolensky et al. 2000), organic, and stable isotopic (Herd et al. 2011) variation between subsamples of this meteorite. However, the intrameteorite differences of whole rock values and apparently poor correlation of whole rock lithium isotope compositions with petrologic type (Pogge von Strandmann et al. 2011) must add caution to this interpretation.
Leachate Versus Whole Rock Values
Previous work on the Murchison meteorite has established differences in lithium isotope ratios between minerals produced by aqueous alteration and their putative anhydrous mineral precursors (Sephton et al. 2004). Primary minerals are predominantly silicates (Krot et al. 2006) and aqueous alteration on the meteorite parent body generated a number of low-temperature minerals including carbonates, hydroxides, and oxyhydroxides; phyllosilicates are also a major alteration phase in type 2s (Zolensky et al. 1993).
The unaltered materials and secondary alteration products in aqueously processed meteorites can be isolated using selective leaching methods. The primary anhydrous silicates are immune to acetic acid leaching (Singleton and Lavkulich 1987) and it is presumed therefore that there were no acetic acid-leachable phases present before aqueous processing on the meteorite parent body. Carbonates and poorly crystalline Fe-(oxy)hydroxides are leached by acetic acid and the leach process is relatively specific to these secondary minerals (Berger et al. 2008; Wimpenny et al. 2010).
The meteorites studied here differ in the concentration of lithium in acetic acid leachates (Table 1). The type 1 chondrite Orgueil has almost an order of magnitude less lithium in its acetic acid leachate than the type 2s, Murchison and Tagish Lake. One possible explanation is that the minerals produced in the specific alteration scenarios for the different meteorites associate with lithium to different extents. The production of oxyhydroxides requires low concentrations of oxygen dissolved in water and it is temperature that controls the aqueous solubility of oxygen. Above approximately 60 °C magnetite formation is preferred while below this temperature oxyhdroxides are favored (Olowe et al. 1991), and thermodynamic modeling indicates that this is true for the range of applicable Eh (and pH) values. Leaching with acetic acid will not attack magnetite (e.g., Berger et al. 2008), so the data can be explained by type 2 chondrites being altered at relatively low temperatures producing acetic acid-leachable oxyhydroxides which are associated with lithium (Kim et al. 2008). The suggestion that variations in iron mineralogy act as a control on lithium concentration in acetic acid leachates are supported by substantial evidence of abundant magnetite in Orgueil (e.g., Kerridge and Chatterji 1968).
The data support previous assertions (Sephton et al. 2004; Seitz et al. 2012) that whole rock values are averages of a number of components with distinct lithium isotopic compositions (Fig. 1). Although the meteorite samples in this study differ in whole rock values by only 1‰ in δ7Li, individual fractions display a range close to 10‰ (Table 1). These observations are consistent with previous work where a spread in Li isotope compositions of over 14‰ was observed for Murchison with acetic acid-leachable phases representing the most isotopically heavy component followed by the phyllosilicate matrix and, finally, the anhydrous chondrules being the most isotopically light (Sephton et al. 2004).
The type 1 chondrite Orgueil displays much less difference in δ7Li between acetic acid leachate and whole rock (Δ7Liacetic-whole) relative to the type 2 chondrites, Tagish Lake, and Murchison (Table 1). The fractionation of lithium isotopes is partly temperature-dependent (Vigier et al. 2008) and the isotopic-differential between phases in a sample may reflect the temperature of alteration. Suggested aqueous alteration temperatures on the type 1 CI chondrite parent body occupy a range of values (<50 °C, Leshin et al. 1997; 50–150 °C, Zolensky et al. 1993; 150 °C, Clayton and Mayeda 1984, 1999). Inferred alteration temperatures for the type 2 CM chondrite parent body also occupy a range extending between 20 and 71 °C (Guo and Eiler 2007) which is somewhat similar to the temperature range suggested for the type 2 Tagish Lake parent body (50–100 °C, Leshin et al. 2001). Assuming that the type 1 Orgueil was subjected to higher temperatures than the type 2s Tagish Lake and Murchison, fractionation of isotopes would have been less pronounced at higher temperatures (e.g., Vigier et al. 2008) and the smaller Δ7Liacetic-whole values for Orgueil would be consistent with a temperature effect. Correlating lithium isotope fractionation to temperature will inevitably work best for closed systems and variations in whole rock values may indicate some open system behavior. Temperature effects may still be recognizable in partly open systems, as implied by terrestrial studies (James et al. 1999). Detailed mass-balance studies by analysis of all meteorite constituents would be helpful to appreciate how closed or open the meteorite lithium isotope system was during alteration although such an investigation would require relatively large samples that are rarely available for precious meteorites.
Hence, our data suggest that Δ7Liacetic-whole values provide a mechanism for determining past aqueous processes, with values for the type 1 and type 2 chondrites analyzed in this study differing by almost an order of magnitude. The correlation of Δ7Liacetic-whole values with putative levels of parent body processing strongly suggests that this internal heterogeneity is produced by alteration (Sephton et al. 2004). Common degrees of lithium isotopic fractionation for individual petrologic types imply that temperature may be an important factor.
There are substantial differences in the lithium concentrations in acetic acid-leachable phases in the type 1 and type 2 chondrites analyzed here. The most plausible explanation is that different temperatures of alteration have led to the production of crystalline Fe-oxides, such as magnetite, in the type 1 chondrite Orgueil and poorly crystalline Fe-oxyhydroxides in the type 2 chondrites Murchison and Tagish Lake. Crystalline Fe-oxide phases are not attacked by acetic acid.
Whole rock δ7Li values vary between petrologic types with type 1s such as Orgueil displaying slightly more positive values than type 2s such as Murchison. If whole rock δ7Li values can be used as indicators of aqueous processing the data are consistent with different levels of alteration, as reflected in carbonate-poor and carbonate-rich portions of the type 2 ungrouped Tagish Lake, were occurring on the same parent body in close proximity.
Our data allow the first comparison of δ7Li measurements of whole rock and leachate from more than one meteorite. The type 1 chondrite Orgueil displays much less difference in δ7Li between whole rock and leachate relative to the type 2 chondrites, Tagish Lake, and Murchison. The fractionation of lithium isotopes is, to some extent, temperature-dependent and the isotopic-differential between phases in a sample may reflect the temperature of alteration. The lithium isotopic differential between whole rock and acetic acid leachate appears to be an effective indicator of parent body alteration and, possibly, petrologic type.
The authors are grateful for the valuable inputs of Martin Lee and two anonymous reviewers along with the detailed recommendations of editor Tim Jull.