Report

A Younger Age for ALH84001 and Its Geochemical Link to Shergottite Sources in Mars

See allHide authors and affiliations

Science  16 Apr 2010:
Vol. 328, Issue 5976, pp. 347-351
DOI: 10.1126/science.1185395

Abstract

Martian meteorite ALH84001 (ALH) is the oldest known igneous rock from Mars and has been used to constrain its early history. Lutetium-hafnium (Lu-Hf) isotope data for ALH indicate an igneous age of 4.091 ± 0.030 billion years, nearly coeval with an interval of heavy bombardment and cessation of the martian core dynamo and magnetic field. The calculated Lu/Hf and Sm/Nd (samarium/neodymium) ratios of the ALH parental magma source indicate that it must have undergone extensive igneous processing associated with the crystallization of a deep magma ocean. This same mantle source region also produced the shergottite magmas (dated 150 to 570 million years ago), possibly indicating uniform igneous processes in Mars for nearly 4 billion years.

The orthopyroxene cumulate ALH84001 is unique in having an igneous crystallization age more than 2 billion years older than any other martian meteorite (111), providing critical timing constraints on the formation of Mars. Its igneous crystallization age has been difficult to determine because of its complex postcrystallization history of aqueous alteration and shock metamorphism (12, 13). Although crystallization ages obtained on ALH range from 3.9 to 4.56 billion years ago (Ga) (110), the weighted average of Rb-Sr and Sm-Nd isotope data (2) yield a generally accepted age of 4.51 ± 0.11 Ga (all errors are ±2σ) (11). This age has important implications for the formation history of martian crust and mantle, crater chronology, and the onset of the Mars magnetic field. An age of 4.51 ± 0.11 Ga requires that a stable crust existed very early in Mars’ history, not long after the onset of solar system formation at 4.568 Ga (14). If ALH represents crustal material formed at 4.51 Ga, then this crust must have survived a period of intense bombardment between 4.25 and 4.10 Ga (15, 16), similar to the late heavy bombardment or terminal lunar cataclysm on the Moon, without suffering the intense brecciation observed in prebombardment lunar rocks. A further implication of this age is the potential presence of a magnetic field and core convection in Mars 27 to 48 million years after core formation (17, 18). In order to better refine the timing of these events and to evaluate potential magmatic source affinities, we applied 176Lu-176Hf and 146,147Sm-142,143Nd chronometry to ALH.

ALH is an igneous cumulate rock that has been affected by shock-induced metamorphism and precipitation of secondary phases including carbonate and magnetite (13, 19). ALH is composed of mosaic-grained orthopyroxene (~97 vol %) crystals, 4 to 5 mm in diameter, that enclose euhedral to subhedral chromite (~2 vol %) locally preserved as clasts within recrystallized granular bands. Interstitial to the mosaic-grained orthopyroxene are chromite, maskelynite, phosphate (apatite and merrillite), augite, olivine, and SiO2 (which total ~1 vol %). The 3.6 g of ALH used in this study (sample allocations ALH84001, 365 and ALH84001, 403) have an exceptionally well-preserved igneous texture with no optical evidence of secondary phases and minimal (5 to 10 vol %) granular banding. One portion was separated into oxide (~100% chromite, sample S1) and nearly pure orthopyroxene-rich (sample S2) fractions. Two other fractions—representing the bulk rock (sample S3) and a bulk fraction after chromite removal (sample S4)—are from the disaggregated and processed material of this portion. A second portion was leached with 2.5 M HCl, resulting in a leachate-residue pair (samples L1 and R1) that was analyzed for 147Sm-143Nd. A third portion was crushed into a powder for high-precision 142Nd/144Nd analysis. Sample preparation, chemical separation, and analytical procedures are described in (20).

A 176Lu-176Hf isochron age of 4.091 ± 0.030 Ga for ALH is defined by samples S1 to S4 (Fig. 1). This age is consistent with 207Pb-206Pb ages of 4.074 ± 0.099 Ga (4) and 4.135 ± 0.012 Ga (5) and a U-Pb age of 4.117 ± 0.003 Ga (5). The Lu-Hf age is substantially younger than the respective Sm-Nd and Rb-Sr ages of 4.50 ± 0.13 Ga and 4.55 ± 0.30 Ga (2, 11). The measured Lu-Hf, Pb-Pb, and U-Pb ages are similar to but slightly older than the respective Rb-Sr and Pb-Pb ages of secondary carbonate phases, 3.90 ± 0.04 Ga and 4.04 ± 0.10 Ga (9), as well as the U-Pb age of whitlockite and apatite, 4.018 ± 0.081 Ga (10). The 39Ar-40Ar ages of 3.92 ± 0.10 Ga and 4.1 ± 0.2 Ga (6, 8) are similar to the Lu-Hf age, but shock disturbances, trapped atmospheric components, and 39Ar recoil make comparisons between these isotope systems difficult. The 147Sm-143Nd data of samples S2 to S4 indicate an “age” of 4.405 ± 0.026 Ga, ~315 million years older than the Lu-Hf age and consistent with the previous Rb-Sr and Sm-Nd age determinations (2, 11). The 147Sm-143Nd age of 4.889 ± 0.020 Ga defined by samples L1 and R1 is older than the age of the solar system and has an initial 143Nd/144Nd less than the solar system initial value (fig. S1). The counterclockwise rotation of these older isochrons requires either open-system behavior of Sm and Nd or decoupling of Sm-Nd isotope systematics during subsolidus alterations.

Fig. 1

Lu-Hf isochron for ALH. Inset shows Hf isotope data relative to CHUR at 4.091 Ga. Data are listed in Table 2; errors are 2σ. The isochron was calculated using IsoPlot (34) and a 176Lu decay constant of 1.865−11 year−1 (35).

In ALH, 58% and 78% by weight of Sm and Nd, respectively, reside in the phosphate phases whitlockite and apatite (Table 1). The Sm-Nd apparent isochron defined by samples S2 to S4 is controlled by distribution of light rare earth element (LREE)–rich phosphates, as shown by the strong correlation (R2 = 0.99998) between measured 143Nd/144Nd and Nd concentrations for each fraction (fig. S2). Because phosphates are reactive in low-pH and weathering environments, there was great potential for redistribution of Sm, Nd, and Sr in ALH resulting in spurious age and initial isotope ratios for both the Sm-Nd and Rb-Sr isotope systems. The relatively high scatter of 232Th-208Pb data (5) and a younger average U-Pb age of 4.018 ± 0.081 Ga (10) of apatite and whitlockite are consistent with disturbance of phosphate phases after igneous crystallization. For counterclockwise rotation of a Sm-Nd isochron, the low-Sm/Nd phase, phosphate in this case, must lose Nd relative to Sm, and the high-Sm/Nd phase orthopyroxene must have its Sm/Nd ratio decreased or unchanged. The precise mechanism of this alteration is unknown, but disturbance of the Sm-Nd system in ALH is demonstrated by the ages being too old for the leachate-residue pair coupled with a 143Nd/144Nd less than the solar system initial value. In contrast, orthopyroxene and chromite are the dominant Lu and Hf reservoirs (Table 1), which implies that phosphate disturbance will not affect the Lu-Hf system. Although 3% Lu resides in phosphate, its mobility would have little impact on the slope of the isochron. Because of its resistance to disturbance resulting from phosphate alteration and mutual agreement between U-Pb and Pb-Pb ages, we conclude that the Lu-Hf isochron defines the true igneous crystallization age.

Table 1

Lu, Hf, Sm, and Nd concentrations and their calculated absolute amounts in each listed phase. See (20) for data reduction and analytical procedures. Modal proportions from (19); plagioclase and phosphate data from (36); orthopyroxene and chromite data from this study (20).

View this table:

The Lu-Hf age of 4.091 ± 0.030 Ga precludes it from being a remnant of primordial crust formed during solidification of an early martian magma ocean (MO). Instead, the magma that crystallized to produce ALH was derived from mantle reservoirs that have evolved isotopically since about 4.54 to 4.46 Ga (21). The initial 176Hf/177Hf of ALH cast in ε176HfCHUR notation (22) is –4.64 ± 1.04 (Table 2). This indicates that the 176Hf/177Hf of its parental magma source is less than CHUR (Fig. 2A). The calculated source 176Lu/177Hf—assuming a two-stage mantle evolution model (23), a source formation age of 4.513 Ga, a chondritic bulk Mars, and a Mars formation age of 4.567 Ga (20)—is 0.0183 ± 0.0036. This is similar to but substantially lower than the average source 176Lu/177Hf of 0.02795 ± 0.00010 (weighted mean) calculated for the distinct martian meteorite group of enriched shergottites (24) (Fig. 2A). The calculated source 147Sm/144Nd using the bulk-rock sample S3 is 0.172, slightly lower than the average source 147Sm/144Nd of 0.185 for the enriched shergottites (21). Also, the measured 142Nd/144Nd of ALH [ε142Nd = –0.23 ± 0.05 relative to the terrestrial standard (20)] is slightly lower than (but within error of) the enriched shergottite average of –0.19 ± 0.05. Combined, ALH and the enriched shergottites have distinctly lower ε142Nd than all other martian meteorites that have ε142Nd values of –0.23 ± 0.03 to +0.65 ± 0.05 (21). The ALH carbonate 87Sr/86Sr is identical to its measured bulk rock at 3.90 ± 0.04 Ga, indicating that the carbonate Sr was primarily, if not wholly, derived from ALH (9). Given this, the calculated source 87Rb/86Sr is 0.41 to 0.35, similar to the average source of enriched shergottites of ~0.36 (23). Finally, similarities between ALH and enriched shergottite primary magmas are also indicated by the incompatible trace element abundances of coexisting melts with the cumulate phases in ALH (25). Combined, these observations indicate that ALH is derived from a source that is more incompatible trace element–enriched than shergottites. When source Sm/Nd and Lu/Hf ratios of shergottites and ALH are plotted together, they define a mixing array between hypothetical incompatible trace element–depleted and –enriched end members (Fig. 2B).

Table 2

Lu-Hf and Sm-Nd isotope data measured in this study. See (20) for data reduction and analytical procedures.

View this table:
Fig. 2

(A) Hf isotope evolution diagram relative to CHUR (33) of sources for enriched shergottites RBT 04262 (RBT), NWA 4468, Shergotty (Sher), Zagami (Zag), LAR 06319 (LAR), Los Angeles, and NWA 856 (24, 30, 3740) as well as ALH. There are different recent interpretations of the ages of the shergottites. We have calculated all of the source 176Lu/177Hf and 147Sm/144Nd ratios of shergottites assuming that their igneous crystallization ages are given by their internal isochron ages (0.15 to 0.57 Ga) (11, 23, 24, 29, 3840), not their older Pb-Pb ages (~4.1 Ga) (4, 38, 39); see (20) for further discussion. The isotopic evolution of the sources for the plotted samples is based on a two-stage model (23, 29) where the sources formed from a chondritic reservoir at 4.513 Ga. The 176Lu/177Hf source values are indicated by the numbers referencing the isotopic evolution lines. (B) Mixing diagram for shergottites and ALH 147Sm/144Nd and 176Lu/177Hf source compositions. Red dots, shergottites; DS, depleted shergottites DaG 476, QUE 94201, SaU 094, DaG 476, and SaU 008; IS, intermediate shergottites EETA 79001 and ALH 77005; ES, enriched shergottites RBT 04262, LAR 06319, Shergotty, Zagami, NWA 4468, NWA 856, and Los Angeles; ALH, ALH 84001. The black binary mixing line is based on source compositions of residual trapped liquid (RTL) and cumulates in the upper mantle assemblage (UM1) of (30) produced in MO 2000 to 1350 km deep. The red mixing line was calculated with source end-member compositions used in (23). Blue squares are calculated sources for KREEP-rich lunar basalts. Isotope data used for the source calculations of shergottites come from (21, 23, 24, 29, 30, 3740). Labeled mixing proportions (black symbols) are based on the fractions of RTL. (C) Enlargement of boxed area of (B). Solid symbols are defined as in (B) and represent calculated source compositions; open symbols represent the 147Sm/144Nd and 176Lu/177Hf compositions of liquid in equilibrium with cumulates during MO crystallization (30). For MO 2000 to 1350 km deep, the open square represents RTL in the shallow upper mantle after 98% MO crystallization, the open triangle represents RTL in the upper mantle [UM1 of (30)] after 90 to 94% crystallization, and the open circle is RTL in the upper mantle [UM2 of (30)] after ~66% crystallization. Enriched end-member compositions (31) are labeled BD; percent values represent degree of MO crystallization. Ellipse reflects 2σ error of calculated source compositions for ALH.

In principle, the enriched source characteristics constrained by ALH and less so by the enriched shergottites could potentially be derived from variable amounts of crustal assimilation during magma differentiation (26, 27). However, from the coupled relationships between Os and Nd isotopes, shergottites cannot be related by simple mixing of depleted mantle and crust (28). Furthermore, crustal assimilation is not consistent with major element compositions of shergottites and their trace element and isotope systematics (23). From mixing calculations, the ratio of enriched component to depleted component in the martian mantle that yields the measured source 147Sm/144Nd and 176Lu/177Hf of ALH is about 0.6:0.4. Because ALH falls on the well-defined binary mixing array for Sm-Nd and Lu-Hf source compositions, it would also be highly unlikely that assimilation of mineralogically diverse crustal rocks would yield such a tight array for rocks representing parental magmas formed billions of years apart. Thus, the source for ALH was most likely in the martian mantle. This source has trace element and lithophile isotope characteristics similar but not identical to lunar KREEP (potassium–rare earth element–phosphorus) basalt sources and was likely produced by similar mechanisms (Fig. 2, A and B) (23, 29).

The enriched and depleted end-member compositions in Mars likely formed from the progressive crystallization of its MO, where the residual trapped liquid components are enriched in incompatible trace elements relative to the cumulate fractions at any given point during the crystallization process (23, 2931) creating a hybridized mantle source with the degree of enrichment controlled by the relative proportion of trapped liquid relative to cumulate minerals. The 147Sm/144Nd and 176Lu/177Hf ratios of enriched residual liquid and depleted cumulate end-member components not only vary with degree of MO crystallization, but also with depth of equilibration and initial MO depth (Fig. 2C) (30). Because the modeled enriched end-member contribution to ALH parental magma is high (~60%), the composition of the enriched end member is nearly identical to its calculated source composition. The enriched component likely represents deep (200 to 750 km) residual trapped liquid (RTL) equilibrated with cumulates after ~93% MO crystallization (30) (Fig. 2C). This component is unlike that derived from shallow, very late-stage liquids remaining after >98% MO crystallization that explains lunar KREEP (31) (Fig. 2C). Additionally, cumulates in equilibrium with this RTL are a good match for the composition of the depleted mantle end-member composition as evidenced by shergottite source modeling (30) and our finding that mixtures of these two components can generate the observed sources of shergottites and ALH that range in age from 0.165 to 4.091 Ga (Fig. 2B). Therefore, it is likely that similar magmatic source regions in Mars that produced the long-lived Tharsus and Elysium volcanic regions have been producing magmas for at least the past 4 billion years.

The igneous crystallization of ALH occurred at 4.091 ± 0.030 Ga, during a period of intense bombardment slightly prior to and cessation of the Mars global magnetic field (15, 16, 32). As such, the magnetic properties of the igneous phases in ALH (17) do not record early planetary magnetic fields, as implied for a crystallization age of ~4.5 Ga, and must instead reflect conditions after accretion and ~400 million years of cooling in Mars. The Lu-Hf age requires a much shorter time interval between igneous formation and aqueous alteration at 4.04 ± 0.10 to 3.90 ± 0.04 Ga, necessitating a revision of the timing of pre-alteration textures (13). The younger age predicts that the primordial martian crust was likely largely destroyed from intense bombardment at 4.25 to 4.1 Ga (15).

Supporting Online Material

www.sciencemag.org/cgi/content/full/328/5976/347/DC1

Materials and Methods

Figs. S1 to S5

References

References and Notes

  1. C. Meyer, Mars meteorite compendium, NASA-ARES JSC #27672 Revision C (2008).
  2. See supporting material on Science Online.
  3. e176HfCHUR = (176Hf/177HfCHUR/176Hf/177HfCHUR – 1) × 10,000; CHUR refers to chondritic uniform reservoir model values for the Lu-Hf, as well as Sm-Nd, isotope systems. Hf and Nd CHUR reference values are from (33) and references therein.
  4. Supported by NASA Cosmochemistry grants (T.J.L. and A.D.B.), a NASA Astrobiology grant (B.L.B.), the University of Houston Institute for Space Systems Operations (T.J.L.), and the Belgian Fund for Scientific Research (V.D.). We thank three anonymous reviewers for improving the clarity of the manuscript.
View Abstract

Stay Connected to Science

Navigate This Article