We report on the discovery of a new shergottite from Tunisia, Ksar Ghilane (KG) 002. This single stone, weighing 538 g, is a coarse-grained basaltic shergottite, mainly composed of maskelynitized plagioclase (approximately 52 vol%) and pyroxene (approximately 37 vol%). It also contains Fe-rich olivine (approximately 4.5 vol%), large Ca-phosphates, including both merrillites and Cl-apatites (approximately 3.4 vol%), minor amounts of silica or SiO2-normative K-rich glass, pyrrhotite, Ti-magnetite, ilmenite, and accessory baddeleyite. The largest crystals of pyroxene and plagioclase reach sizes of approximately 4 to 5 mm. Pyroxenes (Fs26–96En5–50Wo2–41). They typically range from cores of about Fs29En41Wo30 to rims of about Fs68En14Wo17. Maskelynite is Ab41–49An39–58Or1–7 in composition, but some can be as anorthitic as An93. Olivine (Fa91–96) occurs mainly within symplectitic intergrowths, in paragenesis with ilmenite, or at neighboring areas of symplectites. KG 002 is heavily shocked (S5) as indicated by mosaic extinction of pyroxenes, maskelynitized plagioclase, the occurrence of localized shock melt glass pockets, and low radiogenic He concentration. Oxygen isotopes confirm that it is a normal member of the SNC suite. KG 002 is slightly depleted in LREE and shows a positive Eu anomaly, providing evidence for complex magma genesis and mantle processes on Mars. Noble gases with a composition thought to be characteristic for Martian interior is a dominant component. Measurements of 10Be, 26Al, and 53Mn and comparison with Monte Carlo calculations of production rates indicate that KG 002 has been exposed to cosmic rays most likely as a single meteoroid body of 35–65 cm radius. KG 002 strongly resembles Los Angeles and NWA 2800 basaltic shergottites in element composition, petrography, and mineral chemistry, suggesting a possible launch-pairing. The similar CRE ages of KG 002 and Los Angeles may suggest an ejection event at approximately 3.0 Ma.
Shergottites, nakhlites, and chassignites (SNC), and Allan Hills (ALH) 84001 meteorites are igneous rocks believed to have originated from Mars. In the past and at present, Martian rocks have been studied extensively to provide possible petrologic constraints on the geological history of Mars (e.g., McSween 1994, 2002). Among the major group of Martian rocks, the shergottites, three subtypes can be distinguished: (1) basaltic shergottites, e.g., Shergotty, Zagami, Queen Alexandra Range (QUE) 94201, Los Angeles, Northwest Africa (NWA) 2800, which are pyroxene-plagioclase-dominated basalts; (2) lherzolitic shergottites, e.g., Allan Hills (ALH) A77005, Lewis Cliff (LEW) 88516, Yamato (Y)-793605 that are olivine-pyroxene cumulates derived from a basaltic magma (McSween and Treiman 1998); and (3) olivine-phyric shergottites, e.g., Dar al Gani (DaG) 476, Dhofar 019, Y-980459, Sayh al Uhaymir (SAU) 005, Elephant Moraine (EET) Lithology A A79001 (Goodrich 2002). The origin of the latter is controversially discussed. They might be products of mixing between basaltic shergottite-like magma sources and lherzolitic material, or represent magmas that could have been parental to both (1) and (2), or may be the product of a completely different magma source and/or formation process. This article addresses petrologic, mineral and bulk chemical, cosmogenic radionuclide, noble gas, and oxygen isotope data of a new basaltic shergottite, Ksar Ghilane 002 (KG 002). This meteorite is the 100th registered fragment available from Mars, not taking paired stones into account, and the first to be recovered from Tunisia. The Sahara has become one of the most productive areas for Martian meteorite recovery, providing nearly half of the collection (101 total; The Meteoritical Bulletin Database, December 2011). In the last 10 years, 50 new fragments have been recovered, 35 of them in the Sahara. Such new discoveries thus present an opportunity to improve our knowledge of Martian meteorites.
Ksar Ghilane 002 was collected by José Vicente Casado and David Allepuz on January 13, 2010, in the Ksar Ghilane dense collection area (KG) at latitude 32°48.375′ N and longitude 009°49.970′ E (Fig. 1a). The find site is on the Saharan platform, over Senonian limestone. The meteorite is a single stone weighing 538 g, with dimensions of approximately 10 × 4.5 × 3.5 cm (Fig. 1b). The surface of KG 002 has been subject to hot desert weathering (i.e., sand abrasion), resulting in the removal of the majority of the fusion crust. In hand sample, KG 002 is dominated by light green pyroxene crystals, abundant maskelynite, and rare opaque oxides (sawn surface shown in Fig. 1c). During the fourth meteorite recovery campaign in Tunisia (2011), although the KG 002 recovery area was carefully examined, no further shergottite specimens were found. Most of our results are discussed with and compared with previous data for two other Martian meteorites—Los Angeles and NWA 2800—as both of them have strong similarities to KG 002. The possibility of a launch-pairing among these basaltic shergottites is also discussed.
Textural and mineralogical investigations of KG 002 were performed by optical and electron microscopy on four different thin sections. Detailed petrographic investigations were carried out using (1) a JEOL 6610-LV scanning electron microscope (SEM) equipped with energy dispersive spectrometers (EDS; INCA; Oxford Instruments) at the Interdisciplinary Center for Electron Microscopy and Microanalysis (ICEM) at the Westfälische Wilhelms-Universität Münster using an acceleration voltage of 20 kV and a beam current of 15 nA; and (2) a Zeiss Neon 40 field emission SEM equipped with EDS at the Centre for Research in Nanoengineering (CRnE) at the Universitat Politècnica de Catalunya. Quantitative analyses of mineral compositions were obtained using a JEOL JXA 8900 Superprobe electron probe micro-analyzer (EPMA) operated at an acceleration voltage of 15 kV and beam current of 15 nA for major and minor elements (Westfälische Wilhelms-Universität Münster). Shock melt glasses and phosphates have been analyzed with a defocused beam of 50 and 10 μm, respectively. Natural and synthetic standards were used for calibration purposes. These include jadeite (Na), sanidine (K), hypersthene (Si), disthene (Al), olivine (San Carlos, Mg), diopside (Ca), fayalite (Fe), rutile (Ti), rhodonite (Mn), and Cr2O3 (Cr). Matrix corrections were made according to the Фρ(z) procedure of Armstrong (1991). The modal mineral abundance of KG 002 was obtained by quantitative electron microprobe grid analysis. A total of 735 points were analyzed and the calculated average of these quantitative data was used to evaluate the major element bulk composition.
A 2.065 g fragment from the interior of the meteorite (at least 1 cm away from the surface in all directions) was finely ground using an agate mortar and pestle. Four aliquots of 79.94, 77.79, 61.94, and 58.01 mg of the powdered specimen were used for bulk chemical analyses. Two independent methods were used for sample preparation: acid digestion treatment in a sealed Teflon reactor and alkaline fusion in a zirconium crucible. Analyses were performed by means of inductively coupled plasma-mass spectrometry (ICP-MS) using a Perkin Elmer Elan 6000 instrument and inductively coupled plasma-optical emission spectroscopy (ICP-OES) using a Perkin Elmer Optima 3200 RL instrument (Barcelona). Our procedures are the same as those described by Llorca et al. (2007, 2009). The accuracy on major and trace element concentrations is better than 5% in all cases.
The oxygen isotope composition was analyzed by means of laser fluorination in combination with gas chromatography continuous flow isotope ratio monitoring mass spectrometry (GC-CF-irmMS). Approximately 1 mg of bulk powder was loaded, along with terrestrial MORB glass (δ18O = +5.6‰) and NBS-28 quartz (δ18O = +9.6‰) standards. The material was fluorinated in a stainless steel sample chamber (Sharp 1990) using approximately 20 mbar purified F2 gas (Asprey 1976). A SYNRAD 50 W CO2 laser was used for the heat source. The samples reacted with the F2 to produce fluorides and O2. Excess F2 was removed by reacting the gas mixture with hot (approximately 150 °C) NaCl to form NaF and Cl2. The Cl2 was removed in a cold trap (−196 °C) and the sample O2 was collected in a 13X molecular sieve (−196 °C), which was then released at approximately 150 °C into a He carrier gas stream and transported to a molecular sieve cryofocusing trap (−196 °C). From there, the sample gas was released by submersion in a water bath (92 °C) and transported with He through a THERMO GASBENCH II gas chromatograph (5 Å molecular sieve) and an open split into the source of a THERMO MAT253 multicollector gas mass spectrometer. The gas chromatograph separated interfering NF (from NF3) (Pack et al. 2007) from O2. The reference gas was injected using the GASBENCH. The δ17O and δ18O values were determined by simultaneously analyzing m/z = 32, 33, and 34. The Δ17O was calculated according to:
The Δ17O of NBS28 quartz and MORB glass was assumed to be zero. The external accuracy of GC-CF-irmMS was ±0.2‰ (1σ) for δ18O and the external accuracy of Δ17O (Young et al. 2002) was ±0.05‰ (1σ).
A 67.53 mg bulk fragment of KG 002 was analyzed for all noble gases helium (He), neon (Ne), argon (Ar), krypton (Kr), and xenon (Xe) using an MAP 215-50 noble gas electron source mass spectrometer based at the Max-Planck-Institut für Chemie (MPIC), Mainz. Noble gases were released from the sample using a furnace step-heating technique in steps of 600, 1000, and 1800 °C. Whilst a final 1900 °C step was performed to ensure that all gases had been released from the sample, this information is omitted from the final results, as in all cases, it was equivalent to a blank. Upper limits on blanks were: 4He 4.4 × 10−10 ccSTP, 22Ne 2.4 × 10−12 ccSTP, 36Ar 9 × 10−12 ccSTP, 84Kr 0.3 × 10−12 ccSTP, 132Xe 0.2 × 10−12 ccSTP. Sensitivity and mass discrimination of the instrument were monitored with analysis of a calibration gas of known composition (standard atmospheric, except for 3He/4He approximately 1), and the data were corrected for interference masses.
Long-lived cosmogenic radionuclides were separated from a 177.44 mg bulk fragment following a procedure described by Merchel and Herpers (1999). Basic steps were: crushing; addition of stable isotope carrier (9Be, 35Cl (35Cl = 99%), 55Mn); acid digestion in a Teflon pressure bomb; anion and cation exchange; precipitation of AgCl, Be(OH)2, Al(OH)3, MnO(OH)2, and Fe(OH)3; drying and partial oxidation (BeO, Al2O3, MnO2, Fe2O3); and mixing with metal powder (BeO with Nb; Al2O3, MnO2 and Fe2O3 with Ag). Accelerator mass spectrometry (AMS) measurements of 10Be and 26Al were performed at the new 6 MV AMS-facility DREAMS in Dresden (Germany) (Akhmadaliev et al. 2013), whereas 53Mn was detected at the Maier-Leibnitz-Laboratory (MLL) in Garching (Germany) (Poutivtsev et al. 2010). Additionally, 60Fe was determined at MLL (Knie et al. 2000). Calibration materials are SMD-Be-12 with 10Be/9Be = (1.704 ± 0.030) × 10−12, SMD-Al-11 with 26Al/27Al = (9.66 ± 0.14) × 10−12, Grant GLS with 53Mn/55Mn = (2.83 ± 0.14) × 10−10, and PSI-12 with 60Fe/Fe = (1.25 ± 0.13) × 10−12 diluted from a primary standard (Rugel et al. 2009). Both DREAMS calibration materials are traceable to primary standard-type materials, i.e., via cross-calibration to NIST SRM 4325 (Nishiizumi et al. 2007) and the 26Al round-robin material of Merchel and Bremser (2004). Grant GLS is an in-house calibration material traceable to a primary-type standard resulting from a nuclear reaction (Schaefer et al. 2006). However, the original published value of (2.96 ± 0.06) × 10−10 (Schaefer et al. 2006) has been slightly changed due to advanced cross-calibration measurements. Used 26Al and 53Mn standards are generally independent from a certain half-life. Half-lives used for the conversion of isotope ratios into radionuclide activities are (1.387 ± 0.012) Ma for 10Be (Korschinek et al. 2010), (0.705 ± 0.024) Ma for 26Al (Norris et al. 1983), and (3.7 ± 0.4) Ma for 53Mn (Honda and Imamura 1971). Measured concentrations have been compared with depth- and radius-depending production rates from theoretical Monte Carlo calculations (Leya and Masarik 2009) based on the mean bulk chemical composition of KG 002.
Results and Discussion
Petrography and Mineral Chemistry
Our analyses indicate that KG 002 is a coarse-grained basaltic shergottite with an ophitic to subophitic texture and preferred orientation of large, elongated grains (Figs. 2 and 3). It is mainly composed of maskelynitized plagioclase (approximately 52 vol%) and pyroxene (approximately 37 vol%), with Ca-phosphates, including both merrillite and Cl-apatite, minor amounts of silica or SiO2-normative K-rich glass, Fe-rich olivine, pyrrhotite, Ti-magnetite, ilmenite, and trace amounts of baddeleyite (Table 1). The largest crystals of pyroxene and plagioclase in KG 002 reach sizes of approximately 4 and 5 mm, exceeding those of most other basaltic shergottites including Los Angeles (e.g., Mikouchi 2000), except for NWA 2800 (Bunch et al. 2008). The latter strongly resembles KG 002 in texture, grain sizes, and mineral constituents. The bulk rock chemical composition for major elements (calculated based on EPMA data), and representative analyses of silicates and oxides in KG 002 are given in Tables 2 and 3.
(1) This study, calculated from EPMA data normalized to 100%, P2O5- and S-concentrations calculated from modal abundance of phosphates and pyrrhotite; (2) This study, data obtained by ICP-MS and ICP-OES (Si); (3) Rubin et al. (2000; INAA); (4) Jambon et al. (2002; ICP/AES); (5) Warren et al. (2000; electron microprobe); (6) Ikeda et al. (2006; electron microprobe).
Table 3. Chemical composition of silicates and oxides obtained with EPMA. Results in wt%
The large plagioclases (maskelynite) are An37–58Or1–7Ab41–59 in composition (Table 3), whereas the smaller ones, which are interstitial to pyroxenes or in close association with melt pockets, can be as anorthitic as An93. The maskelynite composition is significantly distinct from that of highly evolved NWA 2800 (An55.5–72.3Or0.6–3.8; Bunch et al. 2008), but similar to maskelynite in Los Angeles (An38–56Or1–7Ab43–56; Rubin et al. 2000) and Shergotty (An41–56Or1–3Ab41–56; Smith and Hervig 1979). In addition to maskelynite and smaller plagioclase crystals, amorphous K-feldspars were found to occur within SiO2-normative glass patches together with accessory pyroxene and pyrrhotite (Fig. 4b). Their composition can be calculated as An24–43Or5–34Ab38–57 (Table 3) and thus differs from surrounding plagioclase (or maskelynite). It is likely that both the K-feldspar and SiO2-phases have been converted into an amorphous state during shock metamorphism.
The pyroxenes in KG 002 are strongly zoned and typically crosscut by numerous cracks (Fig. 3). Their compositions range from Fs26–96En3–50Wo2–41, with Mg-rich cores of about Fs29En41Wo30 to Fe-rich rims of about Fs68En14Wo18 (Table 3, Figs. 5 and 6). This zoning is clearly indicated in Figs. 5d and 5f. The range of pyroxene composition, given in Fig. 6, is similar to that of Los Angeles pyroxenes (Mikouchi 2001; Xirouchakis et al. 2002). Approximately 10–15 vol% of the KG 002 sample consists of 50 μm to 2 mm wide patches of a fine-grained vermicular to microgranulitic intergrowth of fayalite, Ca-pyroxene, and silica, in approximately 2:2:1 proportion (Figs. 4c and 5c), a texture previously described as symplectite (e.g., Aramovich et al. 2002). As already proposed by Lindsley et al. (1972), such symplectites may have formed due to the breakdown of pyroxferroite at low pressure and during slow cooling, coupled with a ferroan bulk composition of the host rock. Overall, the modal abundance and composition of these symplectite areas are similar to those described for Los Angeles and NWA 2800 (Rubin et al. 2000; Aramovich et al. 2002; Bunch et al. 2008). However, the apparent size of these patches is significantly larger (up to 2 mm) compared with other shergottites.
Olivine, which constitutes about 4.5 vol% of the whole rock (Table 1), occurs mainly within symplectitic intergrowths, in paragenesis with ilmenite and Ca-phosphates, or at neighboring areas of symplectites (Figs. 4a, 4c, 5b, and 5c). Olivine crystals are Fa91–96 in composition and thus comparable to those described in the Los Angeles shergottite (Fa93–95; Xirouchakis et al. 2002).
Ca-phosphates, including both merrillite and Cl-apatite, are abundant in KG 002 and represent about 3.4 vol% of the whole rock (Table 1). However, the dominating phase is merrillite, with crystal sizes up to 800 μm in apparent diameter (Figs. 4a and 5c). Their average composition is given in Table 3. Fe-Ti oxides, including Ti-magnetite and ilmenite, typically occur interstitial to pyroxenes and plagioclase or are associated with phosphates and symplectite areas (Figs. 4a, 4c, and 5c). Coexisting Fe-Ti oxide compositions indicate oxygen fugacity equilibrium of the rock near the quartz-fayalite-magnetite buffer (McSween 1994). Baddeleyite is present as an accessory phase within KG 002. The grains are usually <5 μm in diameter with the largest grain being 15 × 10 μm in apparent size, and typically occur at the edges of symplectite areas, as enclosed crystals in Ti-magnetite and in close association with merrillite.
Ksar Ghilane 002 is heavily shocked, equivalent to an S5 stage, when applying the classification scheme for ordinary chondrites (Stöffler et al. 1991; Bischoff and Stöffler 1992). Evidence for such high shock includes the mosaic extinction of pyroxenes (Fig. 3c), the occurrence of maskelynite (produced by shock transformation of plagioclase), and localized shock melt glass pockets, some of them containing pyroxene- and plagioclase-normative schlieren (Fig. 5a). Similar shock melt pockets have been previously described in the lherzolitic shergottite NWA 1950 (Walton and Herd 2007) and in the shergottite EETA79001 (Fritz et al. 2005; Walton et al. 2010). Fritz et al. (2005) determined a shock pressure of 39 ± 4 GPa for lithology A of EETA79001.
The bulk composition of KG 002 has been measured for 45 elements. Results for major and minor elements are compiled in Table 2, where two other basaltic shergottites, Los Angeles and Dhofar 378, are included for comparison. It is well known that basaltic shergottites, lherzolitic shergottites, and olivine-phyric shergottites are separable in TiO2, Al2O3, Cr2O3, MgO, CaO, and Na2O abundances (e.g., Shirai and Ebihara 2004). Major element contents of KG 002 fall within the range of basaltic shergottites, in accordance with the petrography and mineral chemistry discussed above. In particular, Al2O3 (12.25%), MgO (4.47%), CaO (10.3%), Na2O (2.14%), and K2O (0.23%) contents are well within the basaltic shergottite range and far from the other shergottite groups (Lodders 1998). Element abundances (calculated based on EPMA data) are in agreement with ICP data (see Table 2 for comparison). Note that the bulk rock chemical compositions of KG 002, Los Angeles, and Dhofar 378 overlap within uncertainties (Table 2), which suggests a possible genetic relationship. Figure 7 is a plot of Ca/Si versus Mg/Si ratios for Martian meteorites (where the positions of the various shergottite subclasses are indicated). The shergottites spread along a line with endmembers of high Ca/Si and low Mg/Si ratios versus low Ca/Si and high Mg/Si ratios. As clearly confirmed in Fig. 7, KG 002 belongs to the group of basaltic shergottites and represents the most extreme case on the highly evolved high Ca/Si, low Mg/Si side. According to the scheme proposed by Irving (2011), and based on its CaO content, Mg/(Mg+Fe), and bulk rock REE pattern (see below), KG 002 belongs to the group of enriched mafic, diabasic shergottites. Other members of this group are, among others, Shergotty, Zagami, Los Angeles, and Dho 378.
Trace element abundances of KG 002 are compiled in Table 4, together with those of Los Angeles for comparison. Concerning alteration, hot desert meteorites often display strong K, Ba, Sr, Cs, U, and Ce enrichments (e.g., Barrat et al. 1999; Stelzner et al. 1999; Zipfel et al. 2000). The Th/U ratio of KG 002 is 3.4, which is in the lower range of values reported for Martian meteorite falls (Th/U ratios of about 3.6–4.3). However, a number of shergottites from hot deserts have even lower Th/U ratios (McLennan 2003). The K/La ratio of KG 002 is 886, which is within the range for most shergottite falls (approximately 700–1000) and distinct from heavily weathered shergottites, which show K/La ratios up to 2000. By comparison, Ba and Sr abundances in KG 002 are only slightly outside the trend defined by nonweathered shergottites in Ba versus La and Sr versus Nd plots, respectively (Fig. 8), but distinct from strongly weathered shergottites in hot deserts, like DaG 476, DaG 489, and Dhofar 019 (Jambon et al. 2002). We therefore conclude that KG 002 is only moderately weathered.
Table 4. Trace element abundances (in μg g−1) of KG 002 obtained by ICP-MS. Data for Los Angeles are included for comparison (Jambon et al. 2002)
7.30 ± 0.06
37.9 ± 0.5
201 ± 0.9
30.4 ± 0.1
32.2 ± 0.2
22.1 ± 0.4
89.2 ± 0.5
7.51 ± 0.07
108.7 ± 0.2
17.0 ± 0.1
152 ± 2
3.47 ± 0.03
0.26 ± 0.02
0.05 ± 0.01
112 ± 1
2.11 ± 0.01
6.31 ± 0.04
0.74 ± 0.01
3.66 ± 0.01
1.37 ± 0.01
0.76 ± 0.01
2.13 ± 0.01
0.44 ± 0.01
2.91 ± 0.01
0.60 ± 0.01
1.77 ± 0.02
0.25 ± 0.01
1.68 ± 0.01
0.24 ± 0.01
4.57 ± 0.01
1.39 ± 0.01
0.54 ± 0.01
0.16 ± 0.01
The compatible trace element abundances (such as Ni, Co, Cr, and Cu) of KG 002 are strikingly similar to those reported for Los Angeles (Table 4), which in turn are distinct from other basaltic shergottites (Lodders 1998), thus providing additional evidence for their strong affinity. The REE pattern of KG 002 (Fig. 9) exhibits a pattern similar to Los Angeles and other slightly depleted LREE shergottites including the less evolved Zagami and Shergotty. The positive Ce anomaly (Ce/Ce* ~ 1.2) is probably related to terrestrial weathering, in accordance with the moderate U, Ba, and Sr enrichments discussed above. This anomaly is very rare among Martian meteorites, but has been observed e.g., in the strongly weathered Dhofar 019 shergottite (Neal et al. 2001; Taylor et al. 2002). Ce anomalies are also found in Antarctic shergottites, such as QUE 94201, EETA 79001, and ALHA 77005, and are thought to result from the partial oxidation of Ce3+ to Ce4+, where Ce4+ is relatively insoluble compared with the rest of the trivalent REE and is less mobile than the other REE during terrestrial weathering (Crozaz et al. 2003). In addition, the REE pattern of KG 002 is characterized by a pronounced positive Eu anomaly (Eu/Eu* ~ 1.4), similar to the basaltic shergottite Dhofar 378, but in contrast to the highly evolved Los Angeles shergottite (Fig. 9), providing additional evidence for complex magma genesis and mantle processes on Mars.
Two aliquots of KG 002 were analyzed, yielding δ18O values of +5.06‰ and +5.23‰. The corresponding δ17O values were +2.91‰ and +3.14‰. The Δ17O values were +0.25‰ and +0.40‰ with a mean of +0.32‰, identical within uncertainty. This value falls in the field typical for Martian meteorites (Clayton 1993; Franchi et al. 1999) (Fig. 10).
Noble Gas Analysis
Noble gas results for KG 002 are summarized in Tables 5 and 6, where they are compared with literature data for Los Angeles. Noble gas data for NWA 2800 are not available at this time in the literature. Whilst low concentrations of He and roughly half the Ne budget are released in the 1800 °C step, release of the heavier noble gases Ar, Kr, and Xe occurs preferentially at this temperature (Table 5).
Table 5. Noble gas concentrations for KG 002. Previous data reported for the Los Angeles shergottite are also shown. Concentrations are: 3–4He, 40Ar in 10−8 ccSTP g−, 20–22Ne and 36–38Ar in 10−10 ccSTP g−1, 84Kr and 132Xe in 10−12 ccSTP g−1
Overall, KG 002 shows low He concentrations of 3He approximately 3 × 10−8 ccSTP/g and 4He approximately 2 × 10−7 ccSTP g−1 compared with previous analyses of Martian shergottites, where the majority show a range of 3He of 1–7 × 10−8 ccSTP g−1 and 4He 2 × 10−7 to 3 × 10−6 ccSTP g−1 (see refs. in Fig. 11 (Ne) for previous shergottite noble gas data). Whereas 3He is mainly cosmogenic in origin (and thus relates to the CRE age), 4He concentrations reflect both cosmogenic and radiogenic sources. If we assume a range of cosmogenic (4He/3He)c ratios from 4.0 to 6.0 (Alexeev 1998; Schwenzer et al. 2008), and subtract the associated “cosmogenic” from the measured 4He, the corrected “radiogenic” 4He ranges from approximately 3 to 9 × 10−8 ccSTP g−1. Using the U and Th concentrations from Table 4, we can calculate an apparent “crystallization” age range of 0.7–2.5 Ma. This range is close to the calculated CRE age (see below) and hence consistent with essentially complete loss of radiogenic He accumulated at the time of ejection. Radiogenic He loss is a common occurrence in many Martian meteorites (e.g., Schwenzer et al. 2008), and probably relates to shock metamorphism—(see discussion in the Petrography and Mineral Chemistry section).
For Ne, the 22Ne concentration of 8 × 10−9 ccSTP g−1 is within the range observed in previous studies of shergottites of approximately 1–13 × 10−9 ccSTP g−1 (see refs. in Fig. 11 for previous shergottite data). Figure 11 is a three isotope plot of 20Ne/22Ne versus 21Ne/22Ne with our step-wise heating data for KG 002, alongside shergottite literature data and various endmember components including “trapped” components Earth's atmosphere (EA), Martian atmosphere (MA), solar wind (SW), as well as cosmogenic neon produced by galactic cosmic rays (GCR), by low-energy solar flare protons (solar cosmic rays, SCR), and cosmogenic Ne produced from sodium (Na) (for associated references, see Fig. 11). Our KG 002 results plot within the range observed in previous shergottite analyses and indicate contributions from GCR, Na, and possibly mixing with a minor amount of SCR, as well as mixing with an atmospheric component. As the most significant shift toward terrestrial atmosphere is observed in the lowest temperature step (600 °C), this may very well represent mixing with terrestrial air due to contamination, rather than Martian atmosphere (Fig. 11). The cosmogenic (21Ne/22Ne)c ratio for the sample is 0.783 ± 0.006, consistent with the range observed in previously studied Martian meteorites. Whilst it has been suggested that such ratios require contribution from SCR-Ne, comparison with GCR-Ne production rates based on Leya and Masarik (2009) and the chemical composition of KG 002 (Tables 2 and 4) shows that for KG 002—while a small presence cannot be excluded—SCR-Ne is not required (see Fig. 12). This is because of the extremely low abundance of Mg and the high abundance of Na (approximately 1.8%), contributing Ne with extremely low 21Ne/22Ne (Fig. 11). For example, the (22Ne/21Ne)c ratio modeled for KG 002 using Leya and Masarik (2009) production rates shows a significant 0.10 reduction if Na is omitted. Furthermore, the 1000 °C temperature step plots right on the mixing trend from common GCR to Na spallation, with no evidence for enhanced 20Ne/22Ne as seen in SCR produced Ne.
For the heavier noble gases, it is difficult to distinguish between separate components within KG 002 when relying solely on isotopic ratios. For example, the 129Xe/132Xe ratio may indicate mixing between major components of Martian atmosphere and/or Martian interior and/or Earth's atmosphere (Table 6). Combining this with the heavy noble gas elemental ratios may provide a better insight. Figure 13 is a plot of 129Xe/132Xe ratios versus elemental ratios of trapped 36Art/132Xe and 84Kr/132Xe, including shergottite literature data, Earth's atmosphere (EA), elementally fractionated atmosphere (EFA), Martian atmosphere (MA), Martian interior (MI), and further distinct interior components “Chass-S” and “Chass-E” (for associated references, see Fig. 13; Table 6 captions). This plot indicates that Martian interior may be a dominant noble gas component in KG 002, which is highlighted in the 1800 °C release (Fig. 13). Alternatively, we may have identified an additional distinct interior component with an elevated 40Ar/36Ar ratio and 129Xe/132Xe identical to the Martian interior 1.029 ± 0.019 as defined by Ott (1988), which was also released at the highest temperature step. It is difficult to assess the effect of EFA on observed ratios, especially as the releases for 600 and 1000 °C have low amounts of gas and therefore large analytical uncertainties (Fig. 13).
Cosmic Ray Exposure Ages
Cosmic ray exposure (CRE) ages for Martian meteorites found in the literature have commonly been calculated using the empirical 22Ne/21Ne-corrected production rate equations of Eugster and Michel (1995). Using this approach and assuming that KG 002 is a basaltic shergottite (and thus has production systematics similar to that of eucrites [Eugster et al. 1997]), combined with an average of the major element abundances given in Table 2 for KG 002, the trace element data in Table 4, and the cosmogenic concentrations given in Table 5, we obtain T3 (3He), T21 (21Ne), and T38 (38Ar), ages of 1.97 ± 0.06, 4.42 ± 0.11, and 2.90 ± 0.29 Ma, respectively. Whilst the 3He concentration was assumed to be entirely cosmogenic, the T21 age was calculated from the cosmogenic 21Nec concentration, as obtained assuming the trapped component to have EA composition (Eberhardt et al. 1965) and the cosmogenic component to have a typical GCR 20Ne/22Ne ratio (0.83, e.g., Garrison et al. 1995). The T38 age was calculated from the cosmogenic 38Arc concentration, assuming a trapped component with the 38Ar/36Ar ratio of EA (Lee et al. 2006), and a cosmogenic (38Ar/36Ar)c ratio of approximately 1.5 (Wieler 2002). Note that the errors include uncertainties related to the concentrations of cosmogenic gas as well as the shielding parameter, but not such related to chemical composition. The nominal ages given above show a large scatter. However, it is important to note that the determination of CRE ages for meteorites with unusual chemical compositions is not straightforward. In fact, using the shielding-dependent production rates from Leya and Masarik (2009) and neglecting any possible SCR contribution (minor, if anything, as discussed above), we obtain about 30% shorter CRE ages. This is discussed in more detail below, in the context of comparison with the shergottite Los Angeles and the radionuclide data.
Kr isotopes were partitioned into trapped, cosmogenic, and neutron capture components, using a trapped 83Kr/84Kr ratio of 0.2025 ± 0.0025 and a cosmogenic composition derived from an analysis of eucrite Stannern (Bogard et al. 1971). A rather large meteoroid size is indicated by the presence of excess 80Kr of neutron capture origin (80Krn). The concentration of 1.01 ± 0.20 × 10−12 mL g−1 (Table 5) is at the high end of what has previously been observed for Martian meteorites (Eugster et al. 2002), including Los Angeles (0.93 ± 0.40 × 10−12 mL g−1). Unfortunately, we do not have data for the Br concentration, so it is not possible to deduce a neutron fluence at this time.
For the crystallization age, we can calculate a nominal K-Ar age of 279 ± 27 Ma based on the total 40Ar and K concentrations in the meteorite, assuming that all 40Ar is radiogenic (Tables 2 and 5). However, many Ar-Ar ages calculated for shergottites are older compared with other chronometers—a feature which has been ascribed to excess 40Ar (e.g., Bogard et al. 2009). By assuming that all trapped 36Ar is related to trapped EA (i.e., subtraction of the equivalent 40Ar), we can calculate a K-Ar age of 119 ± 30 Ma. As detailed above, an U-Th-He age for this meteorite is not viable due to the low He concentrations probably caused by He loss during shock metamorphism.
Comparison with Los Angeles
The available data for Los Angeles (Garrison and Bogard 2000; Busemann and Eugster 2002; Eugster et al. 2002) are also displayed in Tables 5 and 6, along with the KG 002 results. For NWA 2800, there are no noble gas data in the literature at this time. There is remarkable similarity between KG 002 and Los Angeles in the concentrations of cosmogenic Ne and Ar (including (22Ne/21Ne)c ratios). However, the situation is different for trapped Ne and Ar (20Net, 36Art), as well as for 40Ar and 84Kr, where abundances for KG 002 are lower compared with all Los Angeles data (Table 5). For 4He, KG 002 is similar to Los Angeles bulk 1 and 2 reported by Eugster et al. (2002) (approximately 2 × 10−7 ccSTP g−1), but significantly lower than the 4He concentration reported by Garrison and Bogard (2000) (approximately 7 × 10−7 ccSTP g−1). Finally, the KG 002 132Xe concentration is similar to that reported for Los Angeles by Garrison and Bogard (2000) (approximately 3 × 10−11 ccSTP g−1), but significantly higher than that reported by Busemann and Eugster (2002) (approximately 8 × 10−12 ccSTP g−1). In addition, Los Angeles may show more influence from both the Martian and terrestrial atmosphere compared with KG 002, with an elevated 129Xe/132Xe ratio of approximately 1.14, 36Art/132Xe of approximately 650, and 84Kr/132Xe ratio of approximately 9.3 (Busemann and Eugster 2002) compared with approximately 1.03, 42, and 1.3, respectively (Table 6; Fig. 13).
A number of crystallization ages have been calculated for Los Angeles, including an Rb-Sr age of 165 ± 11 Ma (Nyquist et al. 2000), Sm-Nd ages of 172 ± 8, and 181 ± 13 Ma (Nyquist et al. 2001; Bouvier et al. 2008), a Lu-Hf age of 159 ± 42 Ma (Bouvier et al. 2008), and an older Pb-Pb age of 4050 ± 70 Ma (Bouvier et al. 2008). Whilst the young ages for Los Angeles (and other shergottites) may result from resetting by shock or be due to the presence of phosphates (Bouvier et al. 2005, 2008), the upper and apparent lower limits on the K-Ar age for KG 002 age of approximately 120 and 280 Ma provide no strong evidence for a true age difference. Thus, both meteorites may have experienced similar resetting events or conditions.
In terms of determining the possibility of a “launch pairing” for Los Angeles and KG 002 based on the noble gas CRE ages, assessing a correlation is not straightforward. Preferred CRE ages for Los Angeles have been reported based on T81 (81Kr) of 3.10 ± 0.70 and 3.35 ± 0.30 Ma (Terribilini et al. 2000; Eugster et al. 2002). These are similar to our T38 age of 2.9 Ma. Eugster et al. (2002) and Garrison and Bogard (2000) also report T3, T21, and T38 ages for Los Angeles of approximately 1.35–1.9, 3.13–3.6, and 3.23–2.8 Ma, respectively, which cover the same range. The T3 ages are low compared with the T21 and T38 ages, a feature also observed for our KG 002 analysis. Concerning 21Ne, it is important to note that Garrison and Bogard (2000) corrected their 21Nec concentration for production from Na and then, based on this corrected value, calculated a T21 of 3.6 Ma using the (22Ne/21Ne)c ratio of approximately 1.163 from their highest temperature release (1600 °C), as this temperature step should be the least affected by production on Na. Eugster et al. (2002) followed the same approach, arriving at a T21 age of 3.15 Ma. If we perform a similar correction on our total 21Nec concentration based on a 4.4% Na contribution to 21Nec (estimated from the Na content given in Table 2 and “typical” production rates on sodium [Leya and Masarik 2009]) and only use our 1800 °C temperature step for the (22Ne/21Ne)c shielding parameter (1.161 ± 0.012, compared to 1.277 ± 0.009 from the total release), we obtain a T21 age of 3.44 ± 0.11 Ma. This value is in agreement with the (presumably more reliable) T38 age of 2.90 ± 0.29 Ma.
In summary, although the noble gas concentrations and ratios for Los Angeles and KG 002 are somewhat different, the crystallization age data may hint at similar resetting events. More importantly, the CRE ages of both meteorites show a strong resemblance. Based on the production rate systematics of Eugster et al. (2002), an ejection event approximately 3.0 Ma ago is suggested.
Shielding Conditions and CRE Ages from Cosmogenic Radionuclides and the Physical Model of Leya and Masarik (2009)
Cosmogenic radionuclide data of KG 002 in comparison with earlier published ones for Los Angeles are summarized in Table 7. The given uncertainties include AMS measurement uncertainties (statistical and from used calibration materials), but none from stable isotope detection (26Al and 55Mn) nor Fe-measurements for normalizing purposes of 53Mn. The measured 60Fe/Fe of KG 002 is 6.2 × 10−16 (+2 × 10−15; −1.5× 10−17) and was not further interpreted due to the large uncertainty associated with the experimental value. Data given by Nishiizumi et al. (2000) and Berezhnoy et al. (2010) for Los Angeles mostly agree well with 10Be and 53Mn data of KG 002, whereas the higher 26Al activity of KG 002, in comparison with Los Angeles, originates partially from the high abundance of stable Al, accounting for about one third of the total 26Al production in KG 002 according to Monte Carlo-based depth- and radius-dependent production rates (Leya and Masarik 2009). In fact, the 26Al value of (116.1 ± 1.9) dpm kg−1, being the highest value ever measured in a Martian meteorite, can be explained solely by GCR-induced production.
Table 7. Cosmogenic radionuclide data and corresponding 1 σ-uncertainties for KG 002 (this work)
53Mn (dpm/kg Fe)
Data reported for the Los Angeles shergottite by Nishiizumi et al. (2000) measured at LLNL (marked *) and Berezhnoy et al. (2010) measured at PRIMELab and Munich (marked #) are shown for comparison. To make 10Be data directly comparable, the PRIMELab data have been renormalized by a factor of 1.045 to DREAMS data according to the recently found disparity (Merchel et al. 2012), whereas LLNL 10Be/9Be data are generally only about 0.4% lower than DREAMS data, and thus remain as they were published. This renormalization is based on pure experimental between-lab measurement differences using the same 10Be standard (Merchel et al. 2012).
When comparing the cosmogenic radio- and stable nuclides with the above-mentioned Monte Carlo calculations, an irradiation history for KG 002 can be reconstructed step by step:
The cosmogenic 22Ne/21Ne being 1.277 ± 0.009 excludes pre-atmospheric radii ≤ 25 cm (Fig. 12).
Calculated saturation activities of 10Be and 26Al cannot be brought into any overlap with experimental values of KG 002, thus pointing to a shorter CRE than necessary for saturation of 10Be (Fig. 14, upper left). Taking into consideration the CRE age from 38Ar determination and the empirical approach, 2.9 Ma, the 26Al activity is near the saturation level (94.2%), whereas 10Be is undersaturated (Fig. 14, upper right). Generally, CRE ages between 2.8 and 3.7 Ma (Fig. 14, lower left and right) are possible, the lower value asking for a smaller radius (approximately 35–50 cm), the higher values (3.5–3.7 Ma) correspond to a radius of approximately 65 cm. For all scenarios, the investigated sample originates from a position near the center. We note, however, that this scenario may not be unique. An alternative story might involve partial production of 26Al by SCR. However, this would require about 50% of total 26Al being of SCR-origin if comparing (22Ne/21Ne)c and 26Al production rates, which seems rather unlikely. Note, in this context, also the lack of evidence for significant SCR contribution to neon.
Another, also very sensitive, radionuclide for CREs in the 3 Ma region is the one with the longest half-life investigated here, i.e., 53Mn. However, having the highest measurement uncertainty (>20%), mainly due to the difficulties of isobar separation of 53Cr and lack of repeat measurements, the nuclide does not yet allow for constraining the CRE age of KG 002 with higher precision. A reduction in the uncertainty down to the 10% level would exclude the younger ages of the 2.8–3.7 Ma spread.
As the experimental 26Al is near the saturation level, a long terrestrial age influencing the 26Al concentration can be excluded. In essence, from the reasoning above, a simple one-stage irradiation model can explain all three radionuclides and the associated (22Ne/21Ne)c: Ksar Ghilane 002 has been of medium-sized radius (35–65 cm) and travelled 2.8–3.7 Ma through space before entering the Earth's atmosphere. The investigated sample was probably from an effectively shielded position near the center, thus, excluding SCR-induced nuclear reaction products as inferred for other shergottites. In this version, the CRE age of KG 002 from noble gases based on the empirical production rate, and for radionuclides based on the Monte Carlo calculated rates agrees well with the ages calculated for Los Angeles: (3.0 ± 0.4) and 3.3 Ma from 10Be and 2.8–3.1 Ma from several noble gases (Nishiizumi et al. 2000; Terribilini et al. 2000; Berezhnoy et al. 2010). However, the deduced meteoroid radii (20–40 cm), which were originally based on calculations assuming the same chemical composition for Los Angeles as QUE 94201, nearly exclude a common travel in a single meteoroid, asking for individual exposure of both meteoroids after a conjoint ejection event on Mars.
Finally, it is important to realize that significant systematic uncertainties exist in the calculation of cosmic ray exposure ages, in particular for meteorites with unusual chemistry. Surprisingly, whilst we have obtained consistent CRE ages for KG 002 based on the empirical approach for noble gases and the Monte Carlo approach for radionuclides, we obtain significantly lower CRE ages T21 and T38 if we consistently use the Monte Carlo approach and the shielding conditions (near the center) implied from the radionuclides: approximately 2.2 Ma (versus approximately 3.0 Ma). In fact, it is virtually impossible to arrive at approximately 3 Ma with realistic shielding conditions in the Monte Carlo model. Notably, the discrepancy holds both for T21 and T38, which essentially agree with each other in both approaches. Luckily, in some way, if we extend our approach to the noble gas data published for Los Angeles, we find the situation to be the same. Thus, most important in this context (and in a way that is reassuring), CRE ages for the two meteorites are indistinguishable, whatever the true value may be.
Ksar Ghilane 002, recovered in January 2010 in Tunisia, has been listed in the Meteoritical Bulletin Database as the 100th Martian meteorite fragment accepted by the Nomenclature Committee.
Based on oxygen isotope data and mineralogical studies, KG 002 is a coarse-grained basaltic shergottite similar to Los Angeles and NWA 2800 and is probably the feldspar (maskelynite)-richest rock among the Martian meteorites. Besides the major phases maskelynite and pyroxene, the rock also contains Fe-rich olivine, the two Ca-phosphates merrillite and Cl-apatite, silica (and/or SiO2-normative K-rich glass), amorphous K-feldspar, pyrrhotite, Ti-magnetite, ilmenite, and baddeleyite as minor or accessory mineral constituents. A typical feature of KG 002 is the high abundance of symplectitic areas composed of fine-grained intergrowths of fayalite, Ca-pyroxene, and a silica phase, which may have formed by the breakdown of pyroxferroite at low pressure during slow cooling. It should be noted that the symplectites in KG 002 (up to 2 mm in apparent size) are larger than in other Martian meteorites, indicating that the pyroxferroite crystals had to be very large prior to their breakdown.
As observed for most of the shergottites, all plagioclase in KG 002 has been transformed to maskelynite by shock indicating a shock stage of S5. The shock event producing this conversion is probably also responsible for the formation of impact melt patches of variable chemical compositions and He loss in the bulk rock.
Major element concentrations clearly indicate the basaltic composition of the rock. Ksar Ghilane 002 has CaO- and Al2O3-concentrations higher than that of other shergottites, such as Los Angeles and Dhofar 378, consistent with the highest modal feldspar abundance among these rocks. Considering the REEs, KG 002 has a similar pattern to that of Los Angeles showing a slight relative depletion of the light REE; however, the concentrations are about a factor of two lower than those in Los Angeles. The small positive Eu anomaly of KG 002 may indicate a complex magma genesis and might be explained by abundant plagioclase accumulation. On the other hand, the small positive Ce anomaly does not appear to be a primordial feature, but rather probably results from terrestrial weathering.
Our noble gas data suggest that gases with a composition inferred to be typical for the Martian interior (possibly mantle, e.g., Ott 1988) dominate the inventory of trapped noble gases. In addition, there may be some influence from elementally fractionated atmosphere. Although KG 002 does not represent a mantle rock, the source region of the magma (carrying the noble gas isotope signature) might thus be within the mantle. Our apparent crystallization ages are within the previously reported range for shergottites (approximately 100–600 Ma), and may result from resetting events. In terms of a comparison with Los Angeles, whilst we have observed differences in gas concentrations and isotopic ratios, the similar apparent crystallization age/resetting event and strong resemblance in CRE ages may provide evidence for a launch-pairing, which may have happened approximately 3.0 Ma ago (for the empirical production rates) or approximately 2.2 Ma (production rates from the Monte Carlo approach). A CRE age of 2.8–3.7 Ma of KG 002 does overlap with the earlier published ages (10Be, noble gases) of Los Angeles. Based on the Monte Carlo approach, the pre-atmospheric radius of KG 002 should have been between 35 and 65 cm with the investigated sample originating from near the center.
J. L. is grateful to ICREA Academia program. We thank U. Heitmann (Münster) for sample preparation. The DREAMS accelerator team, in particular S. Akhmadaliev, is gratefully thanked for helping with 10Be and 26Al measurements, and T. Faestermann and G. Korschinek for 53Mn and 60Fe measurements at Munich.