NORTHWEST AFRICA 2737
‘Diderot’

Nine rock fragments constituting a single 611 g stone were found in the Moroccan Sahara by meteorite hunters under the organization of Bruno Fectay and Carine Bidaut. These black fragments were not recognized as meteoritic until several years later, at which time a sample was submitted for analysis. The importance of this meteorite was soon evident, and multiple analyses were conducted through a collaboration of research groups in France and elsewhere:

1. Laboratoire des Sciences de la Terre, (Centre National de la Recherche Scientifique—Unité Mixte de Recherche [CNRS–UMR]), École Normale Supérieure de Lyon: P. Beck, Ph. Gillet, B. Van de Moortele, B. Reynard
2. Université de Bretagne Occidentale, Institut Universitaire Européen de la Mer: J-A. Barrat, J. Cotten
3. Institut français de recherche pour l'exploitation de la mer (IFREMER), Centre de Brest (Centre national de la recherche scientifique—Unité mixte de recherche [CNRS-UMR]): M. Bohn
4. Planetary and Space Science Research Institute, Open University, United Kingdom: I.A. Franchi, R.C. Greenwood
5. Johnson Space Center, Houston, Texas
6. University of Tokyo, Japan

With the assigned name of NWA 2737, this olivine–chromite cumulate was classified as a dunite that exhibits many mineralogical, chemical, and petrographical similarities to the first known martian chassignite that fell in Chassigny, France in 1815. A third chassignite designated NWA 8694 (photo courtesy of L. Labenne) has been classified at the Museum National d'Histoire Naturelle, Paris, France (Hewins et al., 2014, 2015). Northwest Africa 2737 is composed of 89.6 vol% shock-blackened, cumulus forsteritic olivine, 3.1 vol% chromite, 1.6 vol% sanidine, 1.0 vol% pyroxene (augite, pigeonite, and orthopyroxene), and 0.2 vol% phosphate. Carbonates are present in both chassignites, with at least some of them showing evidence for a martian origin—shock fractures post-date the carbonate formation (Beck et al., 2006). Both chassignites have similar REE patterns, elements that are carried mostly in their apatite component. A martian origin is attested by the abundance ratio of Fe/Mn in olivine, the bulk Na/Al ratios, and by the O-isotopic ratios.

Northwest Africa 2737 is one of the least terrestrially altered martian finds as shown by its trace element composition (e.g., low Ba and Sr contents) and noble gas signatures. However, in contrast to Chassigny and the nakhlites, NWA 2737 contains trapped noble gases more similar to those in shergottites, as evidenced by the lack of martian mantle Xe, the lack of fission-derived Xe from plutonium, and the enrichment of martian atmospheric Xe (Marty et al., 2006).

While feldspars in Chassigny are mostly composed of plagioclase, with lesser amounts of labradorite and sanidine, the feldspars in NWA 2737 only occur as K-rich sanidine or Na-rich analbite. These differences could have been caused by a relatively low Al concentration in the parental melt, resulting in the delayed nucleation of plagioclase with a consequent buildup of Na in the melt (Beck et al., 2006; Papike et al., 2009). The olivine and chromite grains in both chassignites contain melt inclusions with a unique mineralogy, including the occurrence of hydrous kaersutitic (Ti-rich) amphibole, while those in NWA 2737 also include an alkali feldspar-rich glass. Six distinct melt inclusion species were described in NWA 2737 by He et al. (2010,2013) based on the diverse mineral assemblages, including olivine, orthopyroxene, augite, pigeonite, kaersutite, chlorapatite and fluorapatite, biotite, chromite, pyrrhotite, and feldspathic glass; these minerals occur in melt inclusions that span a wide range of sizes (~5–300 µm). It is thought that some melt inclusions might contain trapped parental magma at various stages of fractionation.

It was determined experimentally that the parental magma of the chassignites may have resembled a terrestrial, silica-saturated hawaiite magma with a higher than terrestrial Mg# [= molar MgO/(MgO + FeO)] and an aluminum content of ~12 wt% (Filiberto, 2008). It was ascertained that the chassignites likely crystallized after ~30% crystallization of mafic phases was attained.

In an effort to estimate the original magma composition of the cumulate chassignites, He et al. (2013) examined the ubiquitous, sub-mm-sized, trapped magmatic inclusions in olivine grains. In consideration of the water diffusion coefficient between amphibole and melt, as well as the extent of melt crystallization that occurred prior to amphibole formation (~45%) and other pertinent assumptions, the water content of the primary melt of NWA 2737 was calculated to be ~0.48–0.67 wt%; the water content of the primary melt of the Chassigny meteorite was previously calculated by McCubbin et al. (2010) to be 0.43–0.84 wt%. Utilizing the MELTS program with the known parameters, the research team determined that NWA 2737 was formed under pressure of ~6.8 kbar. Their deduced parental composition for NWA 2737 based on Ca:Al ratios and Mg# is quite similar to that of the martian basalt Humphrey, measured in situ by the MER rover Spirit in Gusev Crater. A broad range of Al contents has been found in martian parental source rocks—Al is relatively high in basaltic surface rocks and the chassignites, while it is thought to have been sequestered at depth in garnet in the nakhlites and most shergottites; this has implications for the magma ocean scenario of Mars' petrogenetic history.

Several differences between Chassigny and NWA 2737 have been identified by Mikouchi (2005): interstitial chromite has a larger grain size in NWA 2737, feldspar occurs as the sodic plagioclase albite in NWA 2737 rather than the calcic plagioclase anorthite, and all phases of NWA 2737 have a more magnesian composition than those in Chassigny, perhaps reflecting conditions of higher temperature and pressure during fractional crystallization from a less evolved parental melt as suggested by Nekvasil et al. (2005). Petrographic evidence indicates that NWA 2737 was the first cumulate to crystallize at the bottom of the pile, followed by Chassigny next in the sequence, while the order for NWA 8694 remains to be determined (McCubbin et al., 2013). In further studies, Mikouchi et al. (2016) found that significant ambiguities exist among the three known chassignites. For example, although each of the chassignites exhibit a similar cooling rate (0.003–0.1 °C/hr), olivine compositions between them show large variations: NWA 8694 is Fa46, Chassigny is Fa31, and NWA 2737 is Fa21; moreover, each chassignite exhibits a distinct shock history. Therefore, they suggest that each of the chassignites is more likely associated with a separate flow or lobe (possibly within a common extensive igneous unit) rather than a single sequential accumulation. In a subsequent in-depth comparative study of the three chassignites conducted by Hewins et al. (2020), it was concluded that NWA 8694 derives from a parental source melt distinct from the other chassignites. See the Nakhla page for further details on the stratigraphic sequence for nakhlites and chassignites.

Lorand et al. (2012) studied the sulfide mineralogy and native metal assemblages in NWA 2737 and made comparisons to Chassigny. While the bulk sulfide abundance in NWA 2737 is low, troilite (and rare pentlandite) blebs up to tens of µm in size are ubiquitous at primary mineral grain contacts, and as even smaller blebs associated with chromite within melt inclusions. Although the sulfide pyrite is abundant in Chassigny, it was not identified in NWA 2737, and the Cu-sulfide chalcopyrite was virtually nonexistent; both sulfide phases were likely lost during an impact-heating event. Similarly, Fe–Os–Ir alloys that are observed in close association with troilite grains in NWA 2737 are considered to have exsolved from pyrrhotite during subsequent cooling from an impact-heating event. These differences that exist between these two chassignites can be attributed to impact-shock-related reduction processes involving high temperatures, leading to the devolatilization of S and the loss of magnetic properties.

Overall, shock effects in Chassigny are only moderate (S4), but some portions experienced much higher shock pressures (45–55 GPa; S5–S6) as attested by the presence of planar deformation features and localized melting. By contrast, the olivine in NWA 2737 exhibits exceptional darkening which is thought to reflect the disordered lattice state of an incomplete transformation to olivine high pressure polymorphs, as well as the presence of nanophase FeNi-metal particles within the olivine (Reynard et al., 2006). These Fe nanoparticles are thought to have formed through subsolidus reduction of olivine during the high temperature phase (~1300°C) of a shock event corresponding to a minimum shock stage of S5. Employing micro-X-ray diffraction technique for strain-related mosaicity analysis, Li et al. (2021 #2428) verified a shock stage for NWA 2737 olivines of S5–S6 (>30–55 GPa).

A partial cause of the darkening (brown color) in olivines as described by Treiman et al. (2006) derives from the conversion of Fe⁺² to Fe⁺³ in an oxidizing environment, resulting from the shock heating loss of H⁺ that was previously dissolved in a hydrous magma. Only a low ppm abundance of Fe⁺³ would be required to cause the observed darkening. Continued investigations utilizing spectral reflectance techniques, including Mössbauer spectroscopy, indicate that the darkening is primarily the result of shock disseminated nanophase metallic iron particles (Pieters et al., 2007). Other deformed olivine grains in NWA 2737 have been recrystallized to a visually colorless phase. Takenouchi et al. (2015) suggested that the Fe nanoparticles were formed in association with high-pressure polymorphs such as wadsleyite during a high-pressure, high-temperature shock event. Ultimately, the wadsleyite was back-transformed to olivine during a period of high post-shock temperatures, and this olivine now appears brown due to the presence of the Fe nanoparticles. Subsequent studies were conducted by Takenouchi and Mikouchi (2016) of NWA 2737 and several shergottites that contain darkened olivine associated with shock-melt phases. They found higher Fe⁺³ ratios in both brown (or brownish) olivine compared to adjacent colorless olivine located within the same grain, attesting to the formation of Fe nanoparticles by olivine oxidation and reduction during transformation to high-pressure polymorphs such as ringwoodite and wadsleyite.

Cooling rate estimates for both Chassigny and NWA 2737 have been set at 28–30°C/year, consistent with a formation within a thick lava flow or a shallow dike or sill. The Sm–Nd data for both plot on a similar isochron—1.416 (±0.057) b.y. for NWA 2737, and 1.380 (±0.030) b.y. for Chassigny. Both of these chassignites, as well as the nakhlites, also have similar CRE ages of 10–12 m.y. based on 3He, 21Ne and 48Ar. Therefore, it may be argued that the chassignites and nakhlites were ejected during a common impact event from the same igneous region on Mars. Based on Ar–Ar data, it was proposed by Bogard and Garrison (2008) that NWA 2737 may have experienced an intense impact event resulting in its burial ~1–20 m deep under a warm ejecta blanket, while a subsequent, less-intense impact event was responsible for its ejection from Mars ~11 m.y. ago. In a noble gas study of chassignites and nakhlites, Nagao et al. (2019) found that the Xe–Kr systematics among the three chassignites are significantly different reflecting diverse formation conditions (see diagram below). The plot for NWA 2737 shows a mixture of interior and atmospheric noble gases likely acquired during the severe impact shock event, while that for NWA 8694 shows the effects of noble gas fractionation similar to the MIL nakhlites, probably inherited through aqueous alteration processes near the martian surface.

Differences have been noted in the gas retention ages of Chassigny and NWA 2737. After correction for terrestrial contamination, a more accurate determination of the K–Ar crystallization age for NWA 2737 was made, and this revised K–Ar age of 376 (±168) m.y. is significantly younger than that of Chassigny and the nakhlites, and more like that of shergottites (Marty et al., 2005; 2006). However, this discordant age has been attributed to the higher shock metamorphism experienced by NWA 2737. The specimen of NWA 2737 shown above is a 0.48 g cut end fragment. The photo below is an excellent petrographic thin section micrograph of NWA 2737, shown courtesy of Peter Marmet.