Twelve individual fragments constituting a single meteorite, with a combined total weight of 392 g, were visually distinguished by Greg Hupé from an assortment of meteorites he had purchased in Morocco. Samples from different stones were sent for analysis to Northern Arizona University (T. Bunch and J. Wittke) and the University of Washington in Seattle (A. Irving and S. Kuehner). A preliminary analysis found similarities to known angrites, and a sample was sent to the Carnegie Institute, Washington D.C. (D. Rumble III) for O-isotopic analysis. By this method it was verified that these meteorites were in fact a new sampling of the angrite parent body. Because of the importance and uniqueness of this find, a sample from each of the twelve fragments was submitted for analysis. Numerous other pairings have been independently analyzed and given separate NWA series numbers, with the total combined weight of this pairing group being ~7.8 kg.
Only a small number of unique angrites are currently represented in our collections, which some investigators have resolved into four subgroups: basaltic/quenched, sub-volcanic/metamorphic, plutonic/metamorphic, and dunitic. In a study based on a comparison of Hf/Sm ratios for a diverse sampling of both angrites and eucrites, Bouvier et al. (2015) inferred that these meteorite subgroups reflect the existence of three distinct crustal reservoirs on their respective parent bodies. These three reservoirs reflect similar chemical differentiation processes on both parent bodies: 1) subchondritic Hf/Sm ratios for the Angra dos Reis angrite and the cumulate eucrites (such as Moama); 2) chondritic Hf/Sm ratios for the quenched angrites (such as D'Orbigny and Sahara 99555) and the basaltic eucrites; 3) superchondritic Hf/Sm ratios for the sub-volcanic and plutonic angrites (NWA 4590 and NWA 4801, respectively) and the unusual cumulate eucrite Binda. The metamorphic NWA 2999 pairing group was not included in the Bouvier et al. (2015) study. Moreover, since Zhu et al. (2019) determined that the absolute MnCr age for angrites (4.5632 [±0.0003] b.y.) is slightly younger than that calculated for Vesta (4.5648 [±0.0006] b.y.), which indicates a delayed mantlecrust differentiation stage for the APB, they reasoned that the APB was probably larger than Vesta.
In contrast to other angrites, NWA 2999 exhibits a polygonal-granular texture consistent with a relatively slowly-cooled and annealed lithology, more similar to the sub-volcanic and plutonic angrites than to the quenched angrites. Evidence in support of a plutonic origin for NWA 2999 can be found in the homogeneous pyroxene compositions compared to the wider compositional range that exists in some other angrites (Kuehner et al., 2006). However, evidence also exists for an extended residence within a regolithlarge (up to 6 mm) anorthite, spinel, and olivine rock fragments are present within the fine-grained groundmass. Moreover, while other angrites contain only minor FeNi-metal (<2 vol%), the NWA 2999 pairing group contains up to 9 vol% (NWA 3164 pairing) coarse FeS and FeNi-metal having chondritic abundance patterns (Baghdadi et al., 2015). The FeNi-metal in NWA 2999 has elemental ratios that are inconsistent with what would be expected from incomplete core separation (Jambon et al., 2012), and neither could this high abundance of metal have been derived through partial reduction of iron. Instead, it is considered more plausible that the FeNi-metal was incorporated from an exogenous source during an impact event on the angrite parent body. The impactor was most likely a metallic object unrelated to any known chondritic or iron chemical group (Humayun et al., 2007; Jambon et al., 2012).
Consistent with this finding, an increased level of other siderophile elements such as Co, Ir, and Au support the presence of a significant meteoritic component. However, it is unknown if this exogenous FeNi-metal source can also explain the increased Mg content and the reduced concentration of refractory elements (e.g., Ca, Al, and Ti) observed in this angrite. Since a chondritic impactor would also have necessarily carried an O-isotopic composition close to that of the TFL, an alternate scenario was proposed by Gellissen et al. (2007) and then by Irving and Kuehner (2007) to account for the observed anomalous elemental abundances. They suggest that a large impact onto the angrite parent body occurred, perhaps by an evolved iron object, which created a mixture of diverse lithologies from within the angrite target body. These diverse lithologies which constitute NWA 2999 were then deeply buried (~120 cm based on depth profiles of 22Ne/21Ne ratios; Nakashima et al., 2018) where they underwent thermal metamorphism and annealing to produce the observed granulitic texture. The chemical composition of the NWA 2999 pairing group shows that it derives from a picritic source magma, which thereafter experienced further fractional melting, metamorphism, and annealing, along with incorporation of an exogenous metal component (Baghdadi et al., 2015).
Northwest Africa 2999 preserves some unique metamorphic features (previously observed in some terrestrial metamorphic rocks) which initially were thought to reflect a decompression stage followed by rapid cooling. Investigators presumed that these events were initiated during an extensive multi-km-deep thrust faulting event on a large parent body, postulated by some to be Mercury (Irving et al., 2005). These metamorphic features include the presence of clinopyroxenespinel symplectites between plagioclase and olivine clasts (reflecting a decompression phase), and plagioclase coronas surrounding portions of spinel grains (reflecting a rapid cooling phase).
An alternate explanation for these unique metamorphic features has been proposed by Ruzicka and Hutson (2006), who argue that under low-pressure oxidizing conditions at various degrees of melt formation, both plagioclase coronas and clinopyroxenespinel symplectites can be produced as cooling proceeds. Improved models of these symplectite and corona textures by Irving and Kuehner (2007) led them to conclude these features are more likely the result of the percolation of a S-bearing fluid during a metasomatic phase. These unique corona microstructures have been further interpreted by Baghdadi et al. (2012, 2013), who reason that a granulitic peridotitic lithology, which originated either as a slowly-cooled pluton or possibly as an annealed brecciated rock at depth (P <0.9 GPa), was intruded by a hot magma that increased the temperature to 10001200°C. This thermal event resulted in the formation of solid state metamorphic coronas at mineral contacts ("contact metamorphism") over an extended time interval through the following reaction: clinopyroxene + spinel ⇒ olivine + anorthite (with the reverse reaction occurring upon re-cooling). Eventual ejection of this angrite from its likely planetary-sized parent body produced the fracturing observed within the coronas and throughout this meteorite.
Other features consistent with a very rapid melting and cooling event on the angrite PB have been identified in the angrite NWA 4590. Glass present along mineral grain boundaries attests to a late mobilization of primary phases consistent with a decompression event (Kuehner and Irving, 2007). It has been postulated that the angrite meteorites might represent the impact-related dissemination of a more FeO-rich outer layer during the early history of Mercury, thereby explaining the chemical and mineralogical differences observed on Mercury compared to the angrites; e.g., the higher FeO-abundance in angrites compared to that on the present surface of Mercury, and the reversed Fe/Mn values for both olivine and pyroxene as compared to those of other planetary bodies. Nevertheless, even accepting the occurrence of collisional-stripping of a hypothetical FeO-rich basaltic (angritic) crust on Mercury, Hutson et al. (2007) find it implausible that Mercury initially differentiated under oxidizing conditions to form the angritic crust, and then subsequently differentiated under reducing conditions to form the surface that we observe today. They have also argued that other mineralogical features identified in angrites (e.g., reaction coronas), which on one hand may be attributed to rapid decompression on a planetary-sized body such as Mercury, may just as well be consistent with the typical cooling processes that occurred during crystallization of a melt.
In contrast to some other angrites, neither kirschsteinite nor orthopyroxene has been found in NWA 2999, and vesicles are absent. Based on HfW systematics, NWA 2999 formed ~5 m.y. later than Sahara 99555 and D'Orbigny (Markowski et al., 2007). However, Jambon et al. (2012, #1758) contend that due to the exogenous FeNi-metal present in this meteorite, the HfW chronometer is not reliable. It was concluded by Kleine et al. (2009) that both NWA 2999 and AdoR were derived from a parental source magma that had higher HfW than other angrites, likely the result of extended differentiation after core formation. A precise crystallization age based on the MnCr system indicates an age for NWA 2999 of 4.5579 (±0.0011) b.y., indistinguishable from that of AdoR and LEW 86010. As deduced by Shukolyukov and Lugmair (2008), two age clusters encompass all of the angrites studied thus far, and this attests to a very early period of magmatic activity.
A CRE age of 73.4 (±6.6) m.y. was calculated for NWA 2999 by Nakashima et al. (2008), while an age of 69.6 (±11.2) m.y. was calculated for the paired NWA 4931. A more precise noble gas analysis conducted by Nakashima et al. (2018) established a CRE age for NWA 2999 and NWA 4931 of 47.2 (±6.1) m.y. and 51.7 (±6.4) m.y., respectively. Multiple episodes of impact, disruption, and dissemination of the crust can be inferred by the wide range of CRE ages determined for the angrites<0.256 m.y. for thirteen angrites measured to date, possibly representing as many ejection events (Nakashima et al., 2008; Wieler et al., 2016; Nakashima et al., 2018). This range is consistent with a single large parent body enduring multiple impacts over a very long period of time, which would suggest that the parent object resides in a stable orbit (planetary or asteroid belt) permitting continuous sampling over at least the past 56 m.y. Alternatively, Nakashima et al. (2018) consider it plausible that there is currently at least two angrite (daughter) objects occupying distinct orbits: one representing the fine-grained (quenched) angrites with the shorter CRE age range of <0.222 m.y., and another representing the coarse-grained (plutonic) angrites with the longer CRE age range of 1856 m.y. (see diagram below).