NORTHWEST AFRICA 773
(NWA 773 Clan)

Lunar Mingled Breccia
(fragmental breccia with clasts of very low-Ti olivine basalt,
olivine gabbro cumulate, fragmental breccias, and regolith breccias)

standby for northwest africa 773 photo
standby for northwest africa 773 photo
Found September 2000
~26° 49' N., ~12° 49' W.

While visiting the Western Sahara, American meteorite collector M. Killgore purchased three meteorite fragments from local nomads. The three fragments, weighing 50 g, 224 g, and 359 g (totaling 633 g), were all recovered in close proximity, and it is apparent they constitute a single meteorite. The find location was reported to be a desert plain near Dchira, Western Sahara. Subsequent paired finds include Anoual [5.92 g], NWA 2700 [31.7 g], NWA 2727, NWA 2977, NWA 3160 [34 g], NWA 3170 [60 g, photo courtesy of S. Ralew], NWA 3333 [33 g], NWA 6950 [1,649 g], NWA 7007 [91 g, photo courtesy of Kuehner et al., 2012], NWA 8127 [529 g], NWA 10656 [262.5 g], and NWA 10985 [250 g]. All of these paired stones exhibit a unique chemical signature in that they have the highest Sm/Eu ratios of any other basaltic lunar meteorite or basalt studied from the Apollo collection. For more information and photos of this lunar pairing group see Korotev–WUSL.

A detailed petrogenetic model for mare basalts has been presented by J. Day and L. Taylor (2007), a synopsis for which can be found on the NWA 032 page. This model, which demonstrates that NWA 032/479 could be launch paired with the Antarctic LaPaz (LAP) pairing group, was then expounded upon to explore the possibilities that the NWA 773 pairing group might also be derived from the same differentiated stratigraphic magma unit as the NWA and LAP samples (Hallis et al., 2007). Based on chemical compositions, mineralogies, textures, cooling rates, and crystallization and CRE ages, it was argued that the lunar pairing group of NWA 773 may represent the more rapidly cooled, cumulate-rich base of this magma unit, whereas the olivine basalt component, well represented in NWA 3160, derives from the lowermost layer adjacent to local pre-existing rock. The uniformly slow-cooled LAP samples are proposed to have crystallized in the middle of the flow, while the more rapidly cooled NWA 032 is consistent with crystallization at the upper margin. The NWA 6950 pairing member has a crystallization age that is ~100 m.y. older than NWA 773, inferring a possible faster-cooling position in the cumulate pile for this rock (Shaulis et al., 2012). However, in-depth studies of the NWA 032/NWA 4734/LAP pairing group mare basalts conducted by Elardo et al. (2014) led them to conclude that these meteorites were formed in a non-KREEPy source reservoir as opposed to the KREEPy source of the NWA 773 pairing group, ruling out any close relationship between them.

Northwest Africa 773 is primarily composed of two distinct lithologies: a light-green, magnesian olivine gabbro cumulate, and a dark-colored, polymict impact breccia containing a regolith component that comprises gabbroic, basaltic, and volcanic elements. The diverse clasts composing the breccia lithology, considered likely derived from a common magma unit, have been grouped into four categories by Fagan et al. (2014): olivine gabbro cumulate (OC); pyroxene–gabbro clasts (PG); symplectite (hedenbergite+fayalite+SiO2) clasts (S); and alkali-phase-ferroan clasts (AF). Boundary textures between the two lithologies in NWA 773 indicate that the cumulate lithology was a large clast within the breccia lithology (Fagan et al., 2003). Several independent analyses of a number of samples have determined a range of mineral compositions for the cumulate portion. It has a modal composition of 48–66% olivine, 26–40% clinopyroxene (comprising both low-Ca pigeonite and high-Ca augite), 8–15% interstitial plagioclase, 2% orthopyroxene, and ~1.5% K-feldspar, ilmenite, and RE-merrillite, along with trace amounts of Ba-rich K-feldspar (hyalophane), Cr-spinel, and troilite–FeNi-metal assemblages. Both the olivine gabbro cumulate and the breccia lithologies are reported to be enriched in incompatible trace elements. A breccia component enriched in incompatible trace elements has also been described in the paired stones NWA 3160 and NWA 3170.

The polymict breccia portion of NWA 773 has a fragmental texture that contains dispersed fragments of both magnesian and ferroan olivine gabbro, Ti-poor (VLT) olivine-phyric mare basalt, and Fe-rich lithic clasts comprising fayalite gabbros and fayalite granites, both representing late-stage differentiates. Other evolved clasts present in the breccia include FeO-rich symplectites (fayalite+hedenbergite+silica) that can be formed by two processes: 1) breakdown of pyroxferroite (a pyroxenoid formed by rapid cooling of a Mg-depleted melt), and 2) quenching of a silicate liquid. The second formation pathway of direct quenching of a melt is favored for NWA 773 due to the presence of feldspathic components in the silica (Fagan et al., 2003). Other unusual clast types have been identified, one of which contains fayalite, hyalophane, silica, and plagioclase, and another containing silica, K-feldspar, plagioclase, troilite, baddeleyite, and RE-merrillite. Mineral fragments present include fayalite, silica glass, agglutinate glass, and hedenbergitic pyroxene. The partial conversion of feldspar to maskelynite in NWA 773 reflects a low shock stage of S2, while the presence of some minor calcite filling fractures indicates only minor alteration consistent with a weathering grade of W1.

Although variable mineral proportions observed in studied samples initially indicated the cumulate portion was a norite or gabbronorite, the dominance of clinopyroxene over orthopyroxene in the plagioclase–orthopyroxene–clinopyroxene ternary diagram (Stöffler et al., 1980) suggests that this lithology is actually a gabbro. In a similar way, when the high proportion of olivine in this cumulate lithology is entered on the plagioclase–pyroxene–olivine ternary diagram (Stöffler et al., 1980), it indicates that this is an olivine gabbro. The plagioclase content of 14.2 vol% differentiates this rock from a peridotite (requiring <10 vol% plagioclase).

The olivine and pyroxene clasts within the breccia component exhibit a wide range of Fe contents constituting a magmatic sequence progressing from the magnesian olivine gabbro cumulate, to an extreme enrichment in Fe# in the ferroan symplectite and silica-bearing FeO-alkali clasts; these ferroan clasts have been compared by some researchers to terrestrial plutonic tholeiitic lithologies (Fagan, et al., 2002, 2013). The more ferroan olivine gabbro clasts contain no anorthosite (i.e., no highlands component) and has an LREE-enriched pattern with a strong Eu depletion—an unusual composition for VLT basalts, but one which exhibits some similarities to Apollo 14 basalts.

The olivine gabbro lithology lacks solar noble gases, but an enhancement in the solar gas content is observed in the breccia lithology, indicating that only the breccia resided at the lunar surface for any significant time prior to ejection. The CRE ages of these two lithologies support this finding: 5.2 (±0.8) m.y. for the olivine gabbro, and 68 m.y. for the breccia lithology. This age for the breccia is equivalent to a surface residence of 136 m.y. Lorenzetti et al. (2005) argue that after the breccia lithology was exposed near the surface, it was mixed with the cumulate lithology during an impact event, and was subsequently buried at a depth of a few meters for the relatively short span of 100 (±20) m.y. The Moon–Earth transit time is considered to have been <30 t.y., and according to Nishiizumi and Caffee (2010) such a short transit time corresponds to a launch event from a more shallow depth of <1–4.7 m.

The K–Ar chronometer associated with NWA 773 and other pairings reflects an age of 2.75 (±0.3) b.y. (2.865 ±0.031 b.y. calculated from Sm–Nd isochron), possibly indicating a late crystallization age (Burgess et al., 2007). This is the youngest age measured for any lunar rock studied to date. The average Pb–Pb age from baddeleyite grains from NWA 773 and other pairings, reflecting primary crystallization of the various pairings of this clan, is ~3.113 b.y. (Shaulis et al., 2012, 2013). The matching Pb–Pb ages for both magnesian and ferroan olivine gabbro cumulate components indicate that they formed during the same time period, while the comparative ages of the olivine gabbro lithologies and the polymict breccia attest to the fact that the constituents of the latter were derived from the former. Furthermore, the chronological dataset on the whole is consistent with all members of the NWA 773 clan being magmatically related (Shaulis et al., 2013).

A possible scenario for the formation of the olivine gabbro begins with the formation of a lunar magma ocean (LMO) to a depth of 400–500 km, and possibly as deep as 1,000 km (Ranen and Jacobsen, 2004). As cooling proceeded, differences in density began to show their effect. After 75% crystallization of the magma ocean, buoyant plagioclase-rich rock began to crystallize and float to the surface to form the original anorthite-rich plagioclase crust constituting the uppermost ~5–30 km of the magma ocean. At the same time, an increasingly more dense (i.e., a gradual decreasing ratio of Mg to Fe) and mafic compositionally-zoned lower crust was accumulated at depths of ~25–55 km. Thereafter, an unstable configuration resulted as rock of a higher density lay above rock of a lower density, and this resulted in convective overturn. This event in turn initiated the pressure-release remelting of early magma ocean olivine- and orthopyroxene-rich cumulates.

An anomalous KREEP-rich region of limited extent (16% of the surface), known as the Procellarum KREEP Terrane (PKT), was formed during the final phase of LMO solidification by extreme (>99.5%) fractional crystallization which occurred 4.492 (±0.061) b.y. ago. (Korotev 2005). This region is also thought to be the source of an ultramafic upper mantle melt that was the parent magma of the Mg-suite rocks. Through studies of Sm–Nd data from lunar samples, it was determined that solidification of the LMO was complete in 60–200 m.y. after its onset, at least by 4.417 (±0.006) b.y. ago (Boyet and Carlson, 2007; Grange et al., 2009). Based on Pb–Pb dating of zircon crystals, this LMO crystallization interval has been refined to 100 m.y. (Nemchin et al., 2009). Based on zircon crystallization studies, it was determined by Grange et al. (2009) that all magmatic activity in at least some locations (e.g., Serenitatis) had completely ended by 4.2 b.y. ago. According to Crow et al. (2011), the Pb–Pb ages of Apollo zircons show a peak at ~4.33 b.y.

It is possible that the parent magma assimilated material containing a high-K KREEP composition and a high LREE/HREE ratio, or alternatively, material with a high RE-merrillite composition. This assimilation produced the high incompatible element abundances present in the later-formed rocks. In contrast to this assimilation scenario, it was suggested by Shearer et al. (2005) that the observed compositional diversity of KREEP-rich magmas is more consistent with the addition of the KREEP component to the basaltic magma during the melting phase, prior to olivine crystallization. In support of the source mixing model, Borg et al. (2005) concluded that the 87Rb–86Sr ratio of NWA 773 would require 22% KREEP assimilation by the parental magma compared to only a 2% KREEP addition to the source magma. The low Fe content and high Mg# determined for NWA 773 are more consistent with a lower proportion of KREEP as predicted by the source mixing model. Still, it is now being considered by some that the KREEP component was incorporated into those specific lunar samples through a late impact into the Procellarum KREEP Terrane.

The parent magma underwent differentiation by fractional crystallization, and Ti-containing cumulus ilmenite was gravitationally extracted. The magma eventually formed plutons of Mg-suite material (petrogenetically distinct from the Mg-suite material from the PKT region), which then intruded the lower crust in some regions of the Feldspathic Highlands Terrane. This Fe-enriched magma may have been part of a shallow layered intrusive, or possibly a thick differentiated lava flow. Crystallization occurred as the melt cooled from ~1200°C to 1050°C. The upper mantle, composed of ultramafic, olivine-rich dunite or harzburgite, may have contributed melt material to the crystallization process of the Mg-suite rocks. The crystallization sequence from the base of the crust upwards was as follows: dunite (>90 vol% olivine) → troctolite (magnesian olivine + 10–60 vol% plagioclase) → norite (low-Ca orthopyroxene + 10–60 vol% plagioclase) → gabbro (high-Ca clinopyroxene + 10–60 vol% plagioclase) → anorthosite (>90 vol% plagioclase). This lunar crystallization sequence is unlike that of any terrestrial oceanic or continental basalt; in the terrestrial case, gabbroic high-Ca pyroxene crystallizes before noritic low-Ca pyroxene. The International Union of Geological Sciences—Subcommission on the Systematics of Igneous Rocks, having established a Working Party on the classification of lunar rocks, has adopted a Classification System for Lunar Rocks.

Northwest Africa 773 is probably derived from an upper mantle or lower-crustal parent melt that was mixed with a KREEP-rich melt component, possibly within a deep pluton. This mixture later intruded into upper crustal rock where crystallization under relatively rapid cooling conditions occurred within a shallow dike or sill (a hypabyssal rocktype), or possibly within a thick lava pool. Elemental distributions within the cumulate lithology provide evidence of cooling rates consistent with this scenario. Ultimately, differentiation of the melt produced the various components that are incorporated into the NWA 773 breccia lithology.

Jolliff et al. (2003) have found that close compositional similarities exist between NWA 773 and Apollo 14 Green Glass B, Type 1, and they suggest that NWA 773 may have originated from a parent melt in proximity to the source region of these picritic, green volcanic glasses, located within the Procellarum KREEP Terrane (PKT). It was proposed that the high KREEP concentration was incorporated as the melt transited from the mantle to the surface. As the melt cooled near the surface, crystallization of olivine and Ca-rich pyroxene was initiated and a trapped melt component of 15–25 vol% was incorporated. Various surface volcanic basaltic lithologies were mixed at this stage to produce the NWA 773 impact-breccia lithology. The high concentration of the heat-producing elements Th, U and K present in the PKT region could have permitted an extended period of melting and mixing that is consistent with the young age of NWA 773; however, VLT material has not been identified in significant amounts in this region.

In the PSRD article "Damp Moon Rising" by G. Jeffrey Taylor (July 2010), it was described how studies at the Carnegie Institute of Washington (McCubbin et al., 2010) and Okayama University in Japan (Yamashita et al., 2010) employed a technique called 'hydrogen manometry' to construct an OH calibration curve from apatite standards. This curve was then employed to determine the amount of water present in specific lunar meteorites such as NWA 773. Fluorapatite in NWA 773 and pairings contains 0.4 to 0.7 wt% (4,000–7,000 ppm) water, significantly more than previously supposed. This converts to a minimum of 0.7 to 1.7 wt% (7,000–17,000 ppm) water in the KREEP-bearing, late-stage magma from which NWA 773 and pairings were derived. In their study of apatite grains in NWA 773, Tartèse et al. (2014) found significantly higher water contents, and determined that the apatites could be resolved into two distinct groups: one characterized by moderate amounts of H2O in the range of 700–2,500 ppm, and another having high to extremely-high amounts of H2O in the range of 5,400–16,700 ppm. The extraordinarily high water content in some apatite grains of the brecciated lithology of NWA 773 has been conjectured to be related to silicate liquid immiscibility, possibly as a result of depletion of F and Cl in the Si-rich melt fraction.

For their calculations, Yamashita et al. (2010) presumed that apatite would have formed from such an evolved magma only after 90–95% crystallization, so they argued that the original basaltic magma would have proportionally contained 360–850 ppm water. Moreover, given the reasonable scenario in which 10% partial melting of the lunar interior occurred, the unmelted interior source region which hosted NWA 773 and pairings would have contained 7–17 ppm water. Although no longer thought to be dry, the Moon certainly contains much less water than the 500–1,000 ppm incorporated in the Earth.

Through remote sensing technology aboard the Clementine spacecraft, utilizing multispectral reflectance imaging, measurements of the mafic minerals olivine, pyroxene, and plagioclase feldspar, and indirectly, anorthosite, have been performed across nearly the entire lunar surface. Notwithstanding the proposal by Jolliff et al. suggesting the PKT region as a possible source location for NWA 773, a different gabbro-containing site located on the far side of the Moon, in the South Pole–Aitken (SPA) basin, has been identified by Clementine. This 2,500 km-diameter impact structure has had its upper crust completely removed, and a homogeneous melt sheet was formed. This large basin is thought to preserve ancient crustal rock that is mostly uncontaminated by subsequent basin ejecta transported over the Moon's surface (Petro and Pieters (2008). Subsequent impact events onto the SPA basin, such as those which formed the 64-km-diameter Bhabha crater and the 505-km-diameter Apollo basin, have excavated lower-crustal material from depths greater than 20 km. Although SPA is predominantly noritic in composition, a small rise known as ‘Olivine Hill’ is interpreted to be an olivine gabbro lithology. Volcanic activity associated with small mare ponds that occurred after basin formation is consistent with the presence of the VLT basaltic component identified in the NWA 773 breccia.

Calzada-Diaz et al. (2015) compared compositional and age data from a large number of lunar meteorites with elemental remote sensing data obtained by the Lunar Prospector gamma ray spectrometer, primarily for Fe, Ti, and Th, to better constrain the meteorite's source regions. For the basaltic breccia NWA 773, plausible ejection sites were identified in Mare Serenitatis, Mare Crisium, and the western boundary regions of Oceanus Procellarum, while Mare Fecunditatis was found to be inconsistent with the age data (see image below).


Image credit: A. Calzada-Diaz et al.
Meteoritics & Planetary Science, vol. 50, p. 219 (2015)

The lunar meteorites Y-793274, QUE 94281, and EET 87521/96008 share many compositional characteristics with NWA 773, and they may have experienced a similar petrogenesis. With the recovery of NWA 773, representing lower-crustal olivine gabbro, our ability to understand the Moon's early history has been greatly enhanced. The top photo above shows both the cumulate lithology and the breccia which constitutes the NWA 773 meteorite. The specimen on the left is a 0.094 g cut fragment of the black polymict breccia component, pervaded by fragments of the cumulate lithology and other diverse mineral and lithic clasts, while the specimen on the right is a 0.085 g cut fragment of the olivine gabbro cumulate component consisting of green to tan olivine crystals within pyroxene, interspersed with black chromite grains, and transected by a small shock-melt vein. An enlarged photo of this cumulate specimen is shown as well.