Multi‐zone fusion crust formation and classification of the 2004 Auckland meteorite (L6, S5, and W0)

On June 12, 2004, a meteorite passed through Earth's atmosphere and landed under the television in the living room of a house in Auckland, New Zealand. Textural characteristics, the chemistry of olivine (Fa23–24) and orthopyroxene (Fs20.7), and the bulk rock triple oxygen isotopes (δ17O + 3.1; δ18O + 4.2‰) from the interior of the completely unweathered (W0) 1.3 kg meteorite, hereafter referred to as Auckland, suggest it to be a strongly metamorphosed fragment from the interior of a low iron ordinary chondrite (L6) parent asteroid. The occurrence of maskelynite but shock fracturing of olivine and pyroxene indicates Auckland experienced extreme shock metamorphism (S5), likely during Ordovician fragmentation of the asteroid parent. The fusion crust consists of three zones: (1) an innermost zone containing narrow Fe‐Ni‐S‐bearing veins that migrated along pre‐existing shock fractures in olivine and pyroxene; (2) a middle zone in which the meteorite partially melted to form a silicate glass and immiscible blebs of metal and troilite, and is accompanied by unmelted silicate minerals; and (3) an approximately 0.1 mm wide vesicular‐rich outermost layer that largely melted, volatilizing sulfides, before quenching to form glass and olivine. Oxygen isotope values of the bulk rock and/or maskelynite of melted rim and modified substrate are 2–3‰ greater than the meteorite interior and indicate that up to 19% of terrestrial atmospheric O2 was incorporated into the fusion crust during the formation. The fusion crust migrated inwards as ablation occurred, enabling melting, migration, and re‐precipitation ± loss of sulfide and metal components, with the prominent glassy rim therefore forming from an already chemically modified zone.


INTRODUCTION
During passage through Earth's atmosphere, the air in front of a meteor is compacted, heated, and ablates the meteoroid, causing it to lose >75% of its mass and generate a fiery tail (e.g., Ceplecha & ReVelle, 2005;ReVelle, 1979). The resulting millimeter-thick dark fusion crust is a ubiquitous feature of meteorites that provides information on the conditions experienced during the last few seconds of the meteoroid's history (e.g., Brandst€ atter et al., 2008;Genge & Grady, 1999a, 1999bGenge et al., 2023;Pittarello et al., 2019;Ramdohr, 1967;Thaisen & Taylor, 2009;Unsalan et al., 2019). In addition, the fusion crust formation process may be analogous to the formation of micrometeorites, melted ablated spherules (Genge & Grady, 1999a) and chondrules (Hezel et al., 2015). A thin and dark glassy rim is commonly the most prominent feature of a fusion crust; however, these rims do not mark the extent of the zone of modification during fusion crust formation. The glassy exterior of the fusion crust is typically underlain by a thermally altered substrate, with the extent of alteration dependent on meteorite composition (e.g., Genge & Grady, 1999a).
On the morning of June 12, 2004, a meteorite crashed through the roof of a house, bounced off the sofa, hit the ceiling, and came to rest under a television in the Auckland suburb of Ellerslie, New Zealand (https://www. lpi.usra.edu/meteor/metbull.php?nwas=&strewn=&code= 73471). This meteorite (hereafter referred to as "Auckland") underwent no weathering after it landed and therefore the fusion crust is pristine other than the presence of several small chips that occurred during landing. Here, we use scanning electron microscopy (SEM)-energy dispersive X-ray (EDX) spectroscopy, electron backscatter diffraction (EBSD), and triple oxygen isotope analysis to classify Auckland and to show how the margins were isotopically and chemically modified as it passed through Earth's atmosphere. The typically mass-dependent triple oxygen isotopes provide evidence about the Solar Nebula compartments that different meteorite types formed and so are ideal for characterizing different meteorite classes and groups (e.g., Clayton, 1993;Clayton & Mayeda, 1983). A further notable feature of triple oxygen isotopes in meteorites, other than enstatite chondrites, is that they are distinct from those of Earth rocks and atmosphere (e.g., Pack, 2021). This means that incorporation of terrestrial O 2 into a fusion crust can be tracked by this isotope system (e.g., Clayton & Mayeda, 1989;Clayton et al., 1986;Hezel et al., 2015).

METHODS
The Auckland Museum in New Zealand provided several fragments of Auckland (≤5 mm in diameter) for analysis and permitted us to cut through the base to access the interior of the stone. A fragment from the edge of the meteorite and another from the interior were mounted in epoxy resin and polished for analysis in a Zeiss Sigma Field Emission Gun (FEG)-SEM at the University of Otago. Images were gathered using backscattered electron (BSE) imaging, and mineral compositions were measured using EDX analysis with an accelerating voltage of 15 kV, a beam current of 2.7 nA, and a live time of about 30 s. The beam current was monitored with analysis of cobalt metal and the chemical data were standardized against Smithsonian microbeam standards (olivine, hornblende, Cr-augite, chromite, fayalite, labradorite, andesine), with the secondary standards reproduced to ensure high-quality analyses. Chemical maps were made using a pixel dwell time of about 500 ms.
Electron backscatter diffraction maps were acquired at the University of Otago using a Zeiss Sigma VP FEG-SEM equipped with a NordlysF EBSD camera from Oxford Instruments. Acquisition was undertaken on the SYTON-polished sample using an acceleration voltage of 30 kV, an aperture of 300 lm resulting in a beam current of around 60 nA, 70°sample tilt, high vacuum mode, and a step size ranging from 1 to 3 lm. The raw data were processed using the Channel 5 software package.
All oxygen isotopic data are presented in per mil (&) units relative to Vienna Standard Mean Ocean Water and the d x O values are defined as (R sample /R standard À 1) 9 1000 such that R = 18 O/ 16 O or 17 O/ 16 O. Oxygen isotope data were corrected using calibration curves derived from analyses of international standards NBS-28 and NBS-30, and an in-house quartz standard ("Campo Longo Qz"), and were normalized to NBS-28 values. Oxygen isotope analyses were conducted at two facilities. The d 18 O values of the fusion crust and its interior (~2 mm within the fusion crust) were obtained following the CO 2 laser ablation-BrF 5 method of Sharp (1990) at GNS Science, New Zealand (Faure & Brathwaite, 2006). Oxygen was converted to CO 2 before isotopic analysis on a GVI IsoPrime mass spectrometer and samples normalized to the international quartz standard NBS-28 (+9.6&). During these analyses, replicates (n = 8) of NBS-28 quartz had d 18 O values that differed by less than 0.1&. Triple oxygen isotope analyses were conducted at the Queen's Facility for Isotope Research at Queen's University, Canada following the conventional method of Clayton and Mayeda (1963), with some modifications made to allow for measurement of 17 O. Approximately 5 mg of each sample was reacted with BrF 5 in nickel reaction vessels overnight at~600°C. Following this reaction, oxygen was transferred into glass sample tubes containing 5 A molecular sieve beads, which were cooled to liquid nitrogen temperature. Sample O 2 was purified prior to measurement using an ethanol/liquid nitrogen slurry (Clayton & Mayeda, 1983), which removes 14 NF + , a fragment of NF 3 that commonly interferes on m/z 33 (Miller et al., 1999;Pack et al., 2007). Sample O 2 was measured using dual inlet sample introduction and a Thermo-Finnigan MAT253 isotope ratio mass spectrometer. Oxygen isotope data were corrected using calibration curves derived from analyses of international standards NBS-28 and NBS-30, and an in-house quartz standard ("Campo Longo Qz"). Quality assurance and control were maintained by ensuring a minimum of 20% of analyses were of NBS-28 and by generating replicate measurements of unknown samples. The reference materials and replicate analyses were arbitrarily loaded among the 10 reaction vessels of the extraction line to avoid bias in our precision. Accuracies (1r) of 0.077 and 0.020& and precisions of 0.135 and 0.075& are reported for d 18 O and d 17 O values, respectively, based on analyses of NBS-28.

DESCRIPTION OF AUCKLAND
Auckland is an oblate 13 by 10 by 6 cm meteorite with a smooth asymmetric black-red fusion crust (Figure 1a,b). Flow lines and regmaglypts are visible on one surface, which is inferred to have been the front of the meteorite toward its last stages of flight. Where the meteorite was chipped during its landing, a pale speckled interior is exposed (Figure 1c). A slice through the meteorite reveals the fresh interior is surrounded by a slightly bleached rim and the glassy black fusion crust ( Figure 1d).

Auckland Interior
The Auckland interior contains relict chondrules up to about 1 mm in diameter ( Figure 2). Although the boundaries to the chondrules are mostly diffuse, some exhibit relict porphyritic or barred textures (e.g., Lofgren & Lanier, 1990;Weisberg, 1987) and others are largely maskelynite-and/or clinopyroxene-free ( Figure 2). Surrounding the chondrules is an interlocking matrix assemblage dominated by olivine, maskelynite, low-Ca pyroxene, apatite, merrillite, kamacite, high Ca-pyroxene, chromite, taenite, and troilite ( Figure 2b). The olivine and low-Ca pyroxene range in grain size up to~1 mm in areas that have very little maskelynite ( Figure 2).
The interior consists of olivine, low-and high-Ca pyroxene, and maskelynite. The chemical composition of olivine exhibits little variation, with an FeO wt% = 21.8 AE 0.5 (2 SD) and a corresponding Fa# 23-24  (Table 1). This small range includes olivine within the relict chondrules as well as olivine in the surrounding matrix. The low Capyroxene shows very little internal variation (Wo 1.4 AE 0.3 En 77.8 AE 0.6 Fs 20.7 AE 0.3 ) and indexes by EBSD as enstatite (Figure 2c). The maskelynite, confirmed to be an amorphous felspathic phase due to its inability to be indexed by EBSD (Figure 2c), displays very little variation (Ab 80.6 AE 0.8 An 10.7 AE 0.4 Or 8.6 AE 0.8 ; where Ab = 100 9 Na/[Na + Ca + K]; An = 100 9 Ca/[Na + Ca + K]; Or = 100 9 K/[Na + Ca + K]); however, it does exhibit a brighter color in BSE images which are adjacent to low-Ca pyroxene crystals. No crystalline plagioclase was observed.
Auckland contains minor apatite, merrillite, and high Ca-pyroxene. Merrillite grains are approximately 0.75 mm wide and are typically coarser than apatite ( Figure 2). The high Ca-pyroxene is indexed as diopside   by EBSD (Figure 2c) and occurs as~50 lm euhedral grains associated with olivine and orthopyroxene. An exception to this is the relict porphyritic chondrule, which is internally dominated by clinopyroxene (Figure 2).
Oxide and sulfide phases are well distributed (Figure 3). Chromite grains are less common, <200 lm in diameter, and are typically associated with large (up to 400 lm) taenite and kamacite grains (Figure 3). EBSD reveals that coarse kamacite and taenite grains are internally polycrystalline with kamacite consisting of strongly distorted aggregates up to 150 lm in diameter and taenite mostly consisting of small (up to 10 lm) elongated grains (Figure 3b,c), although the latter does not index well. There are also widespread irregularly shaped troilite patches up to 320 lm in diameter, as well as fine-grained taenite and kamacite. In some cases, troilite contains small exsolved blebs of kamacite.

Auckland Fusion Crust
Previous fusion crust-specific studies have subdivided these crusts into two domains: an outermost melted zone and a thermally altered substrate (Genge & Grady, 1999a, 1999bRamsdohr, 1967). However, we find it more convenient to describe the fusion crust on Auckland in three zones that are indicated by their different sulfide occurrence and metal behavior textures of sulfides and metals (Figure 4).
The outermost zone has a thickness of about 0.1 mm and is composed of elongate quench olivine crystals 50 lm in diameter and a vesicle-bearing glass that in places contains small unidentifiable crystals (relatively bright under BSE) that are likely magnetite or w€ ustite (Genge & Grady, 1999a, 1999b (Figure 4b). This zone corresponds to Genge & Grady's (1999a) melted crust. It grades into an about 0.5 mm wide zone consisting of small rounded troilite grains enclosed by olivine, orthopyroxene, and maskelynite ( Figure 4c). This zone is pervaded by amorphous glass veins, likely of feldspar composition, which has reacted with relict crystals in some areas. Round troilite crystals are mostly associated with the glass and contain exsolution lamellae and blebs of kamacite. The edges of olivine crystals are enriched in Fe (Figure 4b). The third zone, which is also gradational, consists of a 1.5-2 mm wide substrate subparallel to the melted crust, in which orthopyroxene, olivine, and chromite are pervaded by a network of thin troilite-rich veinlets ( Figure 4d) and adjacent blebs of kamacite. The kamacite-troilite grains in this zone commonly have thin troilite rims. Maskelynite typically lacks the penetrating Fe-Ni-S veinlets (Figure 4d). Similar vein textures have been observed in the modified substrate in ordinary chondrites (e.g., Genge & Grady, 1999a;Jenniskens et al., 2014). Genge & Grady (1999a, 1999b included the two inner zones as a single thermally altered substrate.

Oxygen Isotopes
The outermost layers of the fusion crust on Auckland have an average d 18 O value of 7.9& (Table 2). Due to the grain size and glassy appearance of maskelynite, it was the only phase that could be isolated from the bulk rock in either the fusion crust or the interior. The d 18 O values of maskelynite (7.4&) and bulk rock sampled within approximately 2 mm of the glassy crust (6.0&) are lower than the value (7.9&) of the outermost melted crust. These d 18 O values overlap with ordinary chondrites, enstatite chondrites, and carbonaceous chondrites ( Figure 5). However, the d 18 O and d 17 O values of maskelynite (5.3 and 3.4&, respectively) and bulk rock (4.2 and 3.1&, respectively) sampled at~20 mm from the margin are 2-3& lower than the values at the rim and show the meteorite to be distinct from enstatite or carbonaceous chondrites and to overlap the ordinary chondrite field ( Figure 5).
Ordinary chondrites can be further subdivided by petrologic type based on the degree of thermal metamorphism (e.g., Huss et al., 2006). Rb-Sr and U-Pb age dating indicates that the L chondrite parent bodies were metamorphosed at~4.4 to 4.5 Ga (e.g., Gopalan & Wetherill, 1968;Gray et al., 1973;Jenniskens et al., 2019), although the metal assemblages present indicate that the parent bodies were not large enough to undergo the core formation process characteristic of moons, planets, and large asteroids. The recrystallized olivine and low-Ca pyroxene and absence of a finegrained matrix surrounding distinct chondrules (Figure 2) indicates that Auckland thoroughly recrystallized in the parent body and can be categorized as petrologic type 5 or 6 (e.g., Huss et al., 2006). The occurrence of minor clear chondrule relicts (Figure 2), abundant plagioclase (now maskelynite) grains >50 lm (e.g., Van Schmus & Ribbe, 1968;Van Schmus & Wood, 1967), and the low modal abundance of recognizable chondrules (e.g., Rubin, 2000) collectively indicate that Auckland belongs to petrologic type 6. Superimposed on the Auckland metamorphic history is one or more shock events, of which the last catastrophically probably disrupted the parent body. Ar-Ar age dating of many L chondrite fragments indicates that a cataclysmic event occurred~470 Ma (e.g., Bogard et al., 1995;Haack et al., 1996;Korochantseva et al., 2007;Meier et al., 2017), although several L chondrites have much older ages, slightly different mineral and oxygen isotope chemistries, and different modeled source locations that indicate more than one L parent body . Compared to Creston, which is an L chondrite lacking a record of thẽ 470 Ma shock event , the mineral and oxygen isotopic chemistry of Auckland are more similar to the well-characterized L5 Park Forest (Simon et al., 2004), which belongs to the~470 Mafragmented parent.
Shock metamorphism-induced textures recorded on the meteorite vary based on mineralogy; for example, olivine and pyroxene in the interior of Auckland appear to have undergone extensive fracturing and remain fractured, kamacite and taenite have recrystallized into smaller sub-grains, and plagioclase is converted entirely to maskelynite. Maskelynite is mostly free of fractures compared to adjacent olivine and orthopyroxene crystals (Figure 4d), likely because it represents post-peak shock quenching of a plagioclase composition liquid (e.g., Burgin et al., 2023;Chen & El Goresy, 2000). St€ offler et al. (2018) report that the occurrence of maskelynite indicates a shock metamorphic (S) 5 stage, which is consistent with the polycrystallinity of large troilite and kamacite grains (Figure 3c,d; Bennett & McSween, 1996). The S5 textures indicate P > 45 GPa (St€ offler et al., 2018), although there is some uncertainty in the accuracy of P given that higher-grade ringwooditebearing assemblages have been inferred to have formed at lower P (Hu & Sharp, 2017). As Auckland was collected immediately after the fall event, it has not been affected by weathering and therefore is assigned a W0 weathering grade (Wlotzka, 1993).

Fusion Crust Formation
The Auckland fusion crust exhibits three zones that developed rapidly as the meteorite was ablated (Figure 6). These zones are distinguished by the textures of metals and sulfides. The thickest and innermost zone in the fusion crust is identified by the extent of the modified substrate (1.5-2.0 mm) and highlighted by the S and Fe element map (Figure 4a). Compared to the unmodified interior, this zone is mostly lacking small-to intermediate-sized sulfides and metals. Instead, the heavily shock fractured olivine and pyroxene grains are now filled with the precipitates from Fe-Ni-S liquids that migrated through these fractures (Figure 4d). This has resulted in the zone containing a troilite-rich vein network sub-parallel to the meteorite rim (Figures 4a  and 6a). Minor kamacite occurs in the troilite veins ( Figure 4d). An attempt to determine whether this zone is enriched or depleted in S, Ni, and Fe relative to the interior, using effective bulk compositions from element maps, was unsuccessful because the resulting bulk composition of the zone is biased by the occurrence of large relict Fe-Ni metals and our small sample size. However, similar fusion zones have been estimated to have undergone some extrusion of metal and sulfide components (Genge & Grady, 1999a;Genge et al., 2023).
Working outwards, another fusion crust zone occurs closer to the rim where the Fe-Ni-S veined substrate ends (Figures 4c and 6). Here, the fine-scale network of Fe-Ni-S veins and relatively large Fe-Ni metal-and troilite-rich areas that were likely present when the meteorite was larger (i.e., when the fusion crust was further away from the meteorite center) are absent. The Fe-Ni metals and sulfide minerals instead form small rounded grains of troilite, kamacite, and taenite associated with olivine and glass with a plagioclase composition (Figure 4b,c). The occurrence of round Fe-sulfide minerals and Fe-Ni metals enclosed within silicate glass indicate that formation conditions were sufficiently hot to generate melts with immiscible components, resulting in the metals and sulfide minerals precipitating as droplets. The exsolution of kamacite in troilite ( Figure 4e) likely represents Fe-Ni exsolution from troilite as the fusion crust cooled or oxidative S-loss during the liquid stage (e.g., Genge & Grady, 1999b). The implications of these textures in Auckland are that the sulfides and metals in this zone cannot have formed from a pristine interior of the meteorite, but rather from reworking of the Fe-Ni-S veined substrate.
The third reaction front in the Auckland fusion crust occurs at the interface of the metal-sulfide-droplet zone and the outermost~0.1 mm of the rock, which consists of quench-textured olivine enclosed in a vesicular glass, with minimal metal or sulfide minerals present (Figures 4b,c and 6). Although it once again proved impossible to get an accurate bulk composition for the domain because of the small amount of material available for analysis, the vesicles likely formed from evaporating compounds during atmospheric entry (e.g., Genge & Grady, 1999b;Genge et al., 2017). In this environment, sulfides should react with terrestrial O to generate SO 2 but leave behind metal. This marginal zone is therefore likely depleted in S.
As Auckland is a fall and has not been exposed to weathering, it is highly likely that the 2-3& increase in d 18 O value of the melted crust and the substrate is a consequence of reaction with Earth's atmosphere (~24&; Wostbrock et al., 2020). Melted crusts are usually very thin ((1 mm) and oxygen isotope data related to fusion crusts in the literature are limited to this portion of the meteorite, despite Ramsdohr (1967) and Genge & Grady (1999a, 1999b (Clayton et al., 1986;Hezel et al., 2015). The stable isotope evidence from Auckland demonstrates that atmospheric alteration penetrated several millimeters from the extensively melted margin.
During atmospheric entry, meteorites develop a fusion crust at very high temperatures. At these temperatures that melting occurs, likely >1000°C, equilibrium isotope fractionation would be negligible and we could expect that the fusion crust would have exchanged oxygen with the atmosphere and would have similar values (~24&). However, this is not the case because of the short residence time in the atmosphere and also because air O 2 exchanges slower than O 2 in H 2 O (Hezel et al., 2015). Simple conservative mixing calculations between two end-members (Auckland d 18 O value +4.2& and atmosphere +24&) show that up to about 19% of atmospheric oxygen was incorporated into the fusion crust and 9% into the thermally modified substrate of Auckland. In comparison, Clayton et al. (1986) and Hezel et al. (2015) report that less than 8% of atmospheric oxygen was incorporated in the fusion crusts in nearly all of their samples. Murchison, a carbonaceous chondrite, was an exception with 26% (Clayton et al., 1986). Hezel et al. (2015) considered the possibility that atmospheric O 2 exchange between fusion crust may be more complex than simple binary mixing and that there may be exchange of O 2 in the stratospheric ozone (d 17 O % d 18 O % 120&; Thiemens, 2006); however, the lack of 17 O data for the Auckland crust and substrate currently preclude us from investigating this possibility.

CONCLUSIONS
The meteorite Auckland, which fell in New Zealand's largest city in 2003 and is shown here to be an L6-S5-W0 ordinary chondrite, developed a mineralogically and chemically distinct multi-domain fusion crust upon interaction with Earth's atmosphere. During the unobserved luminescent phase, Auckland would have undergone extensive ablation. Examination of the fusion crust shows that three reaction zones formed. The fusion crust substrate zone most distal to the meteorite margin contains veinlets of troilite exsolving kamacite that show Fe-Ni-S liquids migrated along pre-existing shock metamorphism-induced fractures in silicate minerals. Closer to the meteorite margin, the meteorite partially melted and formed immiscible liquids that re-condensed as silicate glass or sulfides and metals. The outermost 0.1 mm layer almost melted completely, allowing S to volatilize, before quenching as glass and olivine. The d 18 O and d 17 O values of the fusion crust bulk and maskelynite fractions are higher than in the unmodified interior, which indicates increasing incorporation up to 19% terrestrial atmospheric O 2 . Oxygen isotope exchange also occurred in the substrate (~2 mm) incorporating~9% atmospheric O 2 . As small meteorites lose most of their mass while traveling through Earth's atmosphere, the reaction zones within the fusion crust migrate rapidly into the meteorite during the short duration (i.e., seconds) of the fireball phase.
Acknowledgments-The Auckland Museum generously provided samples of Auckland for our study. H. Grenfell helped provide images for Figure 1. We thank M. Genge and an anonymous reviewer for constructive reviews, and M. Zolensky for handling the manuscript. Open access publishing facilitated by University of Otago, as part of the Wiley-University of Otago agreement via the Council of Australian University Librarians.
Data Availability Statement-Data used for this study are available in the manuscript.