Meteor Crater is a bowl-shaped depression 180 meters (m) deep and 1.2 kilometers (km) in diameter located on the southern edge of the Colorado Plateau in north-central Arizona (Shoemaker and Kieffer, 1974). This impact crater formed ~50,000 years ago (Nishiizumi and others, 1991; Phillips and others, 1991) by the impact of a 100,000-ton iron-nickel meteorite, ~30 m in diameter, which struck at a speed estimated to be between 12 and 20 km per second (Shoemaker, 1960; Melosh, 1980; 1989; Melosh and Collins, 2005; Artemieva and Pierazzo, 2009). The crater and surrounding rim have since experienced limited erosion, making Meteor Crater one of the best preserved, young impact craters on Earth (Roddy and others, 1975; Grant and Schultz, 1993, Ramsey, 2002). Structural controls in the target rocks, induced by scissor-faulting of mutually perpendicular joints oriented northwest-southeast and northeast-southwest, give Meteor Crater a pseudo-rectangular shape in map view (fig. 1) (Shoemaker, 1960; Roddy, 1978; Ramsey, 1995).

Meteor Crater has historically been described as having a small amount of impact melt compared to other craters of similar size (Nininger, 1956; Shoemaker, 1960; Melosh, 1989). The apparent relative lack of impact melt has been attributed to everything from the volatile content of the target rock to the velocity of the impactor (Hörz and others, 2002; Melosh and Collins, 2005). Impact melt types at Meteor Crater include (1) ballistically dispersed melt bombs (~centimeter-sized) comprising mixtures of melted target rock and melted projectile; (2) shock-melted Coconino Sandstone rocks (for example, lechatelierite) found in the crater floor beneath alluvium and in the ejecta blanket; and (3) ballistically dispersed metallic spherules (for example, melt droplets from the Canyon Diablo meteorite, the impactor that created Meteor Crater). Though all these different melt types have been studied for decades, investigations by Hörz and others (2002), See and others (2002), and Mittlefehldt and others (2005) provided the first detailed chemical information about ballistically dispersed impact-melt particles at Meteor Crater.

See and others (2002) focused on major elemental analyses of the target rocks at Meteor Crater, and Hörz and others (2002) and Mittlefehldt and others (2005) carried out geochemical analyses of impact-melt particles and metallic spherules. Target rocks used in these studies were sampled either from within the crater or the overturned flap; impact-melt particles and metallic spherules were obtained from the Nininger Collection, archived at the Arizona State University Center for Meteorite Studies (Hörz and others, 2002; See and others, 2002; Mittlefehldt and others, 2005). The ballistically ejected materials were originally collected from the plains surrounding the crater, though no precise coordinates were recorded.

Impact-melt particles from this collection were commonly oxidized and (or) hydrated, likely occurring over the tens of thousands of years being exposed on (or close to) the surface. Pristine glass within these particles is rare. Hörz and others (2002) recovered few fragments that contained pristine glass; they described them as texturally homogeneous, clear in color, but compositionally heterogeneous, both within individual particles and from sample to sample. Based on geochemistry, Hörz and others (2002) divided the impact-melt particles into three compositional groups, inferring that these groups likely represented different target rock proportions (as well as meteoritic component) and therefore different depths of melting. Hörz and others (2002) suggested that either all three compositionally distinct melts were derived from a depth of <30 m, or that one group (a mixture of Moenkopi and Kaibab Formation rocks and a small amount of meteorite) was sourced at shallow depth and the other two (higher content of meteorite and Kaibab Formation rocks, with the addition of Coconino Sandstone rocks) originated from depths >90 m.

Additional geochemical analyses on impact-melt particles, including trace elements, were carried out by Mittlefehldt and others (2005). Their results indicated a melt zone that contained variable amounts of Moenkopi Formation, Kaibab Formation, and meteorite, with little to no input from Coconino Sandstone or Toroweap Formation rocks. It is important to note that analytical techniques differed between the Mittlefehldt and Hörz studies: Mittlefehldt and others (2005) measured bulk chemical analyses of the impact-melt particles, whereas Hörz and others (2002) used X-ray elemental maps and quantitative individual mineral and glass measurements. Different techniques may result in discrepancies between estimated depth of melting or input of differing target rock proportions.

The addition of impact melt analyses with known locations—both lateral distances from the crater and vertical depths within the ejecta blanket (compared to the previously studied impact melts that were collected on the surface without their locations recorded)—can be used to test how robust the melting depth hypotheses of Hörz and others (2002) and Mittlefehldt and others (2005) are and provide further information on the characteristics of impact-melt deposits at Meteor Crater. Furthermore, impact-melt particles sampled at depth from within the ejecta blanket can help determine whether impact cratering conditions allowed carbonate-rich rocks to undergo melting (retaining some CO₂), as opposed to exclusive devolatilization (all CO₂ vaporized), as well as the prevalence of secondary alteration at depth.

The U.S. Geological Survey (USGS) Astrogeology Science Center curates and houses the Meteor Crater sample collection. This collection includes more than 2,500 m of cuttings from 161 holes that were drilled into the crater’s rim, flanks, and ejecta blanket (Roddy and others, 1975). Depth of drill holes ranges from a few meters down to ~50 m, sampling material every ~0.3 m. A digital database detailing pertinent information for each collected sample (from the 161 drill holes), including drill hole identification number, depth, amount collected, and so on, has been made available through ScienceBase (a USGS trusted long-term data repository) as a published data release (Gaither and others, 2023).

Meteor Crater is an excellent terrestrial analog for studying planetary impact cratering dynamics, such as complex impact crater formation and morphology, impact-induced hydrothermal systems, impact melting of sedimentary target rocks, and mineral shock metamorphism. The USGS Meteor Crater sample collection provides the planetary science community free access to these geologic samples for petrographic, geochemical, structural, and other geologic analyses. Given the modern financial and logistical difficulties inherent in conducting thorough sampling campaigns at impact sites, continued preservation of this geologic collection and increasing access to the samples by the scientific community maximizes the prior financial and scientific investments of the USGS and NASA.

Work presented here utilized the Meteor Crater sample collection to characterize composition and textural information of impact-melt particles generated during the meteorite impact. Through this publication we showcase some of the unique impact-melt particles that were analyzed, including some uncommon features, such as lechatelierite (shock-melted Coconino Sandstone), to promote future research and use of this unique collection. Information on requesting samples from this collection can be found in Gaither and others (2023), as well as at https://www.usgs.gov/centers/astrogeology-science-center/science/meteor-crater-sample-collection.

The Meteor Crater sample collection has rock cuttings from 161 drilling locations; we narrowed down our sample selection to those from six drill holes: numbers 42, 45, 60, 94, 95, and 109 (fig. 1). These drill holes were noted to have a higher concentration of impact melt compared to other drill holes examined by Gullikson and others (2016). As part of our initial work described in Gullikson and others (2016), we sieved drill cuttings from these selected drill holes into different sizes, and any particles 5 millimeters (mm) in diameter and greater were separated into their respective lithologies (including impact-melt particles). A subset of these impact-melt particles, which typically ranged 5–10 mm in diameter, were then collected for further analysis. These particles all had a glassy appearance but varied in color (black, red, and yellow/orange). Most particles were collected from depths ranging from ~0.5 to 6 m below the surface. To conserve samples for future research, we limited our selection to 42 impact-melt particles for analysis. The samples chosen were considered to appropriately represent the impact melt population within the Meteor Crater ejecta blanket based on a qualitative assessment of size, shape, and color of the impact-melt particles first identified in the Gullikson and others (2016) work. It should be noted that there were many small (<5 mm), dark colored particles within the ejecta deposits that resembled impact-melt beads (spherical in shape), but upon further analysis they were identified as iron concretions. Due to time limitations, it was beyond the scope of this work to separate small impact-melt particles from these post-impact iron concretions; therefore, we focused on particles of sufficient size (that is, ≥5 mm) whose identification was easier to resolve.

Drill hole number and sample depth were recorded for each particle to preserve their original geolocation. Particles were gently broken into several pieces to expose their interiors, then placed into small cylindrical sample cups, and epoxy was poured into each cup. Cups were then placed in a vacuum to remove any bubbles. After epoxy cured and hardened, samples were extracted from sample cups and polished down until most of the individual fragments were exposed at the surface. We then used increasingly finer-sized grit sandpaper to polish until the sample had a smooth and even surface without any noticeable scratches. Between the change of grit size and after the final polish, samples were placed in an ultrasonic portable cleaner with deionized water to remove any foreign substances (that is, remnants from the polishing process). Each mount comprises a single impact-melt particle, broken into several smaller fragments, with interior surfaces exposed and polished for geochemical and imaging analyses (fig. 2). Sample names were assigned a letter in the order they were collected, such that the first sample identifier is mount A, the second is mount B, and so forth.

We used a scanning electron microscope (SEM) at Northern Arizona University for BSE imaging and energy dispersive X-ray spectrometry (EDS) to characterize our samples both texturally and compositionally. Standards used for calibration are listed in Gullikson and others (2024). “Quick” qualitative analyses, where a spectrum of characteristic X-ray patterns is obtained, were regularly done alongside BSE imaging to identify the phases present (referring to both mineral phases and pristine versus altered glass). For more accurate measurements, we ran full spectral profiles at longer count rates to detail a phase’s composition (specifically elemental percentages).

Using our EDS and BSE data, we identified the most representative samples for pristine glass, banded glass (alteration/hydration texture), metallic inclusions, barium sulfate, and carbonate inclusions for more precise measurements. Using an electron microprobe at Northern Arizona University, we carried out these analyses with wavelength dispersive spectrometry (WDS). WDS allows an increase in count rates, a lower detection limit for light elements (increasing resolution and precision for such elements), and a lower energy resolution that is ideal for providing reproduceable and reliable results for the selected phases.

Prior to analysis, samples were carbon coated to prevent charging (the accumulation of a negative charge on the sample from the incident beam) while in the instrument. For each phase analyzed, a specific set of standards was used for calibration. Calibration was tested periodically during data collection to ensure that the instrument's precision did not diminish over time. Instrument settings were changed depending on which phase was being analyzed. For example, glass analysis settings were 20.1 nanoampere (nA) beam current and 14.8 volts; settings for the analysis of carbonate inclusions were 10 nA beam current and 15 volts; and metallic inclusions were 50 nA beam current and 20 volts.

Ideally, oxide concentrations for spot analyses should equal 100 percent. There are a number of reasons why these analyses might yield higher or lower percentages, such as an unfocused incident electron beam when taking measurements, beam positioning (overlapping two phases), alteration (hydration), or that another element is present that was not included in the calibration. Therefore, analyses that resulted in totals of 97 percent or less, or 102 percent or more, were considered unreliable and not used. We anticipated our banded glass samples to return low totals because of alteration, though we still carried out these measurements for comparison to pristine glass.

Hydrated/altered impact-melt particles analyzed by previous researchers (for example, Hörz and others, 2002) were common, therefore it was surprising to find that nearly 75 percent of our analyzed melts had measurable amounts of pristine glass. Pristine glass compositions resembled a trend identified by Hörz and others (2002), where a distinct compositional “gap” was observed in Fe concentrations. Pristine glasses could be categorized into a “low Fe” (<12 weight percent FeO) group or a “high Fe” (>20 weight percent FeO) group. Target rocks at Meteor Crater are all low in Fe content (<2 weight percent FeO, See and others, 2002), indicating that a significant contribution of Fe from the meteorite was incorporated into the impact melts (Hörz and others, 2002).

Altered glass was identified in ~40 percent of our impact-melt particles. Interestingly, alteration of impact melts appears to be spatially non-uniform. We found that both altered and non-altered samples were present in sample bags from the same drill holes and depth, suggesting no spatial trend or correlation between the location of the impact melt, depth within the ejecta blanket, and whether the sample was altered. Impact melts collected on or near the surface and later analyzed by Hörz and others (2002) were interpreted to have been pervasively hydrated and oxidized long after the impact. However, more than half of our sample suite of impact-melt particles collected from drill cuttings do not appear to have experienced considerable alteration, based on texture and composition. Because altered impact melts occur in proximity to non-altered melts, there is the question of when alteration took place. If alteration is occurring in situ, does this illustrate how percolating fluids disperse when migrating through the subsurface? Compositional heterogeneities in impact melts reflect the chaotic nature of excavation and ejecta emplacement processes, resulting in unique thermal and geochemical histories for each individual impact-melt fragment (French, 1998).

Most impact melts have a crystalline groundmass that is composed of Ca-rich pyroxene and Fe-rich olivine, minerals found only in the impact melts. The Coconino Sandstone, the Kaibab Formation, and the Moenkopi Formation are all sedimentary in origin, and as such, olivine and pyroxene do not naturally occur in the vicinity of Meteor Crater (Shoemaker and Kieffer, 1974; Roddy, 1978; Morris and others, 2000). Therefore, the origin of these minerals is from the melting and mixing of the meteorite and target rocks (French, 1998; Hörz and others, 2002), creating a mafic melt ripe for the crystallization of olivine and pyroxene. The crystal habits of these minerals (skeletal or acicular) suggests these phases nucleated, grew, and quenched in a rapid manner (Lofgren, 1980). Many of the granular pyroxene crystals have inclusions of skeletal olivine, indicating that olivine nucleated first, followed by pyroxene.

The abundance of carbonate inclusions (CIs) and barium sulfate was an unanticipated find. Nearly 70 percent of analyzed samples had carbonate inclusions observed in more than one grain within each mount. We do not have an estimate on the percentage of barium sulfate inclusions within CIs, but it was a common occurrence. The origin of CIs in impact melts at Meteor Crater is still questioned, whether they are melted dolomite (from the Kaibab Formation) that did not volatilize during the time of the impact (Osinski and others, 2008, 2015), or precipitated through low-temperature secondary processes, long after the impact (Hörz and others, 2002, 2015).

Carbonates analyzed by Hörz and others (2002, 2015) render slight differences in composition compared to most of the CIs we analyzed. One primary difference was the concentrations of minor elements, such as Mg and Si, in addition to traces of the meteorite. Hörz and others (2002, 2015) determined that their population of carbonates was calcite, with anomalously high SiO₂ concentrations (2–7 weight percent SiO₂), typically <2 weight percent MgO, and completely void (or below the detection limit) of FeO and NiO. This, in addition to being attributed to vesicle infill, led Hörz and others (2002, 2015) to conclude these carbonates were the product of secondary processes (such as through the weathering of soils and rocks by rainwater, low-temperature calcium carbonates can precipitate). The presence of SiO₂ and Al₂O₃ impurities can be explained by the presence of clay minerals that are typically intermixed with high-desert soils (Sancho and others, 1992).

CIs found in impact melts recovered from deep within the ejecta blanket ranged from ~20 weight percent MgO to negligible amounts (fig. 11), with the average being ~4 weight percent MgO. Some melt particles had a carbonate coating or rind, which yielded slightly higher amounts of SiO₂ and Al₂O₃ when compared to CIs (fig. 12) and averaged 3 weight percent MgO. Minor amounts of FeO and NiO were detected in both the rinds and inclusions. Ni concentrations for most inclusions and rinds were below the detection limit of the electron microprobe (which is ~0.03 weight percent), though of the total number of carbonate inclusions we measured, ~28 percent had NiO concentrations of 0.04–1.3 weight percent. FeO concentration ranged from negligible amounts to around 1.5 weight percent FeO, with one outlier of 3.4 weight percent.

We infer that CIs situated within or along the margins of partially melted lithics (for example, Kaibab Formation clasts, with dolomite and quartz present) likely have an impact-related origin based on their spatial relationship to such lithics (fig. 10B,C). On the other hand, CIs that are not associated with partially melted lithics or individual mineral grains, but instead have barium sulfate inclusions, may differ in their origin. The euhedral, tabular habit of barium sulfate crystals within CIs does not support a pre-impact origin in which these crystals were entrained in the impact melt; rather, they likely precipitated in situ along vesicle walls prior to or concurrently with carbonate infill (Brock-Hon and others, 2012). The occurrence of barium sulfate in arid environments has been observed as infill within pore-void spaces, as well as precipitating as tabular crystals lining circular pore spaces that are partially filled with calcium carbonate, typically around 1–2 m depth within the soil horizon (Mees and Tursina, 2010; Brock-Hon and others, 2012).

Mittlefehldt and others (2005) carried out trace element analyses of target rocks and some impact melts and found that barium concentrations were typically less than 400 μg/g in target rocks, whereas impact melts ranged from <100 to 13,400 μg/g. Impact-melt particles enriched in barium were noted by Mittlefehldt and others (2005) to have secondary minerals infilling vesicles and coating the exterior of the fragments, providing a likely explanation for the influx of barium. A possible source for increased concentrations of both barium and sulfur could be dust. Arid environments commonly have dust coating the ground, which can undergo dissolution, leading to the mobilization of barium and sulfur ions that can then re-precipitate at depth (Brock-Hon and others, 2012).

Isotopic analyses for C and O in CIs (for example, Hörz and others, 2015) and S (for example, Craddock and others, 2008) in barium sulfate inclusions can be used to derive the origin of these phases, whether they are related to the impact or from post-impact secondary processes.

Target rocks at Meteor Crater are rich in quartz (that is, Coconino Sandstone, thin layers of calcareous sandstone within the Kaibab Formation, and siltstone and sandstone comprising the Moenkopi Formation), causing silica to be ubiquitous in impact-melt particles. Quartz grains within each of the target rock groups have experienced a range of pressures and temperatures generated during the contact, compression, and excavation stages of crater formation. These conditions led to different stages of shock metamorphism within the silica structure. Kieffer (1971) details the different classes of shocked silica-rich rocks, ranging from weakly shocked and fractured quartz grains to the conversion of coesite or stishovite (high pressure and temperature silica phases), and finally shock-melted glass (lechatelierite). Much of the texture variations of silica are correlated to these classes (Kieffer, 1971).

Small, rounded to irregular in shape metallic inclusions (MIs) were the last features described in this report. MIs we have analyzed here are most like the impactite metallic particles described by Kelly and others (1974). We observed two and sometimes three different phases within the bounds of each metallic inclusion and labeled them as dark, intermediate, and bright (which is how they appear in the SEM). MIs consist predominately of the bright phase, an Fe-Ni alloy that varies in concentration between these two metals (fig. 15) and displays a skeletal texture (fig. 14) (Vdovykin, 1973). The intermediate phase, an Fe-Ni sulfide, is interstitial among the Fe-Ni alloy. The dark phase is the least common and has been observed both in the interstitial spaces and as inclusions along the interior boundaries of the metal alloy (fig. 14).

Using a SEM and electron microprobe, we characterized 42 impact-melt particles that were acquired from the USGS Astrogeology Science Center Meteor Crater sample collection. We carried out detailed imaging for all 42 particles, EDS compositional analyses for 30 of these samples, and WDS compositional analyses for 21 samples. Geochemical data includes the compositions of glass, carbonate inclusions, metallic inclusions, barium sulfate, groundmass (both olivine and pyroxene), and a miscellaneous category for uncommon phases (for example, zircon, potassium feldspar). The full set of WDS and EDS compositional data and annotated BSE images are available in ScienceBase as a data release (Gullikson and others, 2024) and can be found at https://doi.org/10.5066/P9OGAJ8P.

The complete Meteor Crater sample collection database can be accessed at https://www.usgs.gov/data/meteor-crater-northern-arizona-drill-hole-sample-collection-1970-1973-and-curation-2010-2013 (Gaither and others, 2023). All samples within the collection, including impact-melt particles highlighted in this report, can be acquired for scientific research. More information on how to request samples can be found at the USGS Astrogeology Science Center Meteor Crater sample collection website (https://www.usgs.gov/centers/astrogeology-science-center/science/meteor-crater-sample-collection).