Introduction Cosmic airbursts and impacts produce a wide range of surface effects, with high-altitude airbursts, such as the 1908 Tunguska event, primarily generating blast damage without forming craters [1]. In contrast, low-altitude “touch-down” airbursts may induce surface melting, spherule formation, shocked quartz, and shallow cratering [2]. Due to preservation challenges, few airburst signatures are documented in the geologic record, limiting our understanding of these events. Here, we report findings from a site approximately 5.8 km east of Perkins, Louisiana (30.3980° N, 93.3535° W) (Figure 1), where an anomalous 300-m-long seasonal lake/depression (Figure 2) is associated with extensive deposits of likely impact-related materials (Figure 3). We tested the hypothesis that an airburst/impact at the Perkins site produced these materials. To understand these deposits, we conducted a multi-disciplinary study employing a comprehensive suite of microscopy and geochemical techniques to characterize the material, investigate its origin, and determine whether the Perkins site represents an airburst/impact cratering event. Figure 1: Figure 2: Figure 3:
Research history In ~1938, Walter Leo Fitzenreiter, Jr., father of the lead author (R.F.), first speculated that a seasonal lake on his property was an “impact crater” based solely on its shape and crater-like rim raised ~1 m above the surrounding terrain. Later, in 2006, lead author R.F. began to test his father’s suggestion of a crater and discovered several locations around the lake displaying large semi-consolidated masses of spherules associated with multi-cm-sized fragments of black vesicular glass, which, he reasoned, might result from a cosmic airburst/impact event. Beginning in 2006, lead author R.F. found very large quantities of meltglass and spherules in two deposits near the raised rim of the lake, ~300 m long and 120 m wide. From 2006 through 2024, lead author R.F. also hand-extracted 32 cores (~4 to 13 cm in diameter) and excavated two trenches (~3 m wide). Locations and results are described in the sections below. In 2007, after identifying numerous spherules and fragments of meltglass in the Vee and Pond deposits, R.F. first contacted one of the co-authors (K.E.), who analyzed the spherules and meltglass. His analyses suggested that they resulted from a cosmic impact event, and he recommended additional research. In 2011, R.F. next contacted members of the Comet Research Group, who analyzed these materials to investigate their characteristics and origin.
Information on airbursts Cosmic airbursts can occur at various altitudes, producing distinctly different surface effects. High-altitude airbursts, like the 1908 Tunguska event (~5-10 km altitude) [1], primarily cause widespread devastation through blast waves but typically leave no craters [2]. Even so, the airburst evidence at Tunguska has been reported to include spherules, shocked quartz, glass-filled fractured quartz, melted feldspar, carbon spherules, and glasslike carbon [1]. Such events, called “touch-down airbursts,” occur when a bolide explodes sufficiently close to Earth’s surface that high-velocity fragments, the shockwave, and the thermal pulse form meltglass, create spherules, generate shocked quartz, and potentially create shallow ephemeral craters [2]. Due to the scarcity of documented cases of airbursts and the frequent lack of evidence preserved in the geologic record, the effects of airbursts are still poorly understood. However, understanding them is vitally important because, as Boslough and Crawford [3, 4] wrote, “Low-altitude airbursts are by far the most frequent impact events that have an effect on the ground.” To provide additional insights into airburst processes in the geologic record, we document our discoveries at this new proposed airburst location near Perkins, SW Louisiana (Figure 1).
Methods We employed standard protocols for eight complementary techniques to characterize spherules, meltglass, and shocked quartz (for more details, see Supporting Information, https://zenodo.org/records/15496666 [5]). Optical microscopy (OPT). Transmission microscopy is used for 3D imaging. It also uses crossed polarizers to identify isotropic areas of melted silica in quartz grains. EPI-illumination microscopy (EPI) is used for 3D imaging of spherules and meltglass. It also reveals if a fracture in a quartz grain is filled with material but not the nature of the material. Scanning electron microscopy (SEM): Used to image and analyze spherules and meltglass. It also reveals if quartz fractures are filled with silica without determining if it was melted. Energy dispersive spectroscopy (EDS): A standardless technique used to determine the elemental and oxide compositions of spherules, meltglass, and the quartz fracture-filling material (e.g., melted silica, hydrated silica, carbon, other minerals, or polishing compounds). Laser ablation inductively-coupled plasma mass spectrometry (LA-ICP-MS): A laser beam and mass spectrometer were used to analyze the elemental and isotopic composition of spherules and meltglass. Cathodoluminescence (CL): Used to differentiate between amorphous and crystalline areas of quartz grains; non-luminescent (black) areas indicate melted silica [6–8]. Neutron activation analysis (INAA): This nuclear technique measures the elemental concentrations in a material by irradiating it with neutrons and measuring the resulting gamma rays. Universal stage (u-stage): This technique was used to index candidate shocked quartz grains, previously identified using a petrographic microscope. Indexing some quartz grains with a universal stage was problematic because they appeared thermally distorted. However, indexing was possible in those areas where the lamellae remained planar and parallel. The u-stage is considered the definitive technique for identifying shocked quartz. Argon-argon dating (40Ar/39Ar): This radiometric dating technique can determine the formation age of rocks that underwent rapid cooling to below their closure temperature since impact-related heating likely reset previous ages. Samples were neutron-irradiated for 6 hours, which is appropriate for ages from 1 Ma to 100 Ma.
Study objectives Material composition: analyze the material with a focus on the morphology and composition of Fe- and Si-rich spherules, meltglass, carbon spherules, melted minerals, and fractured quartz grains, with the latter as a widely accepted indicator of shock metamorphism [2, 9–12]. Dating: produce an age-depth model for Trench #1 to determine the age of the spherule-and-meltglass abundance peak and acquire calibrated radiocarbon ages on the carbon-rich meltglass surrounded by very large amounts of spherules. Lastly, determine whether or not the meltglass is co-eval with the spherules and meltglass in Trench #1. Formation conditions: estimate the temperatures and pressures required to produce these materials. Potential formation mechanisms: compare anthropogenesis, authigenesis, volcanism, tectonism, lightning discharges, and cosmic airbursts/impact. Hydrocode modeling: produce a cosmic impact model consistent with the observed evidence.
Results This section documents our analyses of the wide range of materials observed at the Perkins site. To investigate the characteristics of these materials, we used eight analytical techniques, as described in the Methods section. Below, we offer a short introduction before presenting the results for each proxy. Also, to assist in understanding the complex, multidisciplinary data presented in this study, we provide the following Table of Contents of the Results and Discussion sections: Table of Contents Results 1. Sedimentology and Stratigraphy
2. Age of Materials 2.1 Radiocarbon Dating 2.2 Argon-Argon Dating
3. High-Temperature Melted Spherules 3.1 Spherule Abundances 3.2 Spherule Distribution 3.3 Spherule Morphologies
4. Carbon Spherules and Glasslike Carbon
5. Meltglass
6. Low-oxygen Metallic Flakes
7. High-Temperature/Pressure Quartz 7.1 Shocked Quartz Grains with Glass-filled Fractures 7.2 Thermally Warped Quartz Grains 7.3 Miller-Bravais Indices Confirm Shocked Quartz 7.4 Microbrecciated Quartz Grains
8. Magnetic survey
9. Summary of Analytical Results for all Proxies Discussion-Interpretations 10.Age of the Perkins Materials
11.Origin of the Spherules, Meltglass, and Shocked Quartz 11.1 Anthropogenesis 11.2 Railroads and Logging Operations 11.3 Natural Gas Blowout 11.4 Authigenesis 11.5 Tectonism 11.6 Volcanism 11.7 Lightning 11.8 Older Airburst/Impact Event 11.9 Airburst/Impact Event
12.The Lake/Depression as an Impact Crater
13.Melted and Shocked Quartz during an Airburst
14.Hydrocode Model of a Touch-down Airburst
15.Comparison of the Perkins Site to other YDB sites and the K-Pg
1. Sedimentology and stratigraphy Louisiana coastal plain near-surface sediments down to several 100 m typically comprise alternating weakly to well-stratified beds of unconsolidated to semi-consolidated sand, carbonate-cemented sand, limestone, and colored clays [13]. At Perkins, stratigraphic profiles reveal massive, mostly non-laminated, unconsolidated deposits of sand. A sand grain analysis for Trench #1 shows that ~70% of the sediment is fine sand (63-200 μm) ( Supporting Information, Figure S1 [5]), dominated by quartz and limestone grains. Four primary deposits of glass-and-spherule-rich material were discovered (Figure 2): 1) the Vee deposit in the narrow-V-shaped southwestern end of the lake is ~10 m in diameter and ~30 cm thick; 2) the Pond site northwest of the lake is ~10 m in diameter and ~60 cm thick; 3) a smaller discontinuous deposit of meltglass and spherules was found at the Lake site; 4) Trench #1 on the east rim of the lake contains moderate amounts of glass-and-spherule-rich material. In addition, small amounts of the material were found along some of the roadways and paths where the landowner had placed them.
3. High-temperature melted spherules 3.1. Spherule abundances We collected samples from the nine cores, pits, and trenches and then extracted the spherules and meltglass, counted them, and analyzed selected examples. The quantities by depth for the nine locations are shown in Figure 6 and Supporting Information, Table S5 [5]. Based on field measurements and SEM imagery of spherule-rich samples, the estimated volume of glass-and-spherule-rich material in the Vee deposit totals ~23 m3, and the Pond deposit is 46 m3, for a total of 69 m3. Each m3 of the material weighs ~1500 kg, so the total mass is >100 tonnes. Quantification using SEM imagery of aliquots of the deposited material revealed that each kg is estimated to contain >100 billion spherules, so, extrapolating conservatively, the total mass of known material includes hundreds of billions of spherules, with the meltglass equaling ~1% of the total or >1 tonne. Figure 6: The spherule composition from the Vee varies, averaging 7.36 wt% Al, 8.15 wt% Ca, 17.05 wt% Si, and 10.12 wt% Fe with small concentrations of other elements. Scanning electron microscopy with energy dispersive spectroscopy (SEM-EDS) and activation analysis (INAA) reveal that the site’s bulk sediment is a variable average mix of Si (31 wt%), Ca (21 wt%), Al (18 wt%), and Fe (16 wt%), totaling ~86 wt% with the remaining 14 wt% comprised of small concentrations of Mn, S, Ti, Na, Mg, K, and P. 3.2. Spherule distribution The spherule distribution was determined at 11 locations to test whether there was a discernible pattern across the site. The results vary from zero spherules north and south of the lake/depression to a peak abundance of 9200 spherules within the southeastern part of the lake/depression (Figure 7). The isopachs show that spherule concentrations are highest within the lake border but fall off substantially with increasing distance from the lake. Figure 7: The spherules were plotted by depth across the site, revealing that the largest spherule peak concentrations occur in the lake at depths of up to ~6 m below the surface (Figure 8). Away from the rimmed lake/depression, spherule peaks vary in depth from 0.5 to 1.5 m below the surface. The depth of the deepest part of the lake is unknown; however, two wells, #4 and #6, drilled within ~6 m of the shoreline, encountered alternating layers of multi-colored clay (red through gray colors) down to a depth of ~6 m. At that depth, the well encountered water-saturated fine sand that flowed into the well bore, suggesting that the bore had passed below the bottom of the lake bed. This observation indicates the lake basin is steep-sided at ≥45°, atypical for any known area lake and consistent with the site’s lake being an anomalous feature. The rim of the lake/depression is also atypical for any known natural area lake within more than a 50-km radius. Figure 8: 3.3. Spherule morphologies Regarding spherule morphologies, although we refer to these objects as being spherulitic, they display a wide range of morphologies that include rounded, sub-rounded, ovate, oblate, elongated, teardrop-shaped, dumbbell-shaped, clustered, accretionary, hollow, and broken (Figures 9–12). An analysis of ~100 spherules reveals that their diameters range from ~1 to 65 μm with an average of 14 μm ( Supporting Information, Figure S2 [5]). Colors include clear, white, red, orange, purple, brown, and black; they can be opaque, translucent, or transparent. Surface textures vary widely, including smooth, dendritic, platy, soccer-ball-like, reticulated, and irregular. When sectioned or broken, the spherules most commonly display one or more large vesicles, but some are solid. The spherules sometimes display multi-phase morphologies (Figures 9 and 10) and may contain relict grains of quartz. The Fe-rich spherules are often depleted in oxygen, which is consistent with being wüstite and sometimes approaches the composition of native iron (Figure 10). Figure 9: Figure 10: Figure 11: Figure 12: Multi-layered glassy spherules were observed with similar morphologies to those previously reported for impact events [24, 25]. Many of them contain fragments of stoichiometric crystalline quartz surrounded by either melted quartz or Ca-rich aluminosilicate glass. Several contain an overgrowth of ettringite, a likely post-depositional mineral. Summary of Spherule Evidence Deposits contain >100 tonnes of spherules.
Estimated >100 billion spherules per kilogram .
Morphologies include rounded, sub-rounded, dumbbell, hollow , and clustered forms.
Elemental composition : Primarily Fe, Si, Al, and Ca.
Spatial distribution : Spherule concentrations are highest within the lake/depression, declining radially.
Depth profile: Deeper spherule peaks (up to 6 m) in the depression’s center, consistent with ejecta fallback.
4. Carbon spherules and glasslike carbon We observed multiple varieties of these carbon-rich particles, with colors including black, tan, and amber; transmissivity including opaque, translucent, and transparent; and shapes including spherical, oval, and broken. The abundance of carbon spherules in the Vee and Pond meltglass was ~500/kg with diameters ranging from ~100-800 μm, averaging ~600 μm (Figure 13). The glasslike carbon fragments were typically multiple millimeters in size with approximately the same abundance as the carbon spherules. The spherules and glasslike carbon are dominantly carbon (up to ~80 wt%) combined with small concentrations of Fe, Si, and O (Figure 14). Figure 13: Figure 14: Summary of Carbon-Rich Spherules and Glasslike Carbon Carbon spherules (opaque, translucent) and glasslike carbon fragments were found.
Enriched in platinum-group elements (PGE): Os, Ir, Pt.
Meltglass carbon content suggests lower peak formation temperatures than metallic or silicate melts.
5. Meltglass The meltglass elemental abundances vary but are mostly Fe-rich Ca-aluminosilicates. The meltglass from Trench #1 and the Vee deposit averages 9.64 wt% Al, 14.97 wt% Ca, 17.42 wt% Si, and 8.71 wt% Fe, with small concentrations of other elements ( Supporting Information, Tables S6 and S7 [5]). These irregular, broken objects display a wide range of textures, including smooth, rounded, folded, layered, and granular; most are highly vesicular (Figures 15–21). Dimensions range from sub-cm to a 31-cm-long meltglass rock weighing 16.33 kg. Colors vary from clear, white, red, orange, purple, brown, and black. They are primarily opaque but can appear translucent, as with numerous mm-sized, white, vesicular fragments that reached an abundance peak in Trench #1 at 12,794 ± 69 cal BP. Sand grain analyses for Trench #1 reveal that the abundance peak in meltglass corresponds to a nearly doubling of the fine silt component (<40 μm; Supporting Information, Figure S1 [5]). Figure 15: Figure 16: Figure 17: Figure 18: Figure 19: Figure 20: Figure 21: Abundances of meltglass fragments from nine cores, pits, and trenches are shown in Figure 6. Meltglass fragments, like many spherules, are typically composed of aluminosilicates (Al, Ca, Si, and O), although they commonly contain small amounts of other elements, including C, Fe, Zr, Ti, Ni, Co, Cr, and PGEs. Variations in concentrations of these elements affect the spherule morphology. Some fragments comprise >80% carbon and are more similar to glasslike carbon. Many meltglass fragments display high-temperature melted or boiled minerals, including magnetite (1590°C), NiCr-magnetite (1590°C), zircon with baddeleyite (1775 to ≥4400°C), kaolinite (1770°C), clinopyroxene (1150°C), and quartz (~1710 to ≥2230°C) (Table 1). Table 1: Summary of Meltglass Evidence Estimated >1 tonne of meltglass across the site.
Composed mainly of melted Fe- and Ca-aluminosilicates .
Textural evidence : Highly vesicular, layered, folded, tubular, and flowed structures.
Thermal evidence: Mineral inclusions suggest melting temperatures ≥1590–4400°C.
6. Low-oxygen metallic flakes These typically flake-like objects were analyzed using SEM/EDS. They ranged from ~38 to 818 μm in diameter, averaging ~200 μm with an abundance of ~100 particles/kg (Figure 22). They are typically composed of troilite, native iron (Fe), wüstite (FeO), reduced iron (Fe2+; i.e., oxygen-depleted), and chromian magnetite, all of which are common ET materials. The morphologies include flat, flake-like, or curved forms that sometimes appear folded and melted along the edges. Similar oxygen-depleted flakes have been reported to be cosmic dust particles, most likely cometary in origin [27–32]. Figure 22: Summary of Low-Oxygen Metallic Flakes Discovery of oxygen-depleted metallic flakes composed of native iron, troilite, and wüstite.
Morphologies (folded, melted edges) and chemistry consistent with extraterrestrial (ET) origin.
Rare on Earth, common in ET and impact-related materials.
7. High-temperature/high-pressure quartz 7.1. Shocked quartz grains with glass-filled fractures The similarities and differences between these four types of lamellae/fractures are summarized in Table 2. We investigated grains optically in 14 levels from 35-165 cm in Trench #1, in addition to samples from the Vee and Pond deposits. In Trench #1, shocked quartz was observed in the 12,794 cal BP YDB layer, not above or below. Shocked quartz candidate grains were selected, thin-sectioned, and further analyzed. Optical microscopy, SEM, and CL reveal evidence of extensive fractures and lamellae (12 in ~25,000 grains). We observed four types of microstructures in quartz in this study: classical planar deformation features (PDFs), classical planar fractures (PFs), and glass-filled non-planar fractures (herein termed “nPFs”) (Figures 23–28). We also observed but ignored non-shock deformation lamellae (DLs) [9, 34–38], which typically result from tectonic motion, but tectonism does not produce glass along the deformation lamellae. Table 2: Figure 23: Figure 24: Figure 25: Figure 26: Figure 27: Figure 28: Melted silica within quartz grains takes two forms: (i) diaplectic glass that forms when quartz undergoes shock-induced, solid-state transformation into an amorphous state by mechanical pressure rather than through high-temperature melting, and (ii) amorphous silica, also called lechatelierite, that melts at high temperatures at any pressure. Both forms of melted silica within quartz grains are called “glass” in this contribution. The presence of melted silica within shocked quartz is a characteristic that makes glass-filled shocked quartz a reliable indicator of cosmic impact events [9]. Indeed, French and Koeberl [38] emphasized the primary importance of melted silica in PDFs. They wrote, “Most striking, and most diagnostic of shock-wave pressures and impact conditions, are the unusual planar deformation features (PDFs) that have served reliably as an impact indicator for several decades.” Nearly all shocked quartz grains at Perkins display regions of amorphous silica, typically along lamellae and around the grain edges, and most show evidence of having been thermally warped after being shocked. One shocked quartz grain (Figure 23B) was observed within a matrix containing thousands of spherules adhering to its surface, confirming a close relationship between spherule formation and shocked quartz. 7.2.Thermally warped quartz grains When quartz grains undergo shock metamorphism that creates glass-filled PDFs, PFs, and fractures, they may undergo additional transformation when exposed to high temperatures from the impact fireball. This process can partially melt and warp the grains, leaving the lamellae still visible though non-planar, i.e., bent or curved (Figures 23–26). Some quartz grains display regions of isotropy when using crossed polars. Often, the grains are coated with a thin layer of isotropic material, consistent with noncrystalline silica (glass), and they commonly display non-planar and planar fractures filled with melted silica (Figure 26). SEM and SEM-EDS confirm that the red areas are not microcrystalline quartz (e.g., chalcedony), cryptocrystalline quartz (e.g., chert), or quartz overgrowth or cement (hydrated silica), and therefore, are melted silica. 7.3.Miller-Bravais indices confirm shocked quartz PDFs and PFs have been reported to display characteristic angles relative to the plane of the c-axis in quartz [9, 38–40], and these angles can be measured in polished thin-sectioned slides using a universal stage [41–43]. For the Perkins shocked quartz candidates, the measured angles were plotted in Figure 27 using the ANIE software program [42, 44]. The indices observed match the most common crystallographic orientations of planar microstructures in shocked quartz [9]. Summary of Shock Metamorphism in Quartz Presence of planar deformation features (PDFs) , planar fractures (PFs) , and non-planar glass-filled fractures (nPFs) .
Thermally warped quartz grains were observed, indicating post-shock heating .
Universal stage measurements match known shock-induced crystallographic orientations.
Shock evidence rate: ~1% of all quartz grains examined. 7.4.Microbrecciated quartz grains This term indicates that quartz grains have undergone brecciation, a process in which the grains are broken into angular to rounded fragments that are then bonded together by a mineral cement or melted material. This phenomenon is typically observed in high-stress environments, such as fault zones at a macro-scale, volcanic eruptions at a macro-scale, and impact craters at micro- and macro-scales. During an impact event, the cement in quartz grains commonly comprises isotropic, amorphous, and often vesicular silica. Sometimes, the entire quartz grain is melted, and as the molten quartz cools, it re-crystallizes into aggregates, with each globule representing a different crystallite [45, 46]. In some cases, the aggregates crystallize into spherical shapes; however, in other cases, they become sub-spherical, such as those observed in the Vee deposit (Figure 28). If the molten cement is quartz in a quartz grain, the grain is referred to as monomict quartz microbreccia; if the material is composed of other elements, the grain is referred to as polymict quartz microbreccia. Shocked quartz has previously been observed in monomict quartz breccia associated with Libyan Desert Glass [47] and in the YDB-aged site, Abu Hureyra, Syria [26]. Summary of Microbrecciated Quartz Microbrecciation within quartz grains bonded by amorphous silica.
Indicative of extreme stress and shock conditions associated with impacts.
8. Magnetic survey If the lake/depression is an impact crater, we reasoned that there may be associated magnetic total field anomalies due to potentially higher Fe concentrations produced by the impactor. We tested this hypothesis by conducting two magnetic surveys on the lake/depression across two wing-like features, similar to the pattern produced by the Tunguska airburst [48, 49], the Saarland structure airburst [50], and rays of secondary ejecta on the Moon and Mars [51]. We found positive magnetic anomalies associated with each wing (Figure 29), consistent with similar patterns related to other known impact structures that typically show magnetic intensity variations ranging from a few nT to several hundred nT [52]. These magnetic signatures may result from several impact-related processes: (i) The concentration of magnetic minerals through ballistic sorting during ejection and deposition; (ii) the thermal alteration of iron-bearing minerals during the impact event, potentially producing new magnetic phases through heating and rapid cooling; and (iii) possible contributions from impactor material, as suggested by the elevated Ni/Fe and Cr/Fe ratios that we found in the spherules and meltglass. Figure 29: Alternatively, there are multiple relict braid-plain stream channels across the area, so the lake/depression and wing-like features could be due to past fluvial action and only coincidentally associated with the spherule-and-meltglass deposits. However, other local lakes do not display wing-like features or raised rims, as is typical for impact craters. The magnetic data, combined with the morphological and geochemical evidence, provides additional indirect support for the lake/depression being an impact crater, although the magnetic evidence is inconclusive. Summary of Magnetic Survey Results Positive magnetic anomalies were detected along “wings” extending from the depression.
Morphology and magnetic signals are similar to known secondary ejecta patterns on the Moon, Mars, and Tunguska.
Alternative fluvial explanations are possible but questionable due to the absence of similar features nearby.
9. Summary of analytical results for all proxies The table below summarizes the results for the Perkins site (Table 3), and Figure 30 provides a visual flowchart to illustrate the complex, multidisciplinary data presented in this study. Table 3: Figure 30:
Discussion-interpretation The two large glass-and-spherule deposits at the Vee and Pond locations are at the surface and not in situ. However, small quantities of spherules and meltglass exhibiting the same morphologies and geochemistry occur in situ in the ~12,800-year layer of Trench #1. In this section, we consider several issues related to the Perkins site. How old are the spherule-and-meltglass deposits that had been relocated? Are these deposits related to or unrelated to the lake/depression? What processes produced the site’s spherules, meltglass, breccia, and shocked quartz?
10. Age of the Perkins materials The Bayesian-modeled age of the in situ spherule-and-meltglass deposit in Trench #1 is well constrained at 12,794 ± 69 cal BP, and the ages of some meltglass in Vee and Pond deposits overlap at 95% CI. We acquired six radiocarbon dates on carbon-rich material from the Vee and Pond deposits, but the dates did not definitively constrain the age of the deposits. Large carbon spherules from the Vee meltglass date to >28,000 cal BP, and six radiocarbon dates on large peat-like carbon-rich material extracted from the Vee and Pond deposits were all older than 42,000 cal BP. The reason for the old ages is unclear, but we speculate that when the Vee and Pond deposits were excavated by backhoe, older and younger carbon was mixed into the deposits, giving a wide range of dates. The 40Ar/39Ar dating helped eliminate any association with older craters elsewhere and confirmed a young deposit age. The significant uncertainties are mainly due to the presence of only small amounts of radiogenic argon (<5%) and the relatively long (6-hour) neutron irradiation that was conducted to test for ages in the Ma range. Incomplete resetting during impact-related thermal degassing could have led to an overestimate of the age, especially for very young (<100 ka) ages. Temperatures >1000°C were most likely attained during the impact, but these conditions only persisted for a few seconds, perhaps leading to incomplete argon diffusion. Non-radiogenic argon from air or other sources could also have been incorporated into the spherules as fluid inclusions or trapped on the surface of the spherules, leading to low radiogenic argon fractions (<5%). Shorter irradiation durations and additional chemical treatment planned for further work could provide better constraints on the formation age of these spherules. The Vee and Pond glass median ages are slightly younger than for Trench #1, possibly due to contamination by microscopic modern plant roots that penetrated the glass and were impossible to remove during processing for radiocarbon dating. An alternate hypothesis is that three meltglass-forming impact events occurred one or more centuries apart, separated by only a few 100 m. However, this hypothesis is statistically unlikely. Alternatively, several lines of evidence suggest that the deposits in Trench #1, the Vee, and the Pond are the same age. First, the layer of spherules, meltglass, and shocked quartz in Trench #1 is stratigraphically below an Early Archaic projectile point (age range: ~10,000-8,000 cal BP), thus indicating deposition of the material occurred before that (Figure 4, Supporting Information, Tables S1 and S3 [5]). Second, in Trench #1, spherules, meltglass, and shocked quartz grains were associated with sedimentary carbon with a modeled age of 12,794 ± 69 cal BP, consistent with the age range of the YDB impact event from 12,835-12,735 cal BP (Figure 4, Supporting Information, Tables S1 and S2 [5]). Third, five radiocarbon dates on the carbon-rich meltglass from the Vee and Pond deposits range from ~10,000 to 20,000 cal BP, and one sample of carbon spherules dated to ~29,000 cal BP (Figure 5, Supporting Information, Tables S1 and S3 [5]), suggesting that this material was deposited between ~30,000 and 10,000 cal BP. Two of the five ages also overlap with the age of the deposits in Trench #1 and the age range of the YDB impact event at 99% CI. This result suggests that the impact occurred at ~12,800 cal BP, but more research is needed to confirm the inferred age of the event. Although the dating on the deposits is uncertain, the spherules, meltglass, and shocked quartz grains in the well-dated Trench #1 are morphologically and geochemically similar to those in the Vee and Pond deposits, suggesting that all were deposited at the same time.
11. Origin of the spherules, meltglass, and shocked quartz 11.1 Anthropogenesis Iron furnaces and other industries often smelted or refined materials that would yield slag, a waste product of smelting metal ores. This process separates molten rock from the desired metal (e.g., during steel production and iron ore smelting). However, there are no known historical industries near Perkins, Louisiana; the closest facilities are not in Louisiana but in states like Pennsylvania, Ohio, and West Virginia. Smelting or glass manufacturing can produce slag glass, a distinctive type of pressed, layered material used in producing glassware and other items in the 19th century. However, there is no historical record of such industries near Perkins, Louisiana. The area has historical remnants of an old sawmill, a sugar mill, a tongue paper mill, and oil prospecting and sulfur mining. However, none of these historical industries could have produced glass slag. Lastly, we considered that the material might be fly ash from coal-fired power plants or railroad slag. The current landowners maintain that the previous owners dug up the deposits and did not haul in fly ash or other industrial material. Even so, we explored this contamination as a possibility. To do so, we compared the elemental abundances of Perkins spherules and meltglass to 273 fly ash samples from 31 studies representing 38 sites in Australia/New Zealand, Asia, Europe, Africa/Mideast, and North America (Figiure 31A, Supporting Information, Tables S8, and S9 [5]). We also compared the Si-rich Perkins melted material to Si-rich anthropogenic spherules [53]. The fly ash samples generally contain less Cr, Ni, Mn, S, Ti, Ca, Na, Mg, Fe, and Si than the Perkins samples. The Si-rich anthropogenic spherules generally contain less CaO, Al 2 O 3 , FeO, TiO 2 , and MgO than the Perkins material (Figure 31B, Supporting Information, Table S7 [5]). Figure 31: Thus, the Perkins glass-and-spherule deposits are chemically unlike any investigated fly ash and anthropogenic spherule samples. Furthermore, the meltglass and spherules are buried up to ~6.5 m below the surface, making anthropogenic contamination highly unlikely. Furthermore, anthropogenesis from industrial activities is doubtful because all seven radiocarbon dates are older than 10,000 years, ranging from ~10,000 to 30,000 cal BP. In addition, the deposits are geochemically nearly identical to the average of Perkins sediment samples, suggesting the deposits derive from local sediment. Although anthropogenic sources (e.g., power plants, factories, and railroads) are capable of producing spherules and meltglass, they cannot generate temperatures required to melt or boil quartz, zircon, and kaolinite (~1710° to 4400°C), nor are they capable of generating the pressures needed to produce shocked quartz. 11.2 Railroads and logging operations In the early 20th century, large parts of Louisiana underwent extensive logging activities, facilitated by installing temporary railroad tracks, which sometimes included industrial slag (i.e., meltglass) along the railway tracks. Although some of the trackways are still visible at the Perkins site, there is no evidence of any remaining slag alongside them. In addition, there is no known mechanism by which railway activities could boil zircon grains, boil quartz grains, or produce shocked quartz, all of which have been found embedded within meltglass at Perkins. 11.3 Natural gas blowout Local residents speculated that the lake/depression resulted from a natural gas blowout, which sometimes occurs among the oil fields of Louisiana. While this possibility might explain the crater, a blowout could not have generated the high pressures/temperatures required to produce the spherules, meltglass, shocked quartz, and other proxies. 11.4 Authigenesis There are various types of silica, including authigenic ones. Kletetschka et al. [1] studied shock-fractured quartz at Tunguska and reported that the quartz grains often showed isotropic silica coatings and glass-filled fractures similar to those found at the Perkins site. Their analyses confirmed that these features were melted silica rather than cryptocrystalline or microcrystalline quartz. As was previously reported for Tunguska, we often measured oxygen/silicon ratios of 53/47 wt% or less in partially melted grains, significantly lower than quartz overgrowth and quartz cement (i.e., hydrated silica at ~66/34 wt%), thus eliminating authigenesis. In addition, no known authigenic process can produce shocked quartz. 11.5 Tectonism Perkins is in a relatively inactive coastal fault zone, and although fault motion can create breccia, meltglass, deformation lamellae, and macro-scale melting [54, 55], tectonic processes cannot boil quartz and zircon or produce PDFs, glass-coated quartz, or silica-filled fractures in individual quartz grains [35, 56]. Anders et al. [57] noted that deformation lamellae are not fractures but dislocations. Unlike the shock features observed at Perkins, they are never glass-filled and are never shocked. 11.6 Volcanism Spherules, meltglass, and breccia can form from volcanism. However, it is widely accepted that volcanism cannot produce shock metamorphism [35, 56]. Furthermore, the nearest volcanoes are >1000 km away in Central America, which is far too distant to deposit tonnes of material at the Perkins site. 11.7 Lightning This process typically produces tubular fulgurites confined to small areas up to several meters in diameter [26, 58, 59], much smaller than the 300-m lake. One might reasonably expect lightning strikes to be frequent over time on any given part of the planet, and therefore, fulgurites should be ubiquitous. However, we found only one fulgurite in 32 cores and five trenches, and it was not associated with an abundance peak in spherules and meltglass. Furthermore, meltglass was not ubiquitous outside of the inferred impact layer. Although we searched extensively for meltglass fragments in Trench #1, we observed none in three samples, inconsistent with formation by lightning strikes, which are expected to be ubiquitous. There is a potential connection between lightning and impacts. Lightning strikes are inferred to be numerous in impact-related dust clouds, based on observing massive increases in lightning in volcanic plumes. For example, after the submarine eruption of Hunga Tonga, the incidence of volcanic lightning was so high that it constituted up to 80% of all global lightning activity at that time [60]. However, lightning generated during an airburst differs significantly from typical thunderstorm or volcanic lightning in several key aspects. In airburst-generated lightning, charge separation primarily occurs through debris particles rather than water droplets or ice crystals in thunderstorms [61]. The dust/debris cloud produced by the airburst would have distinct electrical conductivity properties due to its unique composition, which includes vaporized impactor material and superheated atmospheric gases [62], combined with extreme temperatures exceeding several thousand Kelvin. While lightning related to an impact event may have contributed to forming some features in our samples, the mechanism would have differed fundamentally from typical terrestrial lightning processes, potentially producing unique signatures in the geological record. Thus, regular lightning as a potential formation mechanism can be excluded. 11.8 Older airburst/impact event We considered whether the Perkins deposits might be reworked from those older impact events. There are five known impact craters in the southeastern United States within 800 km of Perkins: a ~58-Ma crater 300 km away in Texas, a ~84-Ma crater ~700 km away in Alabama, and three ~100-300-Ma craters >800 km away in Tennessee. In addition, Chicxulub, the 65-Ma K-Pg crater, is >1000 kilometers away in Yucatán, Mexico, and the 35-Ma Chesapeake Bay crater is >1700 km away in Virginia. These craters are too distant to have ejected a 16.33 kg meltglass fragment that landed at the Perkins site. In addition, it is unlikely that spherule deposits hundreds of km away from a crater would have been concentrated at >100 billion spherules/kg at Perkins. Additionally, the 40Ar/39Ar data of the spherules constrain the age to <2 Ma, significantly younger than any known major impacts in the region. Thus, the most plausible scenario is that a local impact event deposited the material, most likely within the last 30,000-10,000 years. 11.9 Airburst/impact event A comparison of Perkins melted spherules and meltglass to impact-related microtektites and impact spherules reveals they are compositionally similar (Figure 31B). Thus, the presence of these materials, along with boiled quartz grains, boiled zircon, and shocked quartz, supports a high-temperature, high-pressure cosmic impact event at the Perkins site. However, although these materials were found in situ in Trench #1, that was not the case for the Vee and Pond deposits, so the question remains regarding their original location. Before 1999, one pit was excavated at the Lake deposit site, displaying numerous fragments of meltglass and spherule-rich clusters (site “W” in Figures 7 and 8). We searched but found no other primary deposits away from the lake/depression. Therefore, the most likely source appears to be the Lake deposit site on the NW edge of the lake/depression (Figure 3). The glass-and-spherule deposits’ lateral and vertical distribution patterns are consistent with the lake being ground zero for an airburst/impact, but this issue of their origin remains unclear. To further explore the cosmic impact hypothesis, we performed neutron activation analyses (NAA) on the sediment from Trench #1 in search of elemental abundances indicative of potential extraterrestrial (ET) material. A useful test is to compare the ratios of Ni and Cr to Fe because Ni is typically enriched in ET material ~330 times over crustal abundance [63] and Cr is enriched 38 times [63], whereas Fe is enriched much less at only ~5 times [63]. The results reveal peaks in Ni/Fe and Cr/Fe ratios that are higher than crustal averages but lower than for meteoritic material (Figure 32A), consistent with the dominance of terrestrial material mixed with small amounts of ET material, typically reported at <2 wt% [64–66]. Figure 32: In addition, we calculated the same ratios from SEM-EDS elemental abundances for selected metal-rich areas of Vee meltglass fragments. Ni/Fe ratios in 5 of 10 spherules and 3 of 22 meltglass fragments were higher than the ET average ratios [63], while the others displayed terrestrial ratios [63]. The Cr/Fe ratios for 4 of 10 spherules and 3 of 22 meltglass fragments were higher than the ET average ratios (Figure 31B). Thus, these results suggest that terrestrial melted sediment became mixed with minor amounts of ET material. We also used laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) to measure elemental abundances for 7 Perkins spherules. We then compared these elemental abundances with those of known ET material, which is typically highly enriched compared with crustal abundances, e.g., Ir = 24,000×; Pt = 1500×; Cu = 6×; Fe = 5×; Zn = 4×; and Ni = 330×(Figure 33A) [63]. The results reveal that spherules from the Vee deposit likely comprise mostly melted terrestrial sediments incorporated with small percentages of extraterrestrial material, typically <2 wt% [64–66]. For some PGEs in Perkins spherules, the enrichments are very high: the average Ir concentration is ~20,000× higher than the crustal average, and Pt is ~600× higher. Figure 33: In addition, we compared Ni/Co and Ni/Cr ratios of Perkins spherules and meltglass to material from known impact events, the 145-Ma Morokweng crater and the 65-Ma Chicxulub crater. Ni, Co, and Cr are helpful for comparison because they are typically elevated over crustal abundances in ET material (Ni = 330×; Co = 33×; and Cr = 38×). The Ni/Co and Ni/Cr ratios for Perkins spherules overlap those from an impact spherule layer dating to 2.63-2.49 Ga, and the Perkins meltglass also overlaps the ratios from two known impacts (Figure 33B, 33C). The Perkins compositions are similar to those found at multiple YDB sites across 5 continents [26, 58, 70–80]. Most importantly, the Perkins glass-and-spherule deposits are morphologically and compositionally similar to 28 spherule layers identified worldwide [81, 82]. Although only seven of those layers have been positively linked to known impact craters, it is widely accepted that those 28 spherule layers represent evidence of cosmic impact events, even in the absence of known craters [81, 82]. Many of the varieties of spherules, meltglass, metallic flakes, and quartz grains observed in this study display evidence of high-temperature melting or boiling. Their equilibrium melting to boiling points range from ~1050 to 4400°C, with flux-mediated, non-equilibrium melting typically occurring several hundred degrees lower (Table 1). The evidence is consistent with an extremely high-energy thermal event.
12. The lake/depression as an impact crater There is no conclusive evidence that the Perkins lake/depression is a crater, partially because modern excavations have modified parts of the Perkins lake/depression. However, the collective results suggest that this structure is a possible crater because known impact craters typically display raised rims and are associated with large quantities of high-temperature melted spherules, meltglass, and high-pressure shocked quartz grains [38]. Even so, identification of an impact crater is not required for acceptance as an impact event. Twenty-eight known spherule layers of different ages across the planet are widely accepted as representing impact events, though 21 are not associated with any known craters [81, 82]. Two extensive fields of meltglass, the Libyan Desert glass field and the Australasian tektite field, the world’s largest, are widely accepted as having been produced by impact events, even though no craters have been identified [3, 4]. Further research would be beneficial for testing the hypothesis that the lake/depression resulted from a cosmic impact. These studies could include deep-coring the lake/depression in search of a buried bed of spherules and meltglass or conducting geophysical surveys (e.g., gravity, seismic, or airborne magnetic). In addition, we propose exploring smaller depressions near the primary lake/depression to determine if they are also of potential impact origin. Several previous investigations have concluded that during airbursts at all altitudes, the incoming comet or asteroid is vaporized so that no fragments reach Earth’s surface [83, 84]. Thus, the question arises whether the airburst proposed here is possible or plausible. While most incoming bolide material vaporizes during an airburst, it is indisputable that airburst fragments commonly reach Earth’s surface. For example, fragments of the Sikhote-Alin iron meteorite struck Siberia in 1947, distributing ~8500 pieces totaling more than 23,000 kg across 1.6 square km. This event produced more than 100 small impact craters ranging from 0.5 to 26 m diameter [85]. Argentina’s Campo del Cielo meteorite field contains numerous meteorites that produced >100 shallow craters up to 26.5 m in diameter and 6 m in depth. Shock pressures were sufficiently energetic for the Sikhote-Alin and Campo del Cielo airbursts to produce multiple small, shallow, shock-generated craters in unconsolidated surficial sediments. Although the airburst over Chelyabinsk, Russia, in 2013 was a much smaller event than the ones discussed above, fragments still reached Earth’s surface. Even though the blast occurred at a height of ~29.7 km with an energy of ~500 kt [83, 86–89], the exploding bolide ejected numerous meteorite fragments, two of which weighed 64.7 and 540 kg [90], the largest of which produced a 9-m diameter hole in the frozen surface of a lake. Across Western Europe (2.2 million km2), exploding fragments from eight proposed near-surface airbursts [2] produced shallow craters associated with spherules, meltglass, shocked quartz, and microbrecciated quartz. Co-author K. E. and colleagues investigated most of them across three countries: the Czech Republic [91, 92], Germany [93–109], and Finland [110]. They have been dated to within the last ~11,700 years, averaging one touch-down airburst in Western Europe approximately every 1500 years during the Holocene, making them a geologically common phenomenon. Thus, there is extensive evidence that airburst fragments can strike Earth’s surface with significant force, making surface impacts during airbursts the norm, regardless of whether the airbursts occur at high or low altitudes. If the shock front and thermal pulse are sufficiently intense during an airburst, these conditions can create meltglass, spherules, shocked quartz, and all other evidence observed at the Perkins site.
13. Melted and shocked quartz during an airburst Airbursts are generally considered to be low-shock events, so the question arises as to whether the required conditions could have occurred to produce shocked quartz in a Perkins airburst. In conducting laboratory experiments to investigate shock quartz, Christie et al. [111–113] showed that melted silica could form in fractured quartz cylinders where the confining pressure was 15 kb (1.5 GPa), inferred to result from frictional melting during brittle fracturing. The evidence indicates pressures equal to or greater than this value produced the materials at Perkins. In addition to high pressures, Gratz et al. [114] found that thermally induced shock metamorphism can occur when high temperatures (≥1710°C) create fractures in quartz that are injected with melted silica. Ernstson performed similar laboratory experiments [115, 116] and observed spallation, during which quartz fractures into multiple fragments through interaction with a compression wave. The results of Christie et al. and Ernstson suggest a plausible formation mechanism for the Perkins shocked quartz and microbrecciated quartz. Relatively low-energy airburst/impact shockwave pulses could have created compression fractures in quartz grains, and spallation created tensile fractures. Lastly, after the grains fractured, they filled with melted silica either from the grain surface or from melted silica and vapor carried by the high-pressure, high-velocity airburst cloud. Melted silica within breccia is the primary characteristic distinguishing impact breccia from other geologically deformed breccia [117]. The material at Perkins can be identified as non-authigenic melted silica due to several factors. The glass is determined to be amorphous using crossed polars, thus confirming that it is noncrystalline, and this characteristic eliminates authigenic quartz overgrowth and quartz cement, which are crystalline. Next, SEM imagery detected no micro-spheres in the glass, thus eliminating chert, agate, chalcedony, and other microcrystalline or cryptocrystalline varieties of quartz. Lastly, SEM-EDS indicates that the oxygen/silicon ratio of the glass is ≤53/47 wt%, the stoichiometric ratio of quartz, thus indicating that the glass is not hydrated silica (hyalite at ~66/34 wt% ratio). Therefore, high-temperature melted silica is the only remaining possibility for the Perkins glass. The shocked quartz and vesicular microbrecciated quartz were observed at all three Perkins locations, often fused to spherules and meltglass, suggesting that they formed simultaneously. As discussed above under potential origins, the most parsimonious explanation is that it resulted from a cosmic impact event.
14. Hydrocode model of a touch-down airburst Here, we explore the possibility that the evidence at Perkins resulted from a touch-down airburst, in which the near-surface detonation of an asteroid or comet produced high-velocity fragments that struck the ground with sufficient velocity to create a shallow crater. In addition, the event generated the high-temperature, high-pressure conditions necessary to produce spherules, meltglass, shocked quartz, and the other observed materials. Multiple studies [2, 12, 118, 119] have produced hydrocode models or reported evidence that airbursts/impacts can produce high-velocity fragments that strike the ground with sufficient velocity to produce glass-filled shocked quartz. In addition, a hydrocode model produced by Boslough and Crawford [62, 120, 121] of a 108-Mt airburst demonstrated that such events can excavate unconsolidated sediment, melt surficial sediments at ~5800 K, and create a crater (Figure 34). Boslough and Crawford [62, 121] describe the theoretical process by which airbursts can produce meltglass. This process is essential, so we quote their conclusion in its entirety. For some airbursts, “the hot jet of vaporized projectile (the descending ‘fireball’) makes contact with the Earth’s surface, where it expands radially. During the time of radial expansion, the fireball can maintain temperatures well above the melting temperature of silicate minerals, and its radial velocity can exceed the sound speed in air. We suggest that the surface materials can ablate by radiative/convective melting under these conditions, and then quench rapidly to form glass after the fireball cools and recedes.” Figure 34: For this study, we also performed computer modeling to investigate whether a cosmic airburst/impact could produce spherules, meltglass, and shocked quartz. Initial conditions were first modeled using the online Earth Impact Effects Program ( Supporting Information, Tables S10 and S11 [5]) [122, 123] as follows: a 350-m-wide comet with a density of 250 kg/m3 (equivalent to the density of Shoemaker-Levy 9 [124, 125]) entered Earth’s atmosphere at 45°, traveling at 30 km/s. The impactor initially broke up into multiple fragments as an airburst at an altitude of 108 km, and these fragments struck Earth’s surface as a low-density cloud traveling at 0.975 km/s with an impact energy of 0.64 Mt (Figures 35 and 36). The largest cloud of impactor fragments was modeled as ~100 m in diameter, although it should be noted that this was not a solid object but rather a high-velocity dispersed cloud of debris. This cloud produced a modeled elliptical crater 363 m long and 77 m wide, corresponding well with the 300-m span of the seasonal lake/depression at Perkins. The floor of the modeled crater was underlain by a lens of meltglass tens of meters below the surface, too deep for us to detect with our 6-m test wells. Figure 35: Figure 36: This hydrocode model confirms that near-surface contact airbursts can produce fragments that strike the Earth with sufficient velocity to produce meltglass, spherules, and shocked quartz. The Earth Impact Effects Program predicts that events similar to the proposed Perkins airburst recur every 19,000 years. The parameters used for modeling with Autodyn are shown in Supporting Information, Tables S10 and S11 [5]. This model is non-unique; we created dozens of models with varying parameters under which low-density, dispersed comet fragments can produce shallow craters similar to the one modeled here. The numerous variations indicate that these events are common, far more so than typical crater-forming impact events, and the modeled cratering shows that they can potentially cause considerable surface damage [2, 10]. We suggest the following potential airburst/impact scenario for the Perkins site. A large low-density comet entered the atmosphere at a low angle and initially broke up catastrophically at a high altitude. Multiple fragments descended until one or more impacted into or exploded just above the Earth’s surface as a “touch-down” or “type-2 contact airburst” [126, 127]. The impact or airburst had sufficient energy to create a shallow-rimmed crater or multiple smaller craters that overlapped to become the modern lake/depression (Figure 37). Smaller fragments struck the ground in a shotgun-like blast, forming multiple ephemeral craters across the site. The airburst/impact ejected surficial sediments that formed a raised rim around the crater, an ejecta blanket to the west, and ‘wings’ of impact-related debris. Figure 37:
15. Comparison of the Perkins site to other YDB sites and the K-Pg Several tables below compare the Perkins site to other YDB sites (Table 4) and compare YDB sites to the KPg impact event (Table 5). Table 4: Table 5: