Introduction

Centimeter-sized tektites and microscopic tektites (i.e., microtektites; usually <1 mm) are impact glasses formed by hypervelocity impacts on Earth1,2. Quenched from impact-generated melt and/or vapor, microtektites share similar geochemical compositions with tektites3,4, and their protolith is consistent with being continental surface deposits1,5. Tektites and microtektites formed by the same impact event have wide ranges in color and size. They are high-velocity, distal ejecta distributed in strewn field extending for hundreds to thousands of kilometers6. So far, five major Cenozoic impact strewn fields are recognized on Earth, including the North America (NASF), Central Europe (CESF), Ivory Coast (ICSF), Australasian (AASF), and Belize6,7. In addition, three other possible tektite strewn fields were recently reported, i.e., Ananguites, Geraisites, and Uruguaites strewn fields8,9,10. The source craters of NASF, CESF, and ICSF have been confirmed, and they occur within or adjacent to the uprange portions of the strewn fields6. The spatial correlation between strewn fields and their source craters is generally consistent with that revealed by numerical simulations of oblique impacts11,12. Therefore, the accurate definition of the geographic boundaries of strewn fields is critical for guiding the search for the source crater.

Among the five Cenozoic strewn fields, AASF is the youngest (788.1 ± 3 ka)13 and largest (covering >15% of the Earth’s surface)14 one. Anchored by localities where tektites and microtektites were discovered, the geographic boundary of AASF has a lobate shape, with the long axis extending north–northwest to south–southeast and the two short wings being nearly symmetric about the long axis (Fig. 1). Therefore, it is generally agreed that the impact event that formed AASF was an oblique impact with an impactor trajectory from northwest to southeast15. While the overall spatial density of AASF microtektites decreases at larger distances from the Indochina Peninsula and adjacent waters16, spatial densities of microtektites are heterogeneous at small areas of the strew field17,18, and they do not occur beyond the sketched geographic boundary (Fig. 1). On the basis of the spatial distribution and composition of the tektites and microtektites in AASF, the source crater was estimated to be 10 s to ~100 km in diameter19 and possibly located in Indochina20,21. However, despite decades of research, the precise location of its source crater remains unconfirmed22,23,24,25.

Fig. 1: The geographic boundary of Australasian strewn field (AASF).
Fig. 1: The geographic boundary of Australasian strewn field (AASF).
Full size image

The extension is determined by the locations of tektites, minitektites and microtektites found on land and/or in deep-sea sediments (DSS)16,18,35. The locations of Beijing Plain and Baise Basin are marked. The base map is sourced from Google Satellite data.

China is part of the uprange portion of AASF26,27. The northern boundary of AASF was roughly sketched based on three anchor points where Australasian tektites and microtektites were discovered (Fig. 1), i.e., an ocean drill core to the south of Japan that contained Australasian microtektites28, Baise, Guangxi Province that contained Australasian tektites29, and ocean floor sediments in the Indian Ocean that contained Australasian microtektites and minitektites18. Compared to the large spatial density of tektites and microtektites in the downrange areas, additional occurrences of Australasian tektites and microtektites may be discovered to the north of the current northern boundary, thus the geographic boundary and the potential impact site of this strewn field may be refined.

Australasian tektites were claimed to exist in Tibet30,31,32. Although the reported compositions of these samples are identical to those of AASF tektites, they were purchased samples and stratigraphic occurrences were missing30,31,32. Field investigations have yet to confirm the existence of Australasian tektites in Tibet. Australasian microtektites were reported in the Chinese Loess Plateau33,34, but their compositions were different from those of Australasian tektites and microtektites6. A recent systematic search across the Chinese Loess Plateau investigated 19 loess sections with a total area of over 950 cm2, yielding negative findings35. On the other hand, in the middle and lower Pleistocene sediments from two boreholes in the Beijing Plain (Fig. 1), over 100 microscopic glassy spherules that appear similar to microtektites were reported36,37,38. Although their compositions are dramatically different from those of known tektites and microtektites (e.g., CaO=27.11 wt%; K2O = 2.94 wt%; TiO2 = 0.08 wt%), they were interpreted to be microtektites formed by an unknown impact event rather than those associated with AASF, NASF or ICSF36,38. However, the exact identity of these spherules was not recognized.

It is theoretically plausible that Australasian microtektites may be found in the Beijing Plain, as this plain has preserved continuous thick continental sediments since the Quaternary39, and the northern boundary of AASF is roughly constrained by only three anchor points (Fig. 1). In principle, oblique impacts on planetary surfaces, including the Earth, preferentially deliver impact ejecta towards the downrange areas40, but highly oblique impacts are also capable of launching minor ejecta towards the uprange areas41. This phenomena is particularly profound on Earth, considering that ejecta landing at over 400 km from the impact site may temporarily escape Earth’s atmosphere, and the atmospheric reentry may occur dozens of minutes and/or hours later, so that the Earth rotation permits a global reentry of tektites and microtektites42,43,44. The spatial extension of AASF microtektites is over 11,000 km45, suggesting that AASF might have a global occurrence. Therefore, if AASF microtektites were confirmed in the Beijing Plain, the refined geographic boundary of the strewn field and the spatial density of microtektites would be indicative of the location of the source impact.

In this study, we analyzed core samples from three boreholes in the Beijing Plain. Referring to the magnetostratigraphy and sedimentation rate of the core samples, we examined continuous core samples with sedimentation ages bracketing the formation age of AASF.

Results

Samples analyzed in this study were obtained from three boreholes in the Beijing Plain, i.e., PGZ01, PGZ05, and NBT 1, which were drilled in the southern and southeastern parts of the Beijing Plain in 2014, 2015, and 2018, respectively46,47,48 (Fig. 2). The depths of the three cores are 420 m, 255 m, and 400 m, respectively, and they all record a continuous sedimentation history of fluvial, lacustrine and alluvial deposits from the early Pliocene to the Holocene46,47,48. Stored at the Beijing Geological Survey Institute. The core diameters of PGZ01, PGZ05, and NBT 1 are 110 mm, 90 mm, and 108 mm, respectively46,47,48. The total cross-sectional area of the three boreholes is ~250 cm2, which is substantially larger than that of individual ocean drill cores, in which Australasian microtektites were discovered49,50. This total area is also comparable to that of a box core (225 cm2) and a cylindrical tube (19.6 cm2) sank in the Indian Ocean, from which microtektites, minitektites, and a tektite fragment were recovered18. Therefore, the three boreholes in the Beijing Plain are large enough in the cross section to search for potential Australasian microtektites.

Fig. 2: The locations of boreholes PGZ01, PGZ05 and NBT 1 in the Beijing Plain.
Fig. 2: The locations of boreholes PGZ01, PGZ05 and NBT 1 in the Beijing Plain.
Full size image

The base maps are sourced from Resource and Environmental Science Data Platform.

For each of the three core sections, we first determined the depth ranges of samples that bracketed the possible deposition age of Australasian microtektites. The AASF impact occurred ~15 kyrs13 prior to the Matuyama-Brunhes geomagnetic reversal boundary (MBB; ~773 ka)51,52. The stratigraphic positions of the MBB and the top of the Jaramillo subchron in the three core sections have been determined by magnetostratigraphic investigations46,47,48, so that the average accumulation rate of sediments between the MBB and the top of the Jaramillo subchron can be estimated for each core section. While the deposition ages of the loose Quaternary core samples cannot be efficiently determined via radiometric dating or luminescence dating, the local sedimentation rate can be established with reasonable precision, because samples deposited between the MBB and the top of the Jaramillo subchron are mainly clays and silty sands that indicate relatively stable deposition environment46,47. Therefore, the stratigraphic position of the AASF impact event was deduced based on the average accumulation rate and the ~15 kyrs age difference with respect to the MBB, which corresponds to depths of 79.45 m, 76.66 m and 71.82 m in boreholes PGZ01, PGZ05, and NBT 1, respectively (red line; Fig. 3). On the other hand, Australasian microtektites in deep-sea cores typically concentrate vertically within an interval of 30–50 cm due to disturbances by ocean flows and/or bioturbation28, and the magnetostratigraphic of the MBB in Chinese loess may be shifted by 10 s cm to several meters due to lock-in effect and/or remagnetization53. To bracket the possible vertical occurrence range of Australasian microtektites, this study focused on samples from depths of 79–80 m in borehole PGZ01, 76–77 m in borehole PGZ05, and 71–72 m in borehole NBT 1 (shaded area; Fig. 3). Within the selected depth ranges, samples of PGZ01 are dominated by reddish-brown and brown fine-grained clays with horizontal bedding46, those of borehole PGZ05 consist primarily of brown clay and yellow-brown silty sands deposited in a lacustrine-swamp environment47, and those of borehole NBT 1 comprise gray and gray-black sands48.

Fig. 3: Palaeomagnetic polarity column, magnetic inclination (I0), and depths of the three studied boreholes46,47,48.
Fig. 3: Palaeomagnetic polarity column, magnetic inclination (I0), and depths of the three studied boreholes46–48.
Full size image

The blue solid line denotes the position of the Matuyama-Brunhes geomagnetic reversal boundary, dated to 773 ka51,52. The red solid line denotes the theoretical position of the Australasian microtektite layer in the boreholes, which was deposited at ~788 ka13. Gray shaded areas indicate the depth intervals of samples analyzed in this study, which encompass the theoretical occurrence range of the potential Australasian microtektite layer.

After sample pretreatment (see Methods), we recognized three spherules that appear similar to microtektites in terms of shape, color and luster (Fig. 4). With dimensions of ~160–210 μm and a glassy and/or metallic luster (Fig. 4a, c, e), the spherules generally exhibit round and/or elliptical shapes. The particle shown in Fig. 4a, b (i.e., PGZ01-1) is colorless, transparent and with complete extinction. The other two spherules (i.e., PGZ01-2 and PGZ01-3) are black and opaque (Fig. 4c–f).

Fig. 4: Optical images of the three spherules that appear similar to microtektites.
Fig. 4: Optical images of the three spherules that appear similar to microtektites.
Full size image

a, c, e Images obtained under reflected light. b, d, f Polished sections of the spherules under transmitted light.

SEM observations reveal that the surface of PGZ01-1 is fully occupied by complex microstructures (Fig. 5a, b). PGZ01-2 and PGZ01-3 have much fewer surface microstructures and smoother textures (Fig. 5c, e). Interlocking dendrites are visible on the surface of PGZ01-2 (Fig. 5c), in which banded structures are present on the surface (Fig. 5d). PGZ01-3 is a broken spherule, and a large void about 120 by 60 μm in dimension is visible in this particle, within which fine-grained materials are visible (Fig. 5e). The surface of PGZ01-3 is dominated by quasi-parallel linear structures (Fig. 5f).

Fig. 5: Microstructures on surfaces of the three spherules, obtained via secondary electron imaging using a scanning electron microscope.
Fig. 5: Microstructures on surfaces of the three spherules, obtained via secondary electron imaging using a scanning electron microscope.
Full size image

a, c, e Overall surface morphology of the spherules. b, d, f Enlarged view of the surface microstructures.

PGZ01-1 and PGZ01-2 exhibit homogeneous compositions in the interior as seen in backscatter electron images of the polished sections (Fig. 6a–c), and the interior of PGZ01-2 is composed of interlocking dendrites (Fig. 6d). PGZ01-3 contains phases with different compositions (Fig. 6e), and the darkish phase exhibits dendritic textures (Fig. 6f).

Fig. 6: Backscatter electron images of polished sections of the three spherules.
Fig. 6: Backscatter electron images of polished sections of the three spherules.
Full size image

a, c, e Polished sections of the individual spherules. b, d, f Enlarged view of the polished sections.

PGZ01-1 is mainly composed of CaO (54.5–55.8 wt%), P2O5 (41.4–42.3 wt%) and F (2.8–3.3 wt%; Table 1). PGZ01-2 and PGZ01-3 are dominated by FeO (89.9–95.8 wt%). Minor compositional difference exists between the dark (FeO = ~91.5–93.1 wt%) and light (FeO = ~94.0–95.8 wt%) phases in PGZ01-3 (Fig. 6f).

Table 1 Major element contents (wt%) of the three spherules obtained using an electron microprobe analyzer
Full size table

Discussion

The major element compositions of the spherules are fundamentally different from those of known Australasian microtektites (Fig. 7), such as the extremely low SiO2 content («1 wt%; Table 1). The chemical composition (CaO=54.5–55.8 wt%, P2O5 = 41.4–42.3 wt%, and F = 2.8–3.3 wt%) of PGZ01-1 and its complete extinction under cross-polarizer are consistent with those of fluorapatite54,55,56. The polygonal surface and structures and compositions (FeO = ~89.9–95.8 wt%; Table 1) of PGZ01-2 and PGZ01-3 are consistent with those of the iron-rich cosmic spherules (I-type)57,58. Therefore, the three spherules are not Australasian microtektites.

Fig. 7
Fig. 7
Full size image

Comparison of major oxide contents between the three spherules and known Australasian microtektites4,14,16,28,61,62,63,64,65,66,67.

The non-detection of Australasian microtektites in the three boreholes from the Beijing Plain concurs with the similar discovery in the Chinese Loess Plateau35. There is no evidence to suggest that the northern boundary of AASF extends further north. This result supports the asymmetric distribution of tektites and microtektites formed during oblique impacts15, further indicating that the AASF impact event did not form a global distribution of microtektites. Therefore, compared to larger impact events, such as Chicxulub that formed a near-global distribution of microtektites59, the AASF impact event was less energetic60.

Methods

Sample pretreatment

For the selected section of each core, we performed continuous sample division, yielding eighty-eight subsamples in total. Within the depths of 79.25–79.75 m of PGZ01, 76.45–76.9 m of PGZ05, and 71.55–72 m of NBT 1, the sampling interval was 2.5 cm. At larger depths (i.e., 79–79.25 m and 79.75–80 m of PGZ01, 76–76.45 m and 76.9–77 m of PGZ05, and 71–71.55 m of NBT 1), the sampling interval was 5 cm. Each sub-sample was first milled into fine particles less than ~1 mm using mortars.

Sample pretreatment followed the recommendation used for ocean drill cores6, which was also employed to search for possible Australasian microtektites in the Chinese Loess Plateau35. Pretreatment of PGZ01 samples was conducted at the Planetary Environmental and Astrobiological Research Laboratory, Sun Yat-Sen University. Each subsample was first immersed in water containing acetic acid or ~5 vol.% diluted hydrochloric acid to digest carbonate component. The subsamples were then sieved and washed through a 300-mesh sieve (~48 μm). Hydrogen peroxide was subsequently added to remove organic matter. Finally, after a second wash, the subsamples were dried in a glove box at ambient temperature35. Pretreatment of PGZ05 and NBT 1 samples was performed at Micropreparation Laboratory, University of Pisa. After milling, each subsample was washed sequentially through sieves with mesh sizes of 48 μm, 125 μm and 300 μm16. Afterward, all washed subsamples (>~48 μm) were dried on a low-temperature heating table at ~40 °C.

Optical and secondary electron imaging

We examined the pretreated samples under a binocular. Spherical particles with rounded shapes and vitreous luster were selected. The optical properties of the spherules, e.g., color, transparency, luster and extinction, were characterized using both binocular and polarizing microscopes. Their detailed surface morphology was characterized using a Zeiss Sigma 300 scanning electron microscope equipped with a backscatter electron detector, operating at a voltage of 15 kV and a work distance of ~9 mm.

Backscatter electron and electron microprobe analysis

We prepared epoxy mounts for the spherules to investigate their internal structures and major element compositions using a JEOL JXA-iSP100 electron microprobe analyzer. The working conditions included an accelerating voltage of 15 kV, a beam current of 20 nA, and a defocused beam diameter of 5–10 μm. Peak counting times were 10 s for Na and K, and 20 s for other elements. The detection limits were below 0.05 wt% for F, FeO, Y2O3, Ce2O3, below 0.03 wt% for Na2O, SiO2, P2O5, CaO, TiO2, Cr2O3, MnO, CoO, NiO, below 0.02 wt% for MgO, Al2O3, SO3, K2O, below 0.01 wt % for Cl. All data were corrected using the ZAF (atomic number, absorption, and fluorescence) procedure.