Interior textures, chemical compositions, and noble gas signatures of Antarctic cosmic spherules: Possible sources of spherules with long exposure ages


  • Takahito OSAWA,

    Corresponding author
    1. Neutron Imaging and Activation Analysis Group, Quantum Beam Science Directorate, Japan Atomic Energy Agency (JAEA), Shirakata Shirane 2-4, Tokai-mura, Naka-Gun, Ibaraki 319-1195, Japan
    Search for more papers by this author
  • Yukio YAMAMOTO,

    1. Environmental Science Center, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author
  • Takaaki NOGUCHI,

    1. College of Science, Ibaraki University, 2-1-1 Bunkyo, Mito 310-8512, Japan
    Search for more papers by this author
  • Akari IOSE,

    1. College of Science, Ibaraki University, 2-1-1 Bunkyo, Mito 310-8512, Japan
    Search for more papers by this author
    • Present address: AstraZeneca K.K., Umeda Sky Building Tower East 1-88, 1-Chome, Ohyodo-naka, Kita-ku, Osaka 531-0076, Japan

  • Keisuke NAGAO

    1. Geochemical Research Center, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author

Corresponding author. E-mail:


Abstract– The interior texture and chemical and noble gas composition of 99 cosmic spherules collected from the meteorite ice field around the Yamato Mountains in Antarctica were investigated. Their textures were used to classify the spherules into six different types reflecting the degree of heating: 13 were cryptocrystalline, 40 were barred olivine, 3 were porphyritic A, 24 were porphyritic B, 9 were porphyritic C, and 10 were partially melted spherules. While a correlation exists between the type of spherule and its noble gas content, there is no significant correlation between its chemical composition and noble gas content. Fifteen of the spherules still had detectable amounts of extraterrestrial He, and the majority of them had 3He/4He ratios that were close to that of solar wind (SW). The Ne isotopic composition of 28 of the spherules clustered between implantation-fractionated SW and air. Extraterrestrial Ar, confirmed to be present because it had a 40Ar/36Ar ratio lower than that of terrestrial atmosphere, was found in 35 of the spherules. An enigmatic spherule, labeled M240410, had an extremely high concentration of cosmogenic nuclides. Assuming 4π exposure to galactic and solar cosmic rays as a micrometeoroid and no exposure on the parent body, the cosmic-ray exposure (CRE) age of 393 Myr could be computed using cosmogenic 21Ne. Under these model assumptions, the inferred age suggests that the particle might have been an Edgeworth-Kuiper Belt object. Alternatively, if exposure near the surface of its parent body was dominant, the CRE age of 382 Myr can be estimated from the cosmogenic 38Ar using the production rate of the 2π exposure geometry, and implies that the particle may have originated in the mature regolith of an asteroid.


Micrometeorites are extraterrestrial particles smaller than 1 mm in size and have been found in the polar regions (e.g., Maurette et al. 1986; Taylor et al. 1997) and also deep-sea sediment (e.g., Murray and Renard 1881; Bruun et al. 1995; Yada et al. 1996). Micrometeorites can be roughly classified into two types: irregularly shaped unmelted micrometeorites and completely melted cosmic spherules. Their difference in shape is due to the different temperatures they were exposed to during atmospheric entry. The maximum temperature with deceleration heating depends on their velocity, entry angle, size, and mass. And although cosmic spherules do not get completely vaporized, they are exposed to high temperatures and their volatile elements get considerably depleted, thus making measuring them for volatile elements such as noble gases extremely difficult.

To identify the spherules as being extraterrestrial, cosmic-ray-produced nuclides, i.e., 53Mn, 59Ni, 10Be, and 26Al, were measured by several research groups (e.g., Nishiizumi 1983; Matsuzaki et al. 2000; Nishiizumi et al. 2007). However, cosmogenic nuclides could only be detected in large spherules (over a few tens of μg in weight), and it was very difficult to discover any extraterrestrial signs in typical-size spherules, which are usually smaller than 150 μm in diameter, due to the low concentrations of the target nuclides.

Noble gases are sensitive indicators of extraterrestrial material because of their distinctive isotopic compositions; however, measuring noble gases in cosmic spherules is very difficult due to their extremely low concentrations. Using laser extraction techniques, Osawa et al. (2003a) achieved an ultra-low background method allowing noble gas isotopic measurements of 31 individual cosmic spherules. Osawa et al. (2003a) detected extraterrestrial noble gases in about 40% of Antarctic cosmic spherules with diameters of 35–250 μm, and also discovered an enigmatic spherule with a 40Ar/36Ar ratio of 566.3 ± 14.8. This ratio is much higher than the terrestrial atmospheric value, and given the spherule’s solar He and Ne, the Ar was attributed to extraterrestrial processes. Micrometeorites do not have high 40Ar/36Ar ratio, so evidently some cosmic spherules originate from different parent bodies than unmelted micrometeorites.

Many questions pertaining to the anomalous extraterrestrial dust samples can be raised. Do achondritic micrometeorites exist? Can we recognize micrometeorites from the Kuiper Belt? Can we associate specific types of micrometeorites with classes of meteorites? Gounelle et al. (2005) reported an unambiguous basaltic micrometeorite, this exception seems to prove the rule that achondritic micrometeorites are extremely rare. Assuming that other unusual spherules might be found, we developed a research plan to further investigate this class of extraterrestrial material. An additional motivation for this class of objects is the very fact that they have been melted. The extent of heating during atmospheric entry is determined by a variety of factors, some of which may depend on the parent body. For example, any dust generated from comets with a perihelia <l AU would have an entry velocity >15 km s−1, whereas dust from an asteroid with low eccentricity and low inclination typically has an entry velocity of at most 15 km s−1 (Love and Brownlee 1991). A numerical model of the atmospheric entry of micrometeoroid particles constructed by Love and Brownlee (1991) concluded that most spherules larger than 50 μm in diameter would have originated in asteroids, whereas nearly all small spherules of <30 μm in diameter originate in comets. The orbit of asteroidal dust strongly depends on that of its parent body (Sykes 1990). Hence, the distribution of parent bodies of melted spherules should systematically differ from unmelted micrometeorites that fall within the same size range due to severe mass loss in spherules during atmospheric entry.

To clarify their source and evolutional history, 99 spherules collected from the Meteorite Ice Field around the Yamato Mountains in Antarctica were investigated. The interior texture, chemical composition, and noble gas isotopes of individual cosmic particles were investigated in detail. A highly sensitive mass spectrometer and laser gas-extraction system together with the extremely low background levels and accurate correction of mass interference (Osawa 2004) enabled single spherule noble gas analysis.

Materials and Methods

The spherules were obtained by filtering melted ice collected during the 39th Japan Antarctic Research Expedition in 1998 (Yada et al. 2004). The spherules were collected from three sampling sites—Kuwagata Nunatak #5 and #7 (approximately 2200 m in altitude) and Minami-Yamato Nunataks #2 (approximately 2400 m in altitude)—within the Meteorite Ice Field around the Yamato Mountains in Antarctica. Kuwagata Nunatak is located about 40 km north of Minami-Yamato Nunataks. The Antarctic micrometeorites gathered from the fields were not fresh particles. Based on δ18O variation, Yada et al. (2004) reported that the top of the ice core taken from the Kuwagata area is estimated to be 27,000–33,000 yr old. Using numerical simulation of ice flow, Azuma et al. (1985) conclude that the age of the ice in the Minami-Yamato area must be younger than that in the Kuwagata area.

Scanning electron microscope (SEM) observations of the surface of each spherule were performed using the JEOL (Tokyo, Japan) JSM-5600 LV SEM at Ibaraki University in low vacuum mode (25 Pa in the sample chamber) so as to avoid use of carbon coating. The individual size of the spherules was also determined through SEM observation. All the spherules were embedded separately in soluble acetone glue. They were lapped and polished using several lapping sheets embedded with 1, 3, 9, and 30 μm size Al2O3 abrasive grain.

The internal texture of the spherules was observed using a JSM-5600 SEM in high vacuum mode and the spherules then classified into one of six types (refer to Table 1). The soluble acetone glue vesiculated so rapidly under electron beam irradiation that an Oxford ISIS 300 energy-dispersive spectrometer (EDS) was used instead of a wavelength-dispersive spectrometer (WDS) because the low electron current (approximately 0.3 nA) during EDS analysis resulted in less thermal modification of the embedding glue. Bulk elemental concentrations were determined by averaging three to ten analyses, each obtained by rastering the electron beam (∼10 μm × ∼8 μm) across the sample (Table A1). The averaged Si and CI chondrite normalized elemental abundances of Mg, Al, Ca, and Fe were obtained for each spherule using 100%-normalized analyses. Following the SEM/EDS analysis each embedded spherule, on a small (∼2 mm × ∼2 mm) chip of glass slide, was placed in a 10 mL beaker with 10 mL of acetone added to it. The embedding glue was completely dissolved and removed by acetone at 70 °C. The acetone was replaced a total of three times for each sample to remove completely any remnants of glue that could interfere with the noble gas isotopic composition.

Table 1.   Description of petrographic types of cosmic spherules.
TypeSubtypeAbbreviationDescriptionNumber of samples
Cryptocrystalline CNontransparent glassy spherules containing a small amount of crystallites13
Barred olivine BContaining fern-like-shaped or platy olivine crystal with abundant small (typically <2 μm) magnetite embedded in interstitial glass40
PorphyriticAP_ASmall amounts of olivine and magnetite embedded in glass with a pyroxene-like composition3
PorphyriticBP_BPorphyritic olivine crystals and dendritic magnetite embedded in glass (sometimes with a small amout of low-Ca pyroxene)24
PorphyriticCP_CFine-grained (typically <5 μm) crystals (typically olivine and magnetite) with a small amount of glass and abundant voids
Magnetite rim on the surface of the spherules
Partially melted PMFine-grained (typically <10 μm) crystals (mainly olivine) and mesosteses filled by small (<2 μm) crystals and minor interstitial glass with small voids (<5 μm)
Magnetite rim on the surface of the spherules

Before analysis the weight of each spherule was determined using a precision balance capable of making measurements in increments of 0.1 μg. The weights given in Table 2 are the averages of several measurements. Very light particles cannot be weighed so their weights are assigned as <0.2 μg. The noble gas measurement procedure has already been detailed by Osawa et al. (2003a). Neon mass interference caused by 40Ar++ (m/z = 20) and CO2++ (m/z = 22) was corrected using experimentally determined 40Ar++/ 40Ar+ and CO2++/CO2+ ratios (Osawa 2004). During the Ne analysis, argon and carbon dioxide were removed with a liquid-nitrogen-cooled trap close to the ion source to reduce mass interferences at 20Ne and 22Ne. Blanks were determined by heating an empty sample holder with a laser. Noble gas measurements were carried out in two series of runs with blank corrections being carried out for all the samples. The blanks used in runs 1 and 2 are given in Table 2. The uncertainty in the noble gas content was determined by propagating quadratically counting statistical errors and the variance in the blank subtraction. All data shown in Table 2 are blank corrected. Because these samples are small, and the noble gas concentrations are low, many samples had a negative value for the noble gas concentration after the blank correction. If the addition of a two sigma error left the value negative, this sample is considered to have a concentration indistinguishable from blank. If, on the other hand, the addition of two sigma error resulted in a greater than zero concentration the number is enclosed in parentheses. If the concentration is greater than blank content, the number is not enclosed in parentheses. 4He, 20Ne, 36Ar, 84Kr, and 132Xe contents without parentheses have <12, <13, <24. <24, and <42% uncertainties, respectively.

Table 2.   Noble gas contents of cosmic spherules.
SampleRunWeight (μg)Size (μm × μm)Density (g cm−3)Typea4He20Ne36Ar84Kr132Xe
(10−12 cm3 STP)(10−15 cm3 STP)
  1. Note: Blank = noble gas content is comparable to blank level; n.m. = not measured; n.d. = not detected.

  2. aSpherules are classified into six types based on their interior structures. Abbreviations are shown in Table 1.

  3. bAverage values of six measurements.

  4. cAverage values of three measurements.

M24000510.4 ± 0.399 × 731.2C48204.2Blank(0.20)
M24004918.6 ± 0.4241 × 1662.0BBlankBlankBlankBlank0.27
M24006710.2 ± 0.263 × 502.0BBlankBlankBlankBlankBlank
M2401221<0.262 × 40<3.0P_C(8.5)(0.55)(0.019)Blank(0.12)
M24016011.9 ± 0.2102 × 993.6P_B(6.8)5.10.19Blank(0.20)
M24017210.3 ± 0.156 × 552.8P_BBlank(0.50)BlankBlank0.35
M24019111.3 ± 0.199 × 853.3B(6.5)Blank(0.028)(4.0)(0.20)
M2401951<0.265 × 53<1.9BBlank(0.047)(0.017)Blank0.57
M24019610.8 ± 0.199 × 911.8PM(5.5)(0.67)0.10BlankBlank
M2402001<0.249 × 47<3.4CBlankBlankBlankBlankBlank
M2402021<0.281 × 74<0.8B(0.70)BlankBlankBlank(0.20)
M2402052<0.258 × 57<2.0B5.1BlankBlankBlankBlank
M24021721.3 ± 0.487 × 863.8CBlankBlankBlankBlankBlank
M2402331<0.241 × 40<5.8BBlankBlankBlankBlankBlank
M2402342<0.263 × 57<1.8P_C(2.5)(0.07)0.063BlankBlank
M2402462<0.255 × 50<2.6BBlankBlankBlankBlankBlank
M24026411.3 ± 0.193 × 833.7P_BBlankBlank(0.014)Blank0.27
M2402682<0.258 × 55<2.1CBlankBlankBlankBlankBlank
M2402782<0.256 × 54<2.3P_B(0.83)Blank(0.036)(0.48)0.67
M2402841<0.254 × 52<2.6P_A(1.4)BlankBlankBlankBlank
M24028610.6 ± 0.573 × 722.8P_B(10.6)0.780.087Blank0.35
M2402871<0.254 × 51<2.6B(1.9)BlankBlank(2.7)0.27
M2402891<0.266 × 62<1.5CBlankBlankBlankBlank(0.12)
M24029011.7 ± 0.3115 × 1062.4B(0.48)Blank(0.021)BlankBlank
M24029120.7 ± 0.377 × 753.0PM(0.37)Blank(0.010)(1.6)0.77
M24029211.6 ± 0.398 × 874.0BBlankBlank0.04102(4.7)0.42
M24029311.9 ± 0.2102 × 924.1BBlankBlankBlankBlankBlank
M24029410.9 ± 0.480 × 753.6P_BBlank1.10.064Blank0.35
M24029810.9 ± 0.473 × 714.5P_BBlankBlankBlank(3.3)0.94
M2403041<0.256 × 52<2.4B(2.0)BlankBlank(3.3)0.57
M2403052<0.239 × 36<7.3P_A3.4BlankBlankBlank0.43
M24030611.3 ± 0.284 × 794.5P_BBlankBlankBlankBlankBlank
M24031720.3 ± 0.358 × 523.5BBlankBlank(0.017)(1.0)Blank
M24031810.2 ± 0.258 × 541.6P_B53200.46Blank(0.20)
M24032720.2 ± 0.269 × 491.9BBlankBlankBlankBlank0.62
M24032810.4 ± 0.464 × 603.4BBlankBlankBlankBlank(0.12)
M24033511.5 ± 0.398 × 953.1B(5.3)Blank(0.014)(5.6)0.42
M24035020.3 ± 0.362 × 483.5BBlankBlankBlankBlankBlank
M24035512.0 ± 0.1108 × 1013.3P_ABlankBlankBlankBlankBlank
M2403581<0.261 × 57<1.9P_B23(0.71)0.043BlankBlank
M24036212.0 ± 0.3109 × 973.4P_B(2.2)0.91(0.018)Blank(0.20)
M2403662<0.262 × 49<2.3BBlankBlankBlankBlankBlank
M2403692<0.275 × 47<1.8C54BlankBlankBlank0.41
M2403782<0.258 × 41<3.3BBlankBlankBlank(0.25)0.99
M24038021.2 ± 0.1111 × 792.8B(0.52)Blank(0.046)2.4(0.30)
M2403981<0.241 × 39<5.9P_B(1.8)BlankBlankBlankBlank
M2404002<0.247 × 45<3.9CBlankBlank(0.010)BlankBlank
M24040210.4 ± 0.371 × 622.3P_BBlankBlankBlankBlank0.50
M24041011.9 ± 0.1120 × 1082.5PM57311.5Blank1.1
M2404131<0.251 × 47<3.2PM(2.3)Blank0.054(3.3)(0.12)
M24041410.9 ± 0.389 × 773.1B(2.2)Blank(0.026)(3.3)0.42
K54009311.4 ± 0.2105 × 962.7BBlankBlankBlankBlank(0.12)
K54012111.3 ± 0.2101 × 952.6B(3.6)Blank(0.030)Blank(0.12)
K54020110.5 ± 0.179 × 722.0BBlankBlankBlankBlank(0.12)
K5402032<0.248 × 48<3.5C(0.71)Blank(0.013)(0.40)0.71
K54020912.1 ± 0.4107 × 1043.4P_B(22)6.90.24Blank0.57
K5402152<0.248 × 39<4.7BBlank1.4Blank3.70.80
K54022210.2 ± 0.262 × 591.7B(2.0)BlankBlankBlank(0.12)
K5402302<0.244 × 43<4.6P_B(1.2)Blank(0.031)(1.7)Blank
K54023210.9 ± 0.377 × 744.2P_B(6.4)110.44BlankBlank
K5402482<0.256 × 55<2.2BBlankBlankBlank(1.8)0.43
K54025210.6 ± 0.379 × 762.6P_B(4.1)Blank(0.013)Blank0.72
K54025512.3 ± 0.2116 × 993.6BBlankBlank(0.017)(5.2)(0.20)
K54025711.1 ± 0.2112 × 1011.7PM(7.5)Blank0.19Blank0.79
K54026111.0 ± 0.291 × 783.3B(2.7)BlankBlankBlank0.42
K54026210.3 ± 0.362 × 612.1C(3.4)BlankBlankBlankBlank
K5402682<0.262 × 55<1.9B(0.42)BlankBlank(1.6)Blank
K54028510.2 ± 0.249 × 492.5P_B(3.4)0.840.040Blank(0.12)
K5402862<0.290 × 66<0.8B538361.92.90.91
K54029020.7 ± 0.481 × 772.7P_B(1.5)4.00.12(0.71)Blank
K54029512.0 ± 0.1112 × 1043.1PM24(0.44)0.063(5.6)0.50
K54030510.5 ± 0.471 × 692.9P_B(1.9)BlankBlankBlank0.72
K54030910.8 ± 0.192 × 882.0B(3.0)BlankBlankBlank0.64
K54031810.3 ± 0.375 × 681.5C(3.1)BlankBlankBlank0.72
K74000311.8 ± 0.388 × 875.1B(3.1)Blank(0.016)BlankBlank
K7400382<0.248 × 48<3.4P_C(2.9)Blank(0.014)(0.79)Blank
K74010710.3 ± 0.357 × 493.9B(3.1)BlankBlankBlankBlank
K74011710.2 ± 0.261 × 591.8P_B1235.80.17Blank0.50
K74011811.5 ± 0.189 × 894.1P_B(8.7)5.60.19BlankBlank
K74011910.2 ± 0.258 × 571.8P_C(14)1.50.090Blank0.57
K7401221<0.248 × 47<3.7P_C405.70.36Blank0.6
K74012612.4 ± 0.2121 × 1202.6PM702.10.057Blank(0.20)
K74012811.6 ± 0.2107 × 1062.5C(2.2)0.860.10Blank0.4
K7401341<0.241 × 38<6.3BBlankBlank(0.013)(4.4)Blank
K7401392<0.246 × 46<3.9PM11Blank(0.039)2.2Blank
K74014812.1 ± 0.3120 × 1022.9B(11)Blank(0.016)8.41.0
K7401501<0.270 × 70<1.1C(3.6)BlankBlankBlankBlank
K7401522<0.247 × 47<3.7P_C29(0.11)BlankBlankBlank
K74015711.1 ± 0.3105 × 1031.9P_C(3.7)Blank0.059Blank0.5
K74016310.6 ± 0.474 × 713.1P_CBlank(0.35)0.10Blank(0.20)
K7401722<0.247 × 47<3.7B(1.5)Blank(0.023)(0.48)Blank
K74017311.5 ± 0.3102 × 883.4P_B48241.59.3(0.20)
K74017410.4 ± 0.486 × 592.0B(5.8)Blank0.037BlankBlank
K7401762<0.241 × 41<5.4P_B(1.0)1.10.10(1.2)(0.23)
K74018011.0 ± 0.3100 × 932.0PM(7.4)(0.52)0.29Blank0.42
K7401812<0.248 × 48<3.5PM171.60.314.21.5
K7401822<0.254 × 54<2.4B(1.7)(0.084)(0.036)4.8Blank
Blankb1    230.770.0356.40.27
Blankc2    3.40.450.0531.90.37


Interior Texture and Classification of Spherules

Figure 1 gives the backscattered electron (BSE) images of the surfaces and interiors of 12 of the spherules. The spherules are classified into six types using the taxonomy proposed by Taylor et al. (2000), with the classification of the spherules used in this work being given in Table 1. Spherules containing only a small amount of fine-grained crystals in their interiors are classified as cryptocrystalline spherules (Figs. 1a–d), and account for 13% of the spherules investigated. As revealed in Fig. 1a their surfaces are often covered by fine-grained dendritic crystals, probably magnetite. Spherule M240217 (Figs. 1c and 1d) is not transparent under an optical microscope and hence is classified as cryptocrystalline, although no crystals can be readily recognized in Fig. 1d.

Figure 1.

 Backscattered electron (BSE) images of the surface and cross sections of spherules. a–d) Cryptocrystalline spherules. e–h) Barred olivine spherules. i,j) Porphyritic subtype A spherule. k,l) Porphyritic subtype B spherule. m,n) A typical porphyritic subtype C spherule. o,p) A porphyritic subtype C spherule containing large voids. q,r) Porphyritic subtype B spherule containing abundant spinels. s,t) A partially melted spherule with a reduction texture. u,v) A partially melted spherule showing a long cosmic-ray exposure (CRE) age (discussed later). w,x) A partially melted spherule. Ol = olivine; Mt = magnetite; Px = pyroxene; Gl = glass; V = void; Mes = mesostasis; Chr = chromite; Sp = spinel.

The majority of the spherules (40%) investigated in this study are barred olivine spherules (Figs. 1e–h; Table 1), and are composed of dense aggregates of comb-like olivine crystals. The interstices of the crystals are composed of abundant quantities of magnetite crystals and glass, with their surfaces being covered by characteristic fern-shaped magnetite crystals.

In this study, the term porphyritic spherules are applied to spherules with a microporphyritic texture, even though Taylor et al. (2000) distinguished porphyritic spherules from relic-grain-bearing ones, and account for 36% of the spherules investigated. We further divide the porphyritic spherules into three subtypes. Porphyritic subtype A spherules contain abundant amounts of glass with small amounts of magnetite and olivine (Figs. 1i and 1j). And while the dark areas in Fig. 1j have a chemical composition similar to low-Ca pyroxene they do not have the stoichiometry of low-Ca pyroxene, leading to the conclusion that they may have only been partially melted. Small amounts of magnetite dendrites and a few protrusions due to underlying phenocrysts can be observed on their surface (Fig. 1i). The most abundant porphyritic spherule subtype is porphyritic subtype B (Figs. 1k and 1l), accounting for 24% of the spherules. Typically, they contain well-developed large magnetite dendrites and olivine phenocrysts on their surfaces and interiors. In Fig. 1l, low-Ca pyroxene can be seen to be overgrowing the olivine phenocrysts. Porphyritic subtype C spherules are characterized by their abundance of small (typically <5 μm in diameter) euhedral olivines in mesostasis and minor amounts of glass (Figs. 1m and 1n). They often contain voids due to vesiculation when they melted (Fig. 1n). In addition to the ferromagnesian spherules, we also found a porphyritic subtype B spherule with a relatively refractory bulk composition. It is composed of spinel phenocrysts and interstitial glass (Figs. 1q and 1r). The phenocrysts are composed of Al-rich cores and Cr-rich rims.

Partially melted spherules represent a transition between porphyritic spherules and scoriaceous spherules. They are characterized by fine-grained (typically <10 μm) crystals (mainly olivine) and mesostasis filled by small (<2 μm) crystals and interstitial glass with small voids (<5 μm). Abundant fine-grained (<2 μm) magnetite crystals exist on their surface (Figs. 1s–x). The presence of mesostasis instead of glass indicates lower degrees of melting of the partially melted spherules than of porphyritic ones. Twenty spherules belong to this type (Table 1).

Because of their low degrees of melting, they could contain abundant relict crystals. For example, K740126 (Figs. 1s and 1t) contains Fe-rich olivine crystals with thin (<2–3 μm thick) magnesian rims, indicative of reduction during melting (Fig. 1t). It also contains euhedral chromite. Its mesostasis and surface contain an abundance of magnetite grains (Figs. 1s and 1t). M240410 with its extremely long cosmic-ray exposure (CRE) age (discussed later) (Figs. 1u and 1v) and M240413 (Figs. 1w and 1x) also belong to this type.

We believe the internal textures described above are related to the abundance of relic crystals when the spherules melted. The cryptocrystalline spherules probably completely melted, thus indicating higher heating temperatures than the liquidus of their constituent phases. Their higher SiO2 content relative to the other spherules might have inhibited the growth of crystals in them during cooling. The barred olivine spherules may also not have contained any relic crystals when they melted. However, depending on their bulk chemical composition, the dendritic olivine and magnetite were crystallized during the cooling. In a flash heating experiment conducted by Connolly et al. (1991), 63–125 μm size fayalite powder produced a glass texture at temperatures of 1500 and 1600 °C but resulted in a barred olivine texture at 1400 °C. Radomsky and Hewins (1990) reported that barred textures form at or just below the temperature at which olivine disappears (within 30 min of being heated). Porphyritic spherules formed during partial melting of their precursor material. Dynamic crystallization experiments performed by Radomsky and Hewins (1990) revealed that porphyritic textures form from near or 30 °C below the temperature at which olivine disappears. Although the three types of porphyritic spherules continuously change, their degree of melting probably decreases from type A through to C. Partially melted spherules were the least melted of all the spherules. The bulk chemical composition and texture of the spherules suggest that the maximum heating temperature during melting decreased in the following order: cryptocrystalline, barred olivine, subtype A to C porphyritic, and partially melted.

Chemical Composition of Spherules

Figure 2 shows (Si+Al)-Mg-Fe ternary diagrams of the bulk chemical composition of the spherules investigated in this study. Their composition is very similar to those retrieved from the South Pole water well (SPWW) (Taylor et al. 2000). The majority of the spherules can be seen to be plotted between the serpentine solid solution line and the olivine solid solution line with Mg/(Mg+Fe) ratios of 0.5–0.8, although the cryptocrystalline ones did tend to have more magnesian and have higher (Si+Al)/(Si+Al+Mg+Fe) ratios than the others.

Figure 2.

 (Si+Al)-Mg-Fe (atomic) ternary diagrams of the bulk chemical compositions of the spherules investigated. a) Cryptocrystalline spherules. b) Barred olivine spherules. c) Porphyritic subtype A spherules. d) Porphyritic subtype B spherules. e) Porphyritic subtype C spherules. f) Partially melted spherules. Spherules appearing in Figs. 1q–x are indicated by arrows.

K740126, with its reduction texture, M240358, containing abundant spinel, and M240410, with its long exposure age (discussed later) do not deviate from the majority of the spherules (Figs. 2d and 2e). Conversely, M240413 deviates from the majority (Fig. 1e). It has the highest (Si+Al)/(Si+Al+Mg+Fe) ratio, reflecting the large amount of glass in the cross section (Fig. 1x).

Figure 3 gives Si and CI chondrite normalized abundance of Ca, Al, Mg, Si, and Fe in the spherules investigated. Although the medians and mean values of Si and CI chondrite normalized Fe abundance in all the types of spherules are lower than unity, cryptocrystalline and porphyritic subtype A are below the 0.5 CI normalized Fe abundance. The medians and mean values of Ca, Al, and Mg are within a factor of two although the Ca and the Al abundances have large variances. The highest Ca and Al values in the porphyritic subtype B spherules are due to the spinel-rich spherule M240358 (Figs. 1q, 1r, and 3d). It contains 12.3 times and 3.5 times CI normalized Ca and Al, respectively, so it may have been caused by a melted Ca- and Al-rich inclusion.

Figure 3.

 Si and CI chondrite normalized abundances of Ca, Al, Mg, Si, and Fe. In this figure, box plots are utilized to depict the distributions of elemental abundances. In these plots, the smallest and the largest values are shown as asterisks, the lower and upper quartiles are represented by the upper and lower edges of the boxes, the medians are horizontal lines within the boxes, and the mean values are represented by stars. a) Cryptocrystalline spherules. b) Barred olivine spherules. c) Porphyritic subtype A spherules. d) Porphyritic subtype B spherules. e) Porphyritic subtype C spherules. f) Partially melted spherules.

With unmelted micrometeorites it is known that the noble gas concentrations are controlled not only by temperature but also by the degree of aqueous alteration in glacier ice (Osawa et al. 2003b). Unmelted micrometeorites collected around the Yamato Mountains have systematically lower Mg/Si ratios than recently deposited unmelted micrometeorites and the Mg/Si ratios correlate with the age of the ice (Terada et al. 2001). These results indicate that aqueous alteration occurs because jarosite, an alteration product of iron oxides and sulfide, is found in many unmelted Yamato Mountains micrometeorites, but not significantly in recently deposited unmelted micrometeorites (Terada et al. 2001; Osawa et al. 2003b). Conversely, the Mg/Si ratios of the spherules have a larger distribution and a higher average value than those of unmelted micrometeorites in the study by Terada et al. (2001). In this work, no significant depletion of Mg is observed, as shown in Fig. 3. Unmelted irregular-shaped micrometeorites are more strongly affected by surface-correlated reactions than rounded spherules due to their high body surface area. As seen in Fig. 1, alteration minerals are found on the surface of some of the spherules. However, the body surface area of a spherule is considerably smaller than that of an unmelted micrometeorite and the effect of aqueous alteration on spherules should be weaker than on unmelted micrometeorites. Indeed there is no significant correlation between the noble gas content and Mg/Si ratio. In this work, no significant sign of noble gas depletion due to secondary alteration in ice is discovered, in contrast to the case of unmelted micrometeorites (Osawa et al. 2003b).

Noble Gas Contents and Spherule Types

There is a weak correlation between the noble gas content and type of spherule. The spherules are categorized into three groups here: (1) cryptocrystalline and barred olivine spherules; (2) porphyritic subtype A and B spherules; and (3) porphyritic subtype C and partially melted spherules. The ratio of the number of spherules with undetectable amount of noble gases is an index of the degree of heating because noble gas contents of a considerable number of samples cannot be determined. Table 3 summarizes the ratios of the three groups. There is a clear correlation between the ratio and the spherule category, indicating the relationship between the degree of heating and morphologic nature of the spherules. The correlation is not due to the bias of the distribution of particle size because the average diameters of each group are 72.6 ± 28.4, 70.6 ± 20.7, and 74.7 ± 27.5 μm, respectively. The correlation presumably reflects different degrees of deceleration heating and/or different parent populations.

Table 3.   Ratios of the number of spherules with undetectable amount of noble gases.
GroupNumber of samplesNumber of blankaRatio of blank
  1. aNumber of blank shown in Table 2.

Cryptocrystalline & barred olivine532447320.450.890.60
Porphyritic A & B2771290.260.440.33
Porphyritic C & partially melted191710.050.370.05

A surprising fact is that many cryptocrystalline type spherules have preserved detectable amounts of noble gases despite severe heating. In evaporation experiments with the melting of FeO-MgO-SiO2-CaO-Al2O3, a cryptocrystalline texture can appear after very severe heating (e.g., heated at 2000 °C for 3.59 min), with more than 35.5% of the mass having been vaporized (Hashimoto 1983). The melting texture found in cryptocrystalline spherules indicates that the maximum temperature during their passage through Earth’s atmosphere is higher than the liquidus temperature. The retentiveness of the noble gases in these spherules, therefore, might be related to their loss of mass during atmospheric entry. After the particle loses much of its mass it is smaller and becomes a more efficient radiator and cools (Love and Brownlee 1991). As any loss of mass greatly shortens the heating duration, detectable amounts of noble gases can be retained in spherules despite high temperature heating.

The geometric mean values of the noble gases in the spherules and the unmelted micrometeorites are summarized in Table 4. As it is known that the distribution of noble gases concentrations in micrometeorites is not normal but logarithmic, geometric mean is more suitable than the arithmetic mean (Osawa et al. 2003b). The volume-normalized noble gas content values are also provided for the sake of comparison. Two methods of calculation were adopted for the spherules measured in this work: (1) only the samples with detectable amount of noble gases (numbers in parenthesis in Table 2 are included) were used to calculation and (2) upper limit of noble gas contents of the samples with undetectable amount of noble gases was assumed as two sigma uncertainties of blank. For example, 0.4 × 10−12 cm3 STP of 4He is assumed for the upper limit of the samples with undetectable amount of 4He. The values for the spherules measured in this work are probably overestimated, particularly in the former case, because only samples with detectable amounts of noble gases were used in calculating the geometric mean values. The geometric mean values of noble gas contents of the Tottuki spherules are significantly higher than those of the Yamato Mountains spherules calculated by the second method. That difference is presumably due to the different size distributions of the samples because the average size of a Tottuki spherule is 101 ± 46 μm in diameter, or 1.39 times larger than the Yamato Mountains spherules. As the volume-normalized contents of the Tottuki spherules are comparable to those of the Yamato Mountains spherules, no large shift in the noble gas content is evident between the two sets of spherules. On the other hand, the normalized geometric mean values for the 4He, 20Ne, and 36Ar content of unmelted micrometeorites are significantly higher than those of the Yamato Mountains spherules. The deficit of 4He, 20Ne, and 36Ar in the spherules is due to heating. If the precursors of the spherules had comparable noble gas contents relative to the unmelted micrometeorites before atmospheric entry, we concluded that about 90% of the Ne and Ar and about 95% of the He must have been lost through heating. Moreover, even the unmelted micrometeorites were severely heated during atmospheric entry; we estimate that >95% of the light noble gases originally preserved in the precursors of the spherules before atmospheric entry were depleted.

Table 4.   Geometric mean values of noble gas contents.
SampleSize rangea (μm)4He20Ne36Ar84Kr132XeVolume normalizedbReference
(10−12 cm3 STP)(10−15 cm3 STP)4He20Ne36Ar84Kr132Xe
  1. Note: AMMs = Antarctic micrometeorites.

  2. aMean diameter and standard deviation.

  3. bNormalized for particle volume and noble gas contents of unmelted AMMs (Yamato Mountains).

  4. cOnly samples with detectable amount of noble gases are used.

  5. dUpper limits of noble gas contents for the samples with undetectable amount of noble gases are assumed.

Spherules (Yamato Mountains)72.5 ± 26.24.771.590.0612.210.380.070.780.181.290.32This workc work (all samples)d
Spherules (Tottuki)101 ± 45.917.20.6460.1022.040.610. et al. (2003a)
Unmelted AMMs (Yamato Mountains)86.1 ± 41.51143.420.5582.871.9911111Osawa et al. (2003b)
Unmelted AMMs (Dome Fuji)118 ± 41.52429.372.2213.85.60.831.071.551.871.10Osawa and Nagao (2002)
Interplanetary dust particles22.3 ± 1080.92.520.265  40.842.227.2  Kehm et al. (2002)

Helium and Neon Isotopic Composition

Table 5 gives noble gas isotopic composition of the spherules analyzed in this work. The helium isotopic ratios can be determined for 26 of the 99 samples, and all of them have 3He/4He ratios higher than that of terrestrial air within one sigma error. Fifteen of them have 3He/4He ratios higher than that of terrestrial air within two sigma error, proving their extraterrestrial origin. Figure 4 gives the 4He content and 3He/4He ratios of the spherules and unmelted micrometeorites (Osawa et al. 2003b). A lot of the spherules have He isotopic ratios close to that of solar wind (SW) of 4.53 ± 0.03 × 10−4 (Heber et al. 2008), as well as the unmelted micrometeorites, indicating that the spherules have preserved solar-derived He despite their severe heating.

Table 5.   Noble gas isotopic compositions of Antarctic cosmic spherules.
Sample3He/4He (10−4)20Ne/22Ne21Ne/22Ne40Ar/36Ar38Ar/36ArET noble gasa
  1. Note: Blank = indistinguishable from blank. All data are corrected for blank. Errors are one sigma.

  2. aPresence of extraterrestrial noble gas.

  3. bOnly the upper limit can be determined due to small amount of 40Ar.

M2400058.2 ± 2.311.2 ± 0.20.031 ± 0.0031.4 ± 0.30.194 ± 0.003Yes
M240067BlankBlankBlank400 ± 223Blank 
M2401226.5 ± 5.410.9 ± 1.00.040 ± 0.018<77b0.163 ± 0.154Yes
M2401608.8 ± 5.310.7 ± 0.20.031 ± 0.003<3b0.187 ± 0.031Yes
M240172Blank9.8 ± 1.20.016 ± 0.021<38bBlankYes
M240191BlankBlankBlank332 ± 430.185 ± 0.122 
M240195BlankBlankBlank258 ± 330.210 ± 0.173 
M2401968.2 ± 5.510.1 ± 1.10.037 ± 0.01434 ± 110.176 ± 0.035Yes
M240202BlankBlankBlank467 ± 437Blank 
M2402059.7 ± 8.5BlankBlankBlankBlank 
M240210BlankBlank0.115 ± 0.081BlankBlank 
M240229BlankBlankBlank223 ± 113Blank 
M240234BlankBlankBlankBlank0.226 ± 0.031 
M240264BlankBlankBlank380 ± 630.261 ± 0.211 
M240284BlankBlankBlank440 ± 318Blank 
M2402866.9 ± 2.910.3 ± 0.50.030 ± 0.008266 ± 110.187 ± 0.039Yes
M240287BlankBlankBlank438 ± 221Blank 
M240289BlankBlankBlank247 ± 217  
M240290BlankBlankBlank310 ± 430.166 ± 0.131 
M240291BlankBlankBlank209 ± 93Blank 
M240292BlankBlankBlank289 ± 240.169 ± 0.079 
M240294Blank10.3 ± 0.60.025 ± 0.010<1b0.197 ± 0.051Yes
M240304BlankBlankBlank417 ± 132Blank 
M240305BlankBlankBlank218 ± 127Blank 
M2403183.3 ± 0.711.5 ± 0.30.031 ± 0.00215 ± 30.194 ± 0.013Yes
M240328BlankBlankBlank232 ± 94Blank 
M240335BlankBlankBlank249 ± 480.129 ± 0.205 
M2403585.0 ± 1.311.5 ± 1.40.026 ± 0.014<31b0.168 ± 0.067Yes
M24036223 ± 2110.7 ± 0.70.027 ± 0.009<119b0.176 ± 0.157Yes
M2403696.3 ± 1.5BlankBlankBlankBlankYes
M240380BlankBlankBlank94 ± 560.218 ± 0.036Yes
M240398BlankBlankBlank344 ± 133Blank 
M240400BlankBlankBlank166 ± 680.169 ± 0.134 
M24041097 ± 118.9 ± 0.10.203 ± 0.0062 ± 10.256 ± 0.005Yes
M2404131060 ± 800BlankBlank127 ± 150.211 ± 0.065Yes
M240414BlankBlankBlank295 ± 300.192 ± 0.113 
K540121BlankBlankBlank292 ± 30Blank 
K540201BlankBlankBlank233 ± 66Blank 
K540203BlankBlankBlank137 ± 102Blank 
K5402093.4 ± 1.310.9 ± 0.50.029 ± 0.003<9b0.184 ± 0.018Yes
K540230BlankBlankBlank105 ± 730.194 ± 0.040Yes
K540232Blank11.0 ± 0.10.031 ± 0.005<1b0.201 ± 0.013Yes
K540252BlankBlankBlank244 ± 62Blank 
K540255BlankBlankBlank260 ± 490.157 ± 0.166 
K540257BlankBlankBlank55 ± 60.176 ± 0.022Yes
K540285Blank10.6 ± 0.70.013 ± 0.008<29b0.185 ± 0.072Yes
K5402863.0 ± 0.511.1 ± 0.10.029 ± 0.0029 ± 30.205 ± 0.002Yes
K540290Blank11.0 ± 0.50.031 ± 0.006Blank0.192 ± 0.020Yes
K5402953.2 ± 0.79.2 ± 1.20.015 ± 0.026140 ± 190.188 ± 0.047Yes
K540305BlankBlankBlank173 ± 134Blank 
K540318BlankBlankBlank270 ± 76Blank 
K740003BlankBlankBlank181 ± 480.179 ± 0.186Yes
K740038BlankBlankBlank199 ± 530.234 ± 0.112 
K7401172.8 ± 0.510.3 ± 0.30.027 ± 0.0049 ± 70.190 ± 0.023Yes
K7401187.2 ± 2.710.7 ± 0.40.034 ± 0.00416 ± 60.200 ± 0.026Yes
K7401193.9 ± 3.510.6 ± 0.70.029 ± 0.01025 ± 130.188 ± 0.044Yes
K7401222.5 ± 0.910.8 ± 0.30.030 ± 0.004<5b0.192 ± 0.015Yes
K7401264.7 ± 0.510.1 ± 0.50.051 ± 0.007<23b0.170 ± 0.055Yes
K740128Blank10.3 ± 1.20.034 ± 0.00925 ± 110.204 ± 0.034Yes
K740134BlankBlankBlank269 ± 52Blank 
K740139BlankBlankBlank93 ± 670.116 ± 0.073Yes
K7401481.10 ± 1.06BlankBlank373 ± 670.167 ± 0.184 
K7401524.3 ± 1.3BlankBlankBlank0.211 ± 0.031Yes
K7401574.5 ± 5.0BlankBlank34 ± 200.192 ± 0.048Yes
K740163Blank11.3 ± 1.80.024 ± 0.023206 ± 100.180 ± 0.032Yes
K7401738.6 ± 2.411.0 ± 0.10.030 ± 0.0035 ± 10.197 ± 0.008Yes
K74017430 ± 19BlankBlank281 ± 330.196 ± 0.084 
K740176Blank11.2 ± 1.4Blank<71b0.166 ± 0.017Yes
K740180Blank10.2 ± 1.60.027 ± 0.0149 ± 40.190 ± 0.017Yes
K7401813.4 ± 2.310.7 ± 0.70.022 ± 0.00837 ± 140.197 ± 0.010Yes
K740182BlankBlankBlank106 ± 620.105 ± 0.081Yes
Figure 4.

4He content against 3He/4He ratio in spherules and unmelted micrometeorites (Osawa et al. 2003b). Errors are one sigma. Dotted line shows the bulk solar wind (SW) composition determined by the GENESIS mission (4.53 × 10−4; Heber et al. 2008). AMMs = Antarctic micrometeorites.

The spherule M240410 has an extraordinarily high 3He/4He ratio (9.7 ± 1.1 × 10−3) that resulted from cosmogenic production of 3He. High isotopic ratios have not been found in unmelted micrometeorites and spherules to date, indicating that this specific spherule may have an exceptional history. The highest 3He/4He ratio reported to date in a micrometeorite is 1.843 ± 0.050 × 10−3 (Stuart et al. 1999), which is much lower than that of M240410.

Figure 5 gives the Ne and He isotopic compositions of cosmic spherules. The Ne isotopic compositions are grouped between the implantation-fractionated SW (IFSW) (Grimberg et al. 2006) and air.

Figure 5.

 a) Three isotope plot of Ne and b) relationship between 3He/4He and 21Ne/22Ne ratio. Errors are one sigma. Solar wind (SW) and implantation-fractionated solar wind (IFSW) data are from Heber et al. (2008) and Benkert et al. (1993), respectively. The composition of galactic cosmic-ray (GCR)-produced Ne is calculated using the chemical composition of M240140 and elemental production rate produced by Leya et al. (2000).

The presence of solar-derived He and Ne in spherules suggests that the spherules have been exposed to SW and/or solar flares before atmospheric entry and they are not simple atmospheric entry ablation fragments of meteorites. Most spherules with preserved detectable amounts of light noble gases were once small particles in interplanetary space during at least part of their exposure to energetic particles.

Most unmelted micrometeorites and spherules contain undetectable amounts of cosmogenic Ne that are either negligible or indistinguishable from the trapped solar component, indicating a short CRE age due to the Poynting–Robertson drag effect. In previous works, a few micrometeorites with high concentrations of cosmogenic Ne have been reported (Osawa and Nagao 2002). In this work, two of the spherules (M240410 and K740126) have high 21Ne/22Ne ratios, reflecting cosmogenic Ne. M240410 in particular has the extraordinarily high 21Ne/22Ne ratio of 0.2031 ± 0.0062, a ratio higher than the previous high value of 0.183 ± 0.018 (Olinger et al. 1990). This ratio reflects the presence of spallation-produced Ne that has been preserved in the particle. The high 21Ne/22Ne ratio of this particle is consistent with the presence of cosmogenic 3He (Fig. 5b).


In a previous work, about 40% of the Tottuki spherules were identified as being extraterrestrial by virtue of their 40Ar/36Ar ratio being lower than the terrestrial value of 296 (Osawa et al. 2003a). In this study, extraterrestrial Ar is found in 35 of the 99 spherules, a greater ratio than for both He and Ne. There are three main reasons why Ar is a sensitive indicator of extraterrestrial origin. First, the sensitivity of our mass spectrometer to Ar is higher than He and Ne. Second, no serious mass interference exists for 36Ar, 38Ar, and 40Ar in the mass spectrometer at a mass resolution of 600, in contrast to Ne (Osawa 2004). Third, the shift of the 40Ar/36Ar ratio is very large because most 40Ar is produced from the electron-capture radioactive decay of 40K. A 40Ar/36Ar ratio lower than that of terrestrial air implies an extraterrestrial origin because generally there are no terrestrial materials with a 40Ar/36Ar ratio lower than 296.

Figure 6a shows the 40Ar/36Ar ratio plotted against the 36Ar content. The spherules can be roughly divided into two groups: those with relatively high 40Ar/36Ar ratios but low 36Ar content and those with relatively low 40Ar/36Ar ratios but high 36Ar content. A lot of the spherules belonging to the former group have 40Ar/36Ar ratios close to the atmospheric value (296), suggesting that they are either terrestrial or that the original Ar in the particles has been lost and subsequently contaminated by atmospheric Ar. Although the chemical compositions and morphologic nature of the spherules selected in this study overlap those of the previously studied spherules from the SPWW collection (Taylor et al. 2000) and are not similar to those of terrestrial materials such as spheritic volcanic glass and microtectites (e.g., Glass 2002), the sample’s Ar isotopic composition does not provide any extraterrestrial evidence. The Ar in the latter group is dominated by extraterrestrial components such as SW and/or primordial trapped components.

Figure 6.

 a) 36Ar content against 40Ar/36Ar ratio for two spherule series and b) relationship between 38Ar/36Ar and 21Ne/22Ne ratio. Errors are one sigma. Tottuki spherules are from Osawa et al. (2003a). Dotted line shows the isotopic ratio of the terrestrial atmosphere of 296. The isotopic composition of solar wind (SW) is from Heber et al. (2008). M240410 has both the signs of cosmogenic 21Ne and 38Ar.

Osawa et al. (2003a) reported an enigmatic spherule, namely To440080, with a high 40Ar/36Ar ratio of 543.7 ± 13.2, presumably reflecting a significant contribution of radiogenic 40Ar. Particles similar to this were not discovered in this work. Although K740148 and M240264 have relatively high 40Ar/36Ar ratios of 373 ± 67 and 380 ± 63, respectively, the values are not comparable to those of To440080. If the two sigma error is taken into consideration, the Ar isotopic ratios of K740148 and M240264 are indistinguishable from the atmospheric value.

Figure 6b shows the relationship between 21Ne/22Ne and 38Ar/36Ar. M240410 has a singularly high 38Ar/36Ar ratio of 0.256 ± 0.005, indicating a large amount of cosmogenic 38Ar. Such a high 38Ar/36Ar ratio has not been found in micrometeorites to date. Cosmogenic 38Ar could not even be identified in the unmelted micrometeorites with high concentrations of cosmogenic 21Ne (Osawa and Nagao 2002). The cosmogenic 38Ar concentration in M240410 is 4.8 ± 0.5 × 10−8 cm3 STP g−1, much higher than that found in typical ordinary chondrites and carbonaceous chondrites.

Extraterrestrial Noble Gases of Spherules

Although the interior textures, the morphologic natures, and the chemical compositions of all the samples attest to their extraterrestrial origin, the noble gas data further attest to their extraterrestrial origin. The criteria we have used to establish whether any given sample contains extraterrestrial noble gases are based on three lines of evidence: (1) the 40Ar/36Ar ratio is lower (by two sigma) than that of the terrestrial atmosphere (296); (2) the 3He/4He is higher than that of the terrestrial atmosphere (1.4 × 10−6) by more than two sigma; and (3) the 20Ne/22Ne ratio is higher than that of the terrestrial atmosphere (9.8) by more than two sigma error.


Cosmic-Ray Exposure Age of Spherules

Most unmelted micrometeorites have CRE ages shorter than 5 Myr so the cosmogenic noble gases cannot be distinguished from the large trapped components (e.g., Osawa and Nagao 2002). However, spallogenic He, Ne, and Ar is definitely in some of the spherules. The cosmogenic 3He, 21Ne, and 38Ar (3Hecos, 21Necos, and 38Arcos, respectively) concentrations and CRE ages of two spherules are shown in Table 6: K740126 only has cosmogenic Ne. The CRE age is dependent on the size of the objects when they were irradiated. We calculate CREs using three models: a 4π exposure on small bodies, a 2π exposure on larger bodies, and exposure to energetic particles as dust falling into the Sun as a result of the Poynting–Robertson effect. For a 4π exposure, we use the cosmogenic 3He and 21Ne and galactic cosmic ray (GCR) production rates from Leya et al. (2000). We assume a 3–5 cm shielding depth within a 5 cm radii meteor. In this model, we assume that the production of cosmogenic nuclides from solar cosmic rays (SCR) is negligible. Although this model for the spherule irradiation is a possibility, the calculation of the CRE age is nonetheless problematic for several reasons. The spherule could have been exposed as a smaller entity before atmospheric entry and we have ignored the production of cosmogenic nuclides by SCR. As these production rates are larger in small bodies, the ages we calculate are likely lower limits. For a 2π exposure, we use the 21Necos and 38Arcos production rates of Hohenberg et al. (1978) and Leya et al. (2001). In this model we obtain a regolith exposure duration assuming that the particles were from grains near the surface of their parent bodies. In the third model, we assume that spallogenic 21Ne is produced by both GCR and SCR protons in which the production rate caused by SCR varies with the inverse-square of the heliocentric distance (Osawa and Nagao 2002). The CRE age calculated in this model assumes exposure as a small particle under the effect of SCR and Poynting–Robertson drag. However, the method can only be used for cosmogenic 21Ne because there are no applicable data for the production rates caused by SCR for 3He and 38Ar. The CREs for each of these three models are shown in Table 6.

Table 6.   Cosmic-ray exposure ages of spherules.
SampleWeight (μg)Size (μm × μm)3He/4He (10−4)21Ne/22Ne38Ar/36Ar3Hecos (10−9 cm3 STP g−1)21Necos (10−9 cm3 STP g−1)38Arcos (10−9 cm3 STP g−1)aCosmic-ray exposure ages (Myr)
4π exposureb2π exposurecGCR + SCRd
  1. Note: GCR = galactic cosmic ray; SCR = solar cosmic ray; n.d. = not detected.

  2. aSomatosensory-evoked potential component (38Ar/36Ar = 0.205) is used as an endmember.

  3. bProduction rates of 4π exposure geometry by GCR are from Leya et al. (2000); 3–5 cm shielding depth of 5 cm radii meteorite is assumed.

  4. cProduction rates of 2π exposure geometry by GCR are from Hohenberg et al. (1978) and Leya et al. (2001); 0 cm shielding depth is assumed.

  5. dCalculated by Model 1 of Osawa and Nagao (2002).

  6. eHeliocentric distance within the regions where spherules are generated.

M2404101.9 ± 0.1120 × 10897 ± 110.203 ± 0.0060.256 ± 0.005279 ± 35326 ± 2148 ± 621 ± 3141.2 ± 11176.3 ± 11382.1 ± 48393 ± 2752 ± 3
K7401262.4 ± 0.2121 × 1204.7 ± 0.50.051 ± 0.0070.170 ± 0.055n.d.1.8 ± 0.6n.d.n.d.1.0 ± 0.41.3 ± 0.4n.d.0.27 + 0.170.27 − 0.121.7 + 0.41.7 − 0.3

Possible Source of M240410

M240410 has exceptionally high 3He/4He, 21Ne/22Ne, and 38Ar/36Ar ratios and the highest concentrations of cosmogenic 3He, 21Ne, and 38Ar. A remarkable point is that the concentration of 21Ne of the spherule (3.26 × 10−7 cm3 STP g−1) is 2.5 times higher than that of the unmelted micrometeorite having the highest 21Necos concentration (1.3 × 10−7 cm3 STP g−1; Osawa and Nagao 2002). This particle is the only sample in which the presence of cosmogenic 38Ar is identified. The presence of the three cosmogenic nuclides clearly revealed an exceptionally long exposure age, independent of the exposure scenario. It is noteworthy that the Ne isotopic composition of the particle lies exactly on the IFSW-GCR line in a three isotope diagram (Fig. 5a) and its 40Ar/36Ar ratio is very low (2 ± 1). The Ne isotopic composition indicates that the spherule contains solar-derived Ne, which is universally observable in micrometeorites (e.g., Osawa and Nagao 2002). The concentration of 20Ne of 1.6 × 10−5 cm3 STP g−1 is difficult to explain without the contribution of solar-derived Ne. The Ne isotopic composition of M240410 can be simply interpreted as being a mix of the average composition of unmelted micrometeorites and cosmogenic nuclides. Judging from the noble gas composition, the particle is not from a typical meteorite fragmented on Earth or in the atmosphere.

A high 21Ne/22Ne ratio is also observed in K740126. Using the model of Osawa and Nagao (2002), K740126 has a CRE age of 0.27 Myr and a heliocentric parent body distance of 1.7 AU. A possible parent body of K740126 is a short-period comet or an asteroid in the main belt, both of which are thought to be the main sources of micrometeorites. M240410 has the very long CRE age of 393 Myr and a long heliocentric parent body distance (52 AU), assuming the model of Osawa and Nagao (2002). This result suggests that the parent body of M240410 is an Edgeworth-Kuiper Belt Object (EKO). Moreover, the spherule might have come from the parent body with a longer heliocentric distance if the majority of cosmogenic nuclides had been lost during atmospheric entry.

Dust production due to the impact of interstellar dust on EKOs is a significant source of interplanetary dust grains (Yamamoto and Mukai 1998). After leaving the EKOs the orbits of the dust grains evolve under the complex influence of the gravitational forces of the giant planets and the Sun, and Poynting–Robertson drag forces. According to the result of numerical model calculations made by Liou et al. (1996), grains with larger diameters than about 9 μm can be destroyed upon colliding with debris and through impacting interstellar dust before reaching the inner solar system, although the possibility remains of grains of dust larger than 50 μm in the Edgeworth-Kuiper Belt entering the inner solar system. If all of the cosmogenic noble gases are due to 4π exposure and its parent body is really an EKO, this particle might have evaded the risk of multiple collisions with interstellar dust. In the dynamical model calculation proposed by Moro-Martín and Malhotra (2003), the mass influx of Edgeworth-Kuiper Belt dust accreted on Earth is 4.1 × 105 kg yr−1 (0.8–150 μm, ρ = 2.7 g cm−3), which corresponds with the 2.9% of total mass influx of 1.4 × 107 kg yr−1 estimated by Yada et al. (2004) or the 1.0% of directly measured mass accretion rate of 4 × 10kg yr−1 (Love and Brownlee 1993). This then might make it reasonable to assume that M240410 arrived from an Edgeworth-Kuiper Belt.

If the spherule is really dust from an Edgeworth-Kuiper Belt, its chemical and mineralogical composition would include important information on a relatively unknown body. The porphyritic particle has an average chemical composition (Fig. 2e) and no prominent characteristics in its texture and shape (Figs. 1u and 1v). The result may imply that there is no significant difference in chemical composition between the rocky material in the EKOs and asteroids located in the main belt.

Another interpretation for the long CRE age is that a large amount of cosmogenic noble gases were produced near the surface of the parent body before the particle broke free. The regolith exposure could be as long as 176 Myr, as estimated from the production rate of 21Necos of 2π exposure geometry caused by GCR (Leya et al. 2001). The concentration of cosmogenic 21Ne of 3.26 × 10−7 cm3 STP g−1 is comparable to the highest values of the solar-flare-irradiated grains separated from Murchison and Murray (CM2) chondrites (3.568 × 10−7 cm3 STP g−1 for Murchison and 4.212 × 10−7 cm3 STP g−1 for Murray), with the longest precompaction exposure of the grains being 143.64 Myr (Hohenberg et al. 1990). However, this type of cosmogenic 21Ne-rich grain is rare. All the irradiated grains in Cold Bokkeveld (CM2) have 21Necos concentrations lower than 1 × 10−7 cm3 STP g−1 (Hohenberg et al. 1990). Caffee et al. (1987) measured the noble gases in solar-flare-irradiated grains in Murchison and three gas-rich meteorites, with the longest precompaction exposure age of 53 Myr being found for Kapoeta (HOW), 33 Myr for Fayetteville (H4), 19 Myr for Weston (H4), and 27 Myr for Murchison (CM2). These values are much shorter than that of M240410. Nakamura et al. (1999) analyzed CM chondrites by laser microprobe, and the results indicated that most parts of clastic matrix and all primary accretionary rocks in CM chondrites experienced a very short precompaction irradiation of <0.1 Myr, and also that cosmogenic Ne had enriched some of the minor portions. The precompaction irradiation age of the cosmogenic Ne rich portions of CM chondrites is 10 Myr at most, and the highest concentration of 21Necos they found is lower than 1/20 that of M240410. Precompaction exposure is often seen in gas-rich meteorites, which are mainly made up of ordinary chondrites, particularly H (Osawa and Nagao 2006). However, M240410 obviously differs from ordinary chondrites. The very low 40Ar/36Ar ratio of 2 ± 1 is distinguishable from gas-rich ordinary chondrites that generally have a 40Ar/36Ar ratio higher than 100 (e.g., Osawa and Nagao 2006).

It is interesting that the chondritic chemical composition and noble gas signature of the spherule are very similar to those of ureilites. If the contribution of solar-derived and cosmogenic noble gases are excluded, the concentrations of 40Ar, 84Kr, and 132Xe, their relative abundances, and very low 40Ar/36Ar ratios, resemble those of ureilites (e.g., Okazaki et al. 2003) rather than other types of meteorites. However, long precompaction exposure has not been seen in ureilites; the highest concentration of 21Ne found in ureilite is 1.63 × 10−7 cm3 STP g−1 from Kenna (Okazaki et al. 2003), which is half that of the 21Necos of M240410.

It should be noted that the estimated surface exposure age of 176 Myr for the spherule is a lower limit because the noble gases had been partially degassed through heating. In addition, the extremely long exposure age of 382 Myr is estimated from the 38Arcos using a production rate of 2π exposure geometry caused by GCR (Hohenberg et al. 1978; Leya et al. 2001). One of the reasons for the large discrepancy between the 21Necos and 38Arcos exposure ages may have been the Fe being depleted during atmospheric entry as it is the main source of cosmogenic 38Ar. Another reason, perhaps the primary reason, is the light noble gases being preferentially depleted. In any case, it is clear that the exposure age of the spherule is much longer than the precompaction age found in the parent bodies of major meteorites. If the precursor of the spherule had spent several million years near the surface of its parent body, the particle can be considered to have been generated from a very rare region of an asteroid where cosmogenic noble gases are extremely rich or from a mature regolith.


The internal texture, chemical composition, and noble gas composition of 99 Antarctic spherules collected from melting blue ice and filtered water at three sampling sites from Kuwagata Nunatak and Minami-Yamato Nunataks in the Yamato Meteorite Ice Field were investigated. Their textures were used to classify them into six different types. The notable findings obtained in this study are listed below.

  • 1 The effects of severe heating during atmospheric entry can be observed in the systematic Fe loss and the low content of noble gases as well as the interior texture of the spherules. At least 95% of the light noble gases had been released through severe heating. The correlation between the noble gas content and the type of spherule also revealed the sign of heating. Mg depletion caused by aqueous alteration in glacier ice was not identifiable despite the long storage duration.
  • 2 As in the case of unmelted micrometeorites, the detectable He and Ne in spherules are dominated by solar-derived components, implying that the spherules are small particles in interplanetary space and not fragments of meteorites fallen to Earth as solar-gas-rich meteorites are quite rare.
  • 3 Cosmic-ray-produced noble gas nuclides are identified in some spherules and the CRE ages calculated. One of the samples, namely M240410, has extraordinarily high concentrations of cosmogenic 21Ne and 38Ar, with the extremely long CRE age of 393 Myr being estimated for the particle. The origin of the spherule may have been an EKO, but neither its chemical composition nor its texture was anomalous. Another plausible reason for the high concentration of cosmogenic nuclides was long surface exposure on its parent body. The calculated exposure age with 38Arcos was 382 Myr in 2π exposure geometry. However, such a long precompaction exposure is yet to be found in known meteorite parent bodies.


Acknowledgments— We would like to thank NIPR for providing the Antarctic micrometeorite samples. We also thank an anonymous referee, S. Taylor, and M. Caffee for their constructive suggestions.

Editorial Handling—

Dr. Marc Caffee


Table A1. Chemical compositions (in wt%) of cosmic spherules.

  1. Note: RGB = relict grain bearing (containing relict crystals that can be identified on backscattered electron images); n.d. = not detected; n.m. = not measured. Chemical compositions of spherules are 100% normalized.