Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/arxiv-cornell-university/cosmic-history-and-a-candidate-parent-asteroid-for-the-quasicrystal-In4UzPMfQj
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- ISSN
- 0012-821X
- eISSN
- ARCH-3330
- DOI
- 10.1016/j.epsl.2018.03.025
- Publisher site
- See Article on Publisher Site
Abstract
The unique CV-type meteorite Khatyrka is the only natural sample in which “quasicrystals” and associated crystalline Cu,Al-alloys, including khatyrkite and cupalite, have been found. They are suspected to have formed in the early Solar System. To better understand the origin of these ex- otic phases, and the relationship of Khatyrka to other CV chondrites, we have measured He and Ne in six individual, ~40-μm-sized olivine grains from Khatyrka. We find a cosmic-ray exposure age of about 2-4 Ma (if the meteoroid was <3 m in diameter, more if it was larger). The U,Th-He ages of the olivine grains suggest that Khatyrka experienced a relatively recent (<600 Ma) shock event, which created pressure and temperature conditions sufficient to form both the quasicrystals and the high-pressure phases found in the meteorite. We propose that the parent body of Khatyrka is the large K-type asteroid 89 Julia, based on its peculiar, but matching reflectance spectrum, evidence for an impact/shock event within the last few 100 Ma (which formed the Julia family), and its location close to strong orbital resonances, so that the Khatyrka meteoroid could plausibly have reached Earth within its rather short cosmic-ray exposure age. crystals (e.g., five-fold). First proposed by 1. INTRODUCTION Levine and Steinhardt (1984), they were first The motivation for this noble gas study is the synthesized and identified in the laboratory by curiosity and fascination with the origin of an Shechtman et al. (1984). The ensuing multi- exotic type of material: quasicrystals. Short for decade search for natural quasicrystals was quasi-periodic crystals, they are materials show- eventually successful when a powder from a ing a quasi-periodic arrangement of atoms, and millimeter-sized rock sample (later called the rotational symmetries forbidden to ordinary “Florence Sample”) from the collection of the Page 1 of 27 Museum of Natural History in Florence, Italy, al. (2014) found high-pressure mineral phases displayed a diffraction pattern with a five-fold (including stishovite and ahrensite) indicating symmetry (Bindi et al., 2009). This first natural Khatyrka was at one point exposed to pressures quasicrystal, with a composition Al Cu Fe , >5 Gigapascals (GPa) and temperatures >1200 63 24 13 was named icosahedrite. In the host rock of the °C, followed by rapid cooling. Synthetic icosa- Florence sample, silicates and oxides are par- hedrite remains stable under these conditions tially inter-grown with exotic copper-alu- (Stagno et al., 2015). These conditions are also minium-alloys (khatyrkite, CuAl , and cupalite, typical for asteroid collisions (Stoeffler et al., CuAl) which contain the quasicrystals. The oxy- 1991). In dynamic shock experiments, Asimow gen isotopic compositions of silicates and ox- et al. (2016) successfully synthesized icosahe- ides in the Florence Sample suggest an extrater- dral phases with a composition close to that of restrial origin (Bindi et al., 2012), as they match icosahedrite, and Hamann et al. (2016) reported the ones found in some carbonaceous chon- the shock-induced formation of khatyrkite drites. The provenance of the Florence Sample (CuAl ), mixing of target and projectile mate- was eventually traced back to a Soviet prospect- rial, and localized melting along grain bound- ing expedition to the Far East of Russia in 1979 aries, resembling the assemblages found in (Bindi and Steinhardt, 2014). To find more of Khatyrka. Lin et al. (2017) argued that a two- the exotic material, an expedition was launched stage formation model is needed to explain the to the Koryak mountain range in the Chukotka mineralogical and petrographic features ob- Autonomous Region in 2011. Eight millimeter- served in Khatyrka: first, an event as early as sized fragments of extraterrestrial origin were 4.564 Ga in which quasicrystals with icosa- found by panning Holocene (>7 ka old) river hedrite composition formed; and, second, a deposits (MacPherson et al., 2013). Again, more recent impact-induced shock that led to Cu,Al-alloys were found attached to, or inter- the formation of a second generation of qua- grown with, extraterrestrial silicates and oxides. sicrystals, with a different composition. Ivanova The oxygen isotopes, chemistry and petrology et al. (2017) suggested the possibility of an an- of the eight new fragments all suggest a CV thropogenic origin of the Cu,Al-alloys from ox (carbonaceous chondrite, Vigarano-type, oxi- mining operations, but this explanation is in- dized subtype) classification for the meteorite, compatible with the chemistry and the thermo- now named Khatyrka (Ruzicka et al., 2014). dynamic conditions needed to explain all obser- The achondritic, diopside-hedenbergite-rich and vations listed above (see also MacPherson et al., Cu,Al-alloy-encrusted Florence Sample sug- 2016). gests that Khatyrka is an unusual chondritic In this work, we aim to better understand the breccia which accreted both achondritic and ex- origin and history of the quasicrystal-bearing otic materials (MacPherson et al., 2013). Khatyrka meteorite and its relationship to other So far, no other meteorite is known to contain meteorites, in particular other CV chondrites. Cu,Al-alloys or quasicrystals. The lack of We do this by measuring the helium and neon 26 24 Mg/ Mg anomalies suggests that the Cu,Al-al- (He, Ne) content of individual Khatyrka olivine loys formed after the primordial Al (t = 0.7 grains, in order to determine their cosmic-ray 1/2 million years, Ma) had decayed away, i.e., at exposure and radiogenic gas retention (based on least ~3 Ma after formation of the oldest con- uranium-thorium-helium = U,Th-He) ages. Due densates in the Solar System 4.567 billion years to the extremely small mass available from (Ga) ago (MacPherson et al., 2013). Hollister et Khatyrka (total mass <0.1 g), destructive analy- Page 2 of 27 ses have to be reduced to the minimum. We in in Fiji/ImageJ (Schindelin et al., 2012), which worked with single grains of ca. 40 μm diame- finds 3D-connected objects with a voxel bright- ter, which were all part of a chondrule from ness above a user-defined 8-bit (256) gray-scale Khatyrka fragment #126 (recovered in the 2011 value (called the “threshold”). We first searched expedition). Because of the very small gas a range of thresholds where the grains are well- amounts expected, we used the high-sensitivity resolved from the background, then searched for “compressor-source” noble gas mass spectrome- the sequence of four thresholds for which the ter at ETH Zurich (Baur, 1999). This unique in- volume change between adjacent gray-scale strument has previously been used in a similar steps was minimal. We then corrected for sur- way to measure He and Ne in individual mineral face resolution effects (i.e., for voxels close to grains returned from asteroid Itokawa (Meier et the actual grain which are only partially in-filled al., 2014). by grain material, and thus have a reduced voxel brightness: they fall below the threshold al- 2. METHODS though they contain a sub-voxel-sized part of the grain) by interpolating between the average 2.1. Volumes from X-ray tomography brightness of unambiguous (internal) grain vox- To determine cosmic-ray exposure and radio- els and background voxels (e.g., if the grain ma- genic gas retention ages, we need noble gas con- terial brightness is 120 and the background centrations (cm STP/g), which requires the de- brightness is 40, a surface-near voxel with termination of the masses of the individual brightness 80 is assumed to be half in-filled grains. This is not possible to do both safely and with grain material). Grain #126-06 was too reliably on a micro-balance for such small (<1 small, and too fragmented to be of use for this µg) grains. To determine their masses, we first study, and was thus not further analyzed. The measure their volumes using high-resolution X- grain volume uncertainties in Table 1 corre- ray tomography (nano-CT), and then multiply spond to the range in volume within the se- these volumes with the densities calculated from quence of four thresholds with minimal volume their mineralogical composition (determined changes (after correction for surface effects). with SEM-EDS) and textbook mineral densities. Grain #126-05 fragmented after nano-CT scan- At the University of Florence, seven olivine ning (as visible from the SEM images, see sec- grains from Khatyrka fragment #126 (named tion 2.2.). Only two of its fragments could be #126-01 through -07, or 1 through 7 for short in transferred to the sample holder for noble gas the figures), all fragments from a chondrule analysis. Their volume was estimated from their (judging from their size and chemistry), were cross-sectional area in the SEM images and an transferred with a micro-manipulator needle empirical relationship between cross-sectional from a TEM grid to a carbon tape. The samples area (A) and volume (V) established by the were then imaged (on the tape) with a GE man- other five grains, A = (0.080±0.009) × V (R = ufactured v|tome|x s micro/nano-CT scanner lo- 0.998), which resulted in a somewhat larger cated at the UChicago PaleoCT facility, using mass uncertainty for these fragments. The vol- the 180kV nano-CT tube, an acceleration volt- ume of grain #126-05 given in Table 1 repre- age of 80 kV, beam current of 70 μA with 500 sents the sum of the two fragments. ms integration time, and no filter. The voxels of the scan are isometric with a size of 5.473 μm. 2.2. Chemical composition, grain masses Grain volumes were then determined from the After nano-CT scanning, the bulk elemental CT images using the “3D-object counter” plug- composition of the grains was qualitatively de- Page 3 of 27 termined by spot analysis on unpolished sur- 202ND/MMN-1 micro-manipulator attached to faces using an Oxford Instruments XMax-50 en- a Nikon Eclipse Microscope at SEAES, Univer- ergy dispersive X-ray (EDX) spectrometer, sity of Manchester (an image of a grain in trans- mounted on a Zeiss Evo 60 Scanning Electron fer is shown in Fig. S2 in the Supplementary Microscope (SEM; acceleration voltage 20 kV, Material). electron current 400 pA) at the Field Museum of Natural History in Chicago. Back-Scattered 2.3. Light noble gas analysis Electron (BSE) images of the grains are shown At the noble gas laboratory of ETH Zurich, the in Fig. S1 in the Supplementary Material. We sample holder with the grains was loaded into a found an average composition of 95±5% sample chamber which remains connected to an forsterite (Mg SiO ; 3.28 g/cm ) and 5±5% fay- 2 4 ultra-low blank extraction line and the “com- alite (Fe SiO ; 4.39 g/cm ), corresponding to a 2 4 pressor-source” noble gas mass spectrometer grain density of 3.33±0.06 g/cm (Table 1). (Baur, 1999) at all times (i.e., no valves are op- Since the individual measurements were done in erated during a measurement run). Noble gases spot analysis mode and are not resolved from were extracted by laser heating using a Nd:YAG their average, we only use the average density to laser (λ=1064 nm), focused on each grain for 60 calculate the grain masses in Table 1. The six seconds, melting and, in some cases, completely grains (or grain fragments) were then transferred vaporizing the grains. Each extraction was to a customized Al sample holder for noble gas monitored using a camera, which showed the in- analysis, using a hydraulic Narishige MMO- tense flickering / glowing of the grains during laser heating. In two cases (grains #126-02 and -05), some residual light emission was present after 60 seconds, suggesting incom- plete degassing. These grains were thus extracted again for an- other 60 seconds in a follow-up analysis run, and the cumulative gases released by each grain were determined in post-processing. After extraction, the released gases were passed through the ex- traction line which features two getters and three cold traps (two of them containing activated charcoal) cooled with liquid ni- trogen. The strong increase in sensitivity provided by the com- pressor source also leads to a very 3 21 Fig. 1: cosmogenic He and Ne in Khatyrka olivine grains. The measured, fast removal of Ar, Kr, and Xe 3 21 purely cosmogenic He (red symbols), measured Ne (orange symbols) and due to ion pumping, and hence cosmogenic Ne (blue symbols) gas amounts (units of 10-15 cm3STP) are only He and Ne can be analyzed. plotted against grain mass (all errors are 1σ). Both linear fits (lines of con - Isotopic ratios and elemental stant concentration) are forced through the origin, and are shown with their 3 4 20 abundances of He, He, Ne, 2σ confidence intervals (dashed lines). Page 4 of 27 21 22 Ne, and Ne were then measured (together termine the possible origins of the Khatyrka me- 16 40 with HD, H O, Ar, and CO for interference teoroid in the asteroid belt was presented in 2 2 corrections, which proved to be negligible) fol- Meier et al. (2017), we will only give the model lowing a protocol developed by Heck et al. parameters here. To determine the migration (2007). In this work, we used a new, slightly range of the Khatyrka meteoroid, we evolved modified data fitting routine which is only ap- the semi-major axes of test particles placed plicable to very low gas amounts, but results in 0.004 Astronomical Units (au) in- and outside smaller uncertainties in the measured gas the ν , 3:1, and 5:2 resonances backwards, using amounts and, in some cases, also to somewhat the formulas given by Vokrouhlicky et al. lower detection limits (a detailed discussion is (2015) for seasonal and diurnal Yarkovsky drift, given in the Supplementary Material). A series and Farinella et al. (1998) for occasional resets of blank runs – indentical to the sample runs ex- of rotation rates and obliquity by collisions. Test cept that no laser is fired – was measured to de- particles were given five different sizes: 0.5, termine the extent their scattering around their 1.0, 1.5, 3.0, and 5.0 m, a compressed density of average. Two standard deviations of that scatter 3.3 g/cm and a porosity of 22% (corresponding added to the average blank define the detection to bulk CV chondrites; the resulting bulk den- - 3 limits for all isotopes, which are (in units of 10 sity used in the model is 2.6 g/cm ), an albedo 15 3 19 -1 -1 cm³STP; 1 cm STP = 2.687×10 atoms) 0.55, of 0.15, a heat capacity of 500 J kg K , and a 3 4 20 21 -1 310, 38, 0.47, and 3.9 for He, He, Ne, Ne, thermal conductivity of either 0.15 or 1.5 W m 22 -1 and Ne, respectively. The average blanks rep- K . Orbital evolution was continued until the resent 8%, 19%, 21%, 34%, and 36% of the de- modeled meteoroid had accumulated the mea- tection limits for these isotopes, respectively. A sured cosmic-ray-produced inventory of noble calibration bottle containing both air-like He, gases (the time for this depends on its radius, and near-atmospheric Ne (Heber et al., 2009) see section 3.2). Collisional disruptions were was used to determine instrumental sensitivity neglected since the collisional life-time for all 3 4 20 for He, He, and Ne, and the instrumental Khatyrka meteoroids considered here is about mass fractionation for Ne isotopes five times longer than their cosmic-ray exposure (0.17±0.06%/amu, based on the measured age (Farinella et al., 1998). Gravitational inter- 20 22 Ne/ Ne ratio). A significant excess signal (ca. actions with planets (apart from resonances) and 8%) on mass 21 was found in calibration runs, other asteroids are neglected. The resonances which was identified as interference from are treated in a statistical way, with test particles 20 20 20 NeH. The NeH/ Ne ratio needed to explain within 0.004 au of resonances removed with a the interference in the calibrations runs is probability scaled to reproduce the resonance- 0.24±0.01‰. The H pressure in the spectrome- specific life-times (Gladman, 1997). All test ter, monitored via measurement of HD, stayed particles ejected by resonances were removed constant over the course of the measurement from the simulation, and replacement test parti- run, implying that the NeH interference re- cles were started until 10 individual migration mained roughly constant. The interference was histories had been accumulated for each reso- subsequently subtracted from mass 21 in the nance and meteoroid size. sample runs. That correction was never more than 3% of the total signal on mass 21. 3. RESULTS 3.1. Noble gas components 2.4. Orbital evolution model During laser melting, all six Khatyrka olivine As the orbital evolution model used here to de- Page 5 of 27 3 21 3 4 grains released He and Ne well above the de- genic”) component ( He/ He ~0.2; e.g. Leya and tection limit (DL; see Table 1, Fig. 1). Also Masarik, 2009). Fig. 1 shows that for He, all six above DL were: He in three grains (#126-02, grains plot within 2σ of their mass uncertainty -04 and -05), Ne in three grains (#126-02, -03, (x-coordinate) on a linear fit forced through the and -05), and Ne in four grains (#126-02, -03, origin. The slope of the line corresponds to an 4 3 3 -04, -05). For the three grains for which He was average cosmogenic He ( He ) concentration cos 3 4 -17 3 -8 3 below DL, a lower limit for the He/ He ratio of 5.5±0.2×10 cm STP/ng (or 10 cm STP/g; was calculated from the ratio of the measured R = 0.993). For the other important cosmogenic 3 4 3 4 21 21 He and the He DL. These He/ He ratios (and isotope, Ne ( Ne ), the scatter around the fit cos -17 3 2 lower limits) are all a factor of at least ~10 line (1.3±0.2×10 cm STP/ng; R = 0.881) is 3 4 higher than the He/ He ratio in the solar wind more pronounced. To calculate the cosmogenic 21 21 or other trapped components, and at least a fac- fraction of Ne ( Ne ), we use a two-compo- cos 3 4 tor of ~1000 higher than the He/ He ratio in the nent deconvolution with a non-cosmogenic 3 21 22 Earth’s atmosphere. This suggests that the He ( Ne/ Ne = 0.03; e.g. Ott, 2014) and a cosmo- 21 22 in Khatyrka olivine is dominated (>90%) by the genic end-member ( Ne/ Ne = 0.89; e.g. Leya cosmic-ray-induced spallation (or “cosmo- and Masarik, 2009). Cosmogenic Ne contrib- 20 4 3 4 20 4 3 4 Fig. 2: Ne/ He vs. He/ He diagram. Main: The Ne / He and He/ He ratios of the six Khatyrka olivine grains cos 20 21 4 ( Ne calculated from Ne ). For grains #126-01, -03, and 07, He was below DL. Therefore, the position of the data cos cos point on the diagram corresponds to the minimum distance to the radiogenic end-member (i.e., the origin): these data points are represented by diamonds instead of circles. Inset: Same axes, but expanded to include the measured ratios in 20 4 4 20 Ne/ He (for grains #126-02 and -05), or upper and lower limits on that ratio where He, or Ne, was below DL (grains #126-03 and -04, respectively; the triangle symbols point towards the unconstrained direction). Noble gas com - ponents in both diagrams are shown as black diamonds connected by dashed lines: radiogenic He at the origin, Q gases (Busemann et al., 2000), and solar wind (SW; Heber et al., 2009), as well as the cosmogenic end-members for 20 4 low and high shielding (Leya and Masarik, 2009). The terrestrial atmosphere plots outside the diagram at Ne/ He = 3.14, in the same direction as elemental fractionation (loss of He vs. Ne). Page 6 of 27 utes between 67.5 and 99.9% of the total mea- CV chondrite. A “shielding parameter” is sured Ne, i.e., the cosmogenic component is needed to reflect the dependence of the cosmo- dominant also for Ne. The nature of the non- genic production rates on the position within the cosmogenic Ne component (e.g., Earth’s atmos- meteoroid and the size of the meteoroid (to- phere, solar wind, Q gases) in the Khatyrka gether referred to as “shielding”). The most olivines cannot be determined, due to the high commonly used shielding parameter, the 20 22 22 21 uncertainties in the measured Ne/ Ne ratios Ne/ Ne ratio of the cosmogenic end-member, resulting from the low gas amounts (Fig. S3 in cannot be used here since it cannot be reliably the Supplementary Material). The measured ele- determined given the large uncertainties in 20 4 20 22 mental Ne/ He ratios are suggestive of an at- Ne/ Ne and the low number of samples (Fig. mospheric origin (Fig. 2, inset). If only the cos- S3 in Supplementary Materials). Instead, we use 20 21 3 21 mogenic Ne (=0.92 × Ne ) is considered, all the cosmogenic He/ Ne ratio, which – for cos samples plot between cosmogenic and radio- forsterite – yields CRE ages showing a clear de- 20 4 3 4 genic end-members in a Ne/ He vs. He/ He pendence on meteoroid size (Fig. 3). The nomi- 3 21 diagram (Fig. 2, main). This suggests that since nal He/ Ne ratios of the grains vary between (and during) Khatyrka’s exposure to cosmic- ~3 and ~7, encompassing the full range ex- rays, no significant fraction of He was lost, e.g. pected from variable shielding. However, since to impact shocks or solar heating. In summary, all six grains are derived from the same, mm- He and Ne in Khatyrka olivine are dominated sized grain (#126), this variation cannot reflect by the cosmogenic component, with a small variable shielding conditions. It does not reflect 21 3 (<33% for Ne, <10% for He) contribution of partial loss of He on the meteoroid or parent atmospheric (or trapped) gases, and some radio- body either, as this would result in variable He 4 21 genic He. and (nearly) constant Ne concentrations, which is the opposite of what is observed (Fig. 1). In- stead, the most likely explanation is that the Ne 3.2. Cosmic-ray exposure age and mete- variability reflects incomplete degassing of Ne oroid radius during laser heating for some of the grains. In- The determination of a cosmic-ray exposure deed, the two grains with the highest Ne con- cos (CRE) age, i.e., the time a meteorite has been centrations (#126-01 and -04) have compatible exposed to galactic cosmic rays (GCR) as a me- 3 21 He/ Ne ratios (of 2.95±0.35 and 3.45±0.33) teoroid (a meter-sized object in interplanetary and Ne concentrations (1.83±0.22 and cos space), requires two key values: the concentra- -8 3 1.78±0.19 × 10 cm STP/g), suggesting that tion of a cosmogenic isotope (e.g., He or cos only these two grains were completely degassed Ne ), and a productionrate (atoms per units of cos for Ne. The error-weighted average of the mass and time) of that isotope from interactions 3 21 He/ Ne ratio of all grains is 3.50±0.46 (2σ), with target elements. While the former can be 3 21 and the ratio of the sums of all He and Ne re- determined by measurement, the latter depends 3 21 leased is 4.03±0.46 (2σ). The true He/ Ne ratio on the chemical composition of the meteoroid, of Khatyrka is thus most likely in the range of its size, and the position of the sample within ~3-4 (shaded region in Fig. 3). This range re- the meteoroid. The model calculations of Leya quires that the Khatyrka meteoroid had a radius and Masarik (2009) allow us to determine He of at least 30 cm (if the samples are derived and Ne production rates in forsteritic olivine in from the meteoroid’s center), and might have a carbonaceous chondritic matrix. This is impor- been up to 5 m in radius, or even more, as the tant because the production of Ne in forsteritic latter is just the largest meteoroid modeled by olivine is 60-80% higher compared to a bulk Page 7 of 27 3 21 3 Fig. 3: CRE age as a function of cosmogenic He/ Ne, shielding and meteoroid radius. Upper part: the He-based 3 21 CRE age of Khatyrka (Leya and Masarik, 2009) as a function of the cosmogenic He/ Ne ratio, shielding depth (symbol 3 21 color) and meteoroid size (symbol shape). Lower part: the He/ Ne ratios measured in the individual grains (filled, 3 21 numbered circles with 1σ errors). The gray shaded area is the range for the error-weighted average of the He/ Ne ra- 3 21 tio (1σ) for all grains. Reading example: if the true He/ Ne ratio of Khatyrka olivine is 3.5, and the meteoroid radius was 150 cm, then the CRE age is ~3 Ma and the shielding depth ~30 cm (light yellow). Leya and Masarik (2009). As shown in Fig. 3, and thus a CRE age in the range of 2-4 Ma. A the cosmogenic He concentration (5.5±0.2 × measurement of the cosmogenic radionuclide -8 3 3 26 10 10 cm STP/g) and the He production rates activities in Khatyrka (e.g., Al and Be), while 3 21 corresponding to a cosmogenic He/ Ne ratio in instrumentally challenging, would significantly the range ~3-4, result in a likely CRE age range improve both the CRE age and pre-atmospheric of 2-4 Ma for meteoroids up to about 1.5 m in size determination. radius. If the meteoroid was larger than this, and 3 21 the true cosmogenic He/ Ne ratio of Khatyrka 3.3. Radiogenic gas retention ages 4 4 olivine is <3.5, the samples could also derive After subtraction of cosmogenic He ( He / cos from a strongly shielded position deep (>1 m) 3 He = 4.5±1.3; Leya and Masarik, 2009), the cos inside a large meteoroid, and only a lower limit 4 remaining He must be from Earth’s atmosphere of >4 Ma can then be given for the CRE age. (air), and / or from radioactive decay of U and However, such large meteoroids entering the Th. Assuming that trapped air is not fractionated 20 4 Earth’s atmosphere often result in air-bursts, in Ne/ He results in a contribution of <15% of and thus contribute only a disproportionately atmospheric He for grain #126-05 (and even small fraction of the flux of meteorites arriving less for the other grains). Therefore, the major- at the surface (Bland and Artemieva, 2006). 4 ity (>85%) of the non-cosmogenic He is radio- Therefore, we favor a smaller meteoroid size, genic. Diffusion of cosmogenic and radiogenic Page 8 of 27 He in olivine at typical equilibrium tempera- grains would be systematically depleted in tures in the asteroid belt (~170 K) is very low, mesostasis compared to a chondrule. To allevi- so that no diffusion correction is necessary, ate the concern that we might be underestimat- even at Ga-scales (Trull et al., 1991). We can ing our U,Th-He ages, we will calculate them thus determine a simple U,Th-He retention age, for both 14-23 ppb U (the “nominal age”) and 5 i.e., the time it would take to accumulate the ra- ppb U (the “maximum age”). diogenic He concentration from the radioactive The nominal ages of five of the grains are decay of U and Th present in the sample. A re- roughly consistent with each other at <400-700 tention age shorter than the age of the solar sys- 4 (Table 1, Fig. 4). The maximum ages for these tem is indicative of a loss of radiogenic He at five grains are approximately <1.5-1.8 Ga (see some point, e.g., due to impact shock-heating, Fig. S4 in the Supplementary Material). Grain or protracted solar heating. If that He-loss was #126-02, with a higher nominal age of <2.5–3.5 complete, the retention age dates the (end of Ga (and a maximum age in excess of the age of the) He-loss event. If not, the U,Th-He age is an the Solar System), might have had a higher frac- upper limit to the age of the (last) He-loss event. tion of (U-enriched) mesostasis than the other Since it is not possible to determine if the He- grains, and / or was only partially degassed. The loss was complete, we consider every U,Th-He consistent, short U,Th-He ages in all but one age given here to be an upper limit age. For this grain seem to indicate that the parent body of reason, we also do not correct for atmospheric 4 Khatyrka experienced a strong gas-loss event, He, because such a correction (<15%) would with a nominal age of <600 Ma (the youngest not lead to a qualitatively different result. upper limit age in the set, rounded to a single To calculate U,Th-He retention ages, we also digit) and a maximum age of <1.8 Ga. need U and Th concentrations. These were not measured directly, to maximize the mass avail- 4. DISCUSSION able for noble gas analyses, but as mentioned 4.1. An improved chronology for the the size and chemistry of the grains suggest they Khatyrka meteorite are fragments of chondrules. Individual chon- Prior to this work, only two points in the history drules of the CV chondrite Allende have bulk ox of Khatyrka were known: the Cu,Al-alloys U concentrations of 14-23 ppb and a Th/U ratio formed at least 2-3 Ma after the condensation of of ~3.5 (Amelin and Krot, 2007), which is also compatible with the respective values in bulk the first solids in the solar system (i.e., less than CV chondrites (Wasson and Kallemeyn, 1988). ca. 4563 Ma ago; Amelin and Krot, 2007), and fragments of the meteorite were eventually de- However, Kööp and Davis (2012) found that U posited in a river sediment in the Far East of is inhomogeneously distributed within chon- Russia, >7 ka ago (MacPherson et al., 2013). drules (of an LL3 chondrite): most of the U is Now, we have established at least two addi- concentrated in the mesostasis (up to 150 ppb tional events: (1) The time when the Khatyrka U), leaving only about 5±5 ppb U in olivine. Al- though we refer to the grain mineralogy as meteoroid was ejected from its parent asteroid, “olivine” here, we only did SEM-EDS spot most likely 2-4 Ma ago; (2) The age of a strong shock event, which led to the loss of radiogenic analyses of the grains (see section 2.2.), and He, about <600 Ma (and perhaps up to <1.8 thus cannot exclude that mesostasis or other Ga) ago. Since the record of cosmogenic He is minerals were present as minor phases in the undisturbed (see section 3.1., Figs. 1 and 2), the grains. We find it unlikely, however, that all Page 9 of 27 He-loss event must have happened before the start of cosmic-ray exposure (this also excludes that the shock event recorded in Khatyrka is recent, e.g., from anthro- pogenic activity). Stoeffler et al. (1991) show that ra- diogenic gas loss in ordi- nary chondrites starts at shock stage S3 (10 GPa), and is complete only when shock stage S5 is reached (35 GPa). Sharp and de Carli (2006) suggest that the pressures associated with shock stages are likely overestimated by Stoeffler et al. (1991) by a factor of two. Conservatively, the shock event experienced by Fig. 4: Cosmic histories of CV chondrites. Main: U,Th-He retention age vs. ( Ne- Khatyrka must thus have based) CRE age for all CV chondrites with published bulk He, Ne, Ar contents (all reached a pressure of at data points are simple averages of all samples measured for a meteorite; data com - least 5–18 GPa, and proba- piled from Choi et al., 2017; Leya et al., 2013; Schultz and Franke, 2004) , and bly close to the upper end Khatyrka. Meteorites mentioned in the main text are labeled. The U,Th-He ages (calculated here for the range U = 14-23 ppb and Th/U = 3.5) are capped at the of that range given that age of the solar system (long-dashed line). All CRE ages are nominal values (given most Khatyrka grains ana- uncertainty = 20%). For the full data set, see Table S1 in Supplementary Material. lyzed have been thoroughly Inset: CRE age histogram for all CV chondrites, with potential peaks at 2 Ma, 5 Ma, reset. This range is consis- 10 Ma and 23 Ma. tent with the pressure >5 3.3 g/cm , an initial porosity of 22% (com- GPa deduced for Khatyrka (Hollister et al., pressed to zero above 5 GPa), a characteristic 2014). Lin et al. (2017) suggested an S4 shock heat capacity of 1 J/gK, and a pre-shock temper- stage for Khatyrka. According to Stoeffler et al. ature of -100 °C (~170 K), is ~-40 °C for 5 GPa (1991), an S4 shock stage entails a post-shock and ~1500 °C for 18 GPa (Sharp and de Carli, temperature increase of only 250–350 °C, much 2006). An impact shock at the upper end of this lower than the minimum shock temperatures range can therefore explain both the He-loss and recorded in some minerals from Khatyrka shock temperatures >1200 °C. Therefore, the (>1200°C). However, higher post-shock temper- event that degassed the Khatyrka olivine grains atures are achievable if the target material has a might well have been the same as the shock significant porosity (e.g., Schmitt, 2000). Oxi- event that formed the second generation of qua- dized CV chondrites have indeed a relatively sicrystals (Lin et al. 2017). high porosity of 21.8±1.7% (Consolmagno et al., 2008). The shock temperature expected for a carbonaceous chondrite with a grain density of Page 10 of 27 4.2. Is Khatyrka related to other CV chon- 4.3. Candidate parent asteroids for drites? Khatyrka By comparing the CRE and U,Th-He ages of Before being delivered to an Earth-crossing or- Khatyrka with their counterparts in other CV bit, meteoroids spend most of their CRE time chondrites, we can potentially identify mete- (typically a few 10 Ma for stony objects) as orites which experienced a similar cosmic his- small bodies in the asteroid belt. Their orbits tory and might thus be “source-paired” with evolve slowly due to gravitational encounters Khatyrka. This could lead to the discovery of and the effect of non-gravitational forces, most more meteorites with Cu,Al-alloys and qua- notably, the Yarkovsky effect (Vokrouhlicky et sicrystals, or at least the precursor materials al., 2015). Eventually, they reach an orbital res- from which they may have formed. Out of 428 onance which ejects them from the asteroid belt, CV chondrites listed in the Meteoritical Bulletin e.g., the 3:1 resonance at 2.50 au, where the me- Database (as of September 2017), only 30 have teoroid orbits the Sun exactly three times for ev- complete He, Ne, Ar records allowing the deter- ery orbit of Jupiter. In Fig. 5 (top panel), we mination of both Ne-based CRE ages and show the position (inclination vs. semi-major U,Th-He retention ages (Fig. 4; Supplementary axis) of the seven most important resonances in Table S1; data from Choi et al., 2017; Leya et the asteroid belt (Gladman, 1997), together with al., 2013; Schultz and Franke, 2004). There are the positions of the largest asteroids (>30 km), six CV chondrites with CRE ages that are for which non-gravitational forces are negligi- roughly compatible with Khatyrka: Acfer 082, ble, between 2.2 and 3.5 au (all asteroid data are Acfer 086, Acfer 272, Grosnaja, NWA 6746, from http:// asteroid.lowell.edu ). K-type aster- and NWA 10670. They all have higher U,Th-He oids, highlighted in green in Fig. 5, have long retention ages than Khatyrka (even if we con- been associated with the CV, CK and CO type sider only the maximum age of 1.8 Ga), except chondrites, based on similar reflectance spectra for Acfer 082 and Acfer 272. These two mete- in the visible and near-infrared (Bell, 1988; Bur- orites, however, are thought to have lost most of bine et al., 2001; Cloutis et al., 2012). their (cosmogenic and radiogenic) He through The distance the meteoroid migrated in the as- intense solar heating during CRE (Scherer and teroid belt (prior to ejection by a resonance) can Schultz, 2000). In contrast, Khatyrka has re- be estimated from the CRE age, the meteoroid tained its cosmogenic He, as discussed in sec- size, and some reasonable assumptions on mete- tions 3.1 and 3.2. Therefore, the shock age of oroid density and thermal properties, which, in five of the six Khatyrka olivines seems to be combination, determine the strength of the non- unique (at least so far) among CV chondrites, gravitational forces altering the orbit. This mi- even if we use the maximum age of 1.8 Ga. This gration distance can then be used to identify only adds to the “uniqueness” of Khatyrka al- possible parent asteroids for Khatyrka. It should ready suggested by the exotic Cu,Al-alloys, still be considered an approximation, because quasicrystals, and the unusual achondritic lithol- the meteoroid might also spend part of its CRE ogy found in the Florence sample. In summary, on an Earth-crossing orbit (after ejection from so far, it seems that Khatyrka does not have a the resonance), and because its initial ejection potential “peer” among the other CV chondrites. from the parent asteroid will also result in an or- bit slightly different from the one of its parent (e.g., an orbital velocity change of ~50 m/s, about the escape velocity of a ~100 km asteroid, Page 11 of 27 Fig. 5: Potential parent asteroids for Khatyrka. Distribution of large asteroids (gray) and K-type asteroids (green) in the asteroid belt in inclination vs. semi-major axis space (top panel, a). Also shown are the positions of the seven most important resonances, and the position of two asteroid families discussed in the main text. The lower two panels show the cumulative probability density distribution (bin size 0.001 au, bin fraction given on right axis) of the initial semi- 4 4 major axis of 5×10 modeled Khatyrka meteoroids (10 for each of the five radii tested), for two surface thermal con - -1 -1 -1 -1 ductivities, 1.5 Wm K (dark shades) and 0.15 Wm K (light shades), in the vicinity of two orbital resonances which could have delivered the Khatyrka meteoroid within 2-4 Ma to Earth, 3:1 (middle panel, b) and 5:2 (bottom panel, c). Large K-type asteroids in this region are shown as green circles (with the symbol size proportional to the diameter, and the asteroid number given for the largest of them). Smaller members of the Julia family are shown as green crosses. results in a semi-major axis change of up to rate close to the one of Saturn). These are the ~0.01-0.02 au). Using a simple model (Meier et only three resonances which can plausibly de- al., 2017), we track the orbital migration of me- liver the Khatyrka meteoroid to an Earth-cross- teoroids with radii of 0.5, 1, 1.5, 3, and 5 me- ing orbit within the short CRE age range of 2-4 ters, with corresponding CRE ages of 2.5, 2.5, Ma (Gladman, 1997). While the large K-type 3.5, 5, and 6 Ma, respectively, in the vicinity of Eos family is overlapping the 9:4 resonance, the the 3:1, 5:2, and ν resonances (asteroids in the median time for meteoroid delivery to Earth secular ν resonance have an orbital precession through that resonance is about 90 Ma (Glad- Page 12 of 27 man, 1997). As a consequence, the delivery of large K-type asteroids are found within 0.05- meteoroids from the 9:4 resonance (and the K- 0.06 au of the ν resonance (thus it is not shown type Eos family located there) is very inefficient as a separate panel in Fig. 5), but there are sev- (di Martino et al., 1997). For a surface thermal eral large K-type asteroids within that distance -1 -1 conductivity K = 1.5 Wm K (measured di- of the 3:1 resonance. Perhaps most interestingly, rectly in CV chondrites; Opeil et al., 2012), the the ~145 km diameter asteroid 89 Julia (the sec- model yields migration distances on the order of ond-largest K-type asteroid after 15 Eunomia), 0.02-0.03 au in the vicinity of the 3:1, 5:2, and has a compact (i.e., young) cratering family as- ν6 resonances. This distance is nearly indepen- sociated to it (Nesvorny et al., 2015). Asteroid dent of meteoroid size: despite the higher migra- 89 Julia and its family are also located relatively tion rate of small meteoroids, the higher produc- close to the ν resonance (both in semi-major tion rate of cosmogenic nuclides in small mete- axis and inclination), providing an additional oroids requires that their measured inventory of channel through which meteoroids could be de- cosmogenic noble gases must have been accu- livered to Earth-crossing orbits. Nesvorny et al. mulated in less time, thereby limiting the dis- (2015) do not give an age for the Julia family, tance they can migrate. but a rough upper limit (ignoring the initial ve- locity distribution imparted on the fragments in In Fig. 5 (middle and lower panels), we show the collision) can be determined from the largest probability density distributions for the initial distance between a small (km-sized) family semi-major axes of the modeled meteoroids member and 89 Julia, about 0.01 au. A charac- -1 -1 (dark shades: K = 1.5 Wm K ). As there are no teristic migration rate for a km-sized K-type as- large K-type asteroids within 0.02-0.03 au of the teroid, at 2.5 au, is within a factor of a few of three resonances, the Khatyrka meteoroid likely -4 10 au/Ma, which thus implies an age of a few had a higher migration rate (although, as men- 100 Ma, compatible with the <600 Ma age of tioned above, the initial “kick” at ejection from the event that reset the U,Th-He clocks in the parent asteroid might have added at least an- Khatyrka. other ~0.02 au). Higher migration rates are pos- sible if the surface is partially dust-covered Interestingly, asteroid 89 Julia displays an addi- and/or strongly fractured, which reduces the sur- tional peculiarity, strengthening its possible face thermal conductivity, resulting in a stronger connection to Khatyrka: while it fits the typical Yarkovsky effect (Vokrouhlicky et al., 2015). K-type spectrum well at short wave-lengths, it Surface thermal conductivities can vary by shows a steep red slope long-ward of 1 μm, and about three orders of magnitude between the a plateau above 1.5 μm (Birlan et al., 2004). -3 - very fine-grained lunar regolith (K ~10 Wm This was recognized already by Gaffey et al. 1 -1 -1 -1 K ) and bare rock (K ~1 Wm K ; Rubincam, (1993), who suggested that a contribution from 1995). As shown in Fig. 5, a reduction of the a Ca-rich clinopyroxene lithology could explain surface thermal conductivity by an order of 89 Julia’s peculiar spectrum (see also Cloutis, -1 -1 magnitude (to K = 0.15 Wm K ; light shades) 2002; while Gaffey et al. 1993 classified 89 Ju- allows the Khatyrka meteoroid to migrate a dis- lia as an “ungrouped S”, its shallow absorption tance of 0.05-0.06 au in 2-4 Ma, and thus to at 1 µm is a better fit to the definition of the K- reach the 5:2 resonance from asteroids 519 Syl- type asteroids. Bus and Binzel (2002) subse- vania and 599 Luisa (the latter asteroid has a quently classified it as K-type). The Florence spectrum closely matching the one of the CV Sample of Khatyrka, in which the first qua- chondrite Mokoia; Burbine et al., 2001). No sicrystal was identified (Bindi et al., 2009), as Page 13 of 27 well as some of the grains recovered in the 2011 four properties: (1) a K-type reflectance spec- expedition, represent an achondritic intergrowth trum matching, at short wavelengths, the CV of Cu,Al-alloys, olivine, spinel and diopside- classification of Khatyrka; (2) a peculiar re- hedenbergite – and the latter is indeed a Ca-rich flectance spectrum at longer wavelengths, sug- clinopyroxene (MacPherson et al., 2013). If the gestive of the presence of a Ca-rich clinopyrox- unusual Florence Sample represents an impor- ene lithology, reminiscent of the unusual achon- tant lithology on the Khatyrka parent asteroid, dritic lithology represented by the “Florence we would expect this asteroids’ reflectance sample” of Khatyrka; (3) proximity to two effi- spectrum to resemble the peculiar one of 89 Ju- cient and fast orbital resonances, 3:1 and ν , ca- lia. While a detailed spectral reflectance study pable of delivering a meteoroid within the CRE of the different Khatyrka materials – including age of 2-4 Ma to Earth; and (4) a strong shock the Cu,Al-alloys – in the visible and near-in- event within the last few 100 Ma (the cratering frared (~0.5 – 2.5 μm) is beyond the scope of event which led to the formation of the Julia as- this work, we encourage further investigation of teroid family), matching the U,Th-He age of ap- the issue. proximately <600 Ma of Khatyrka olivine grains. 5. CONCLUSIONS Acknowledgments: MM thanks J. Gilmour for The analysis of light noble gases (He and Ne) in the hospitality during his visit to SEAES. MM six forsteritic olivine grains from the quasicrys- was supported by a Swiss National Science tal-bearing CV meteorite Khatyrka shows that ox Foundation (SNSF) Ambizione grant. LB was the precursor meteoroid must have traveled in- supported by the “Progetto di Ateneo 2015” is- terplanetary space for 2-4 Ma if it was smaller sued by the Università di Firenze, Italy. PRH than 1.5 m in radius, or, less likely, >4 Ma if it was funded by a grant from the Tawani Founda- was larger and the sample heavily shielded. tion. H. Busemann’s work was supported by While this CRE age is similar to CRE ages in a NCCR PlanetS funded by SNSF. The authors few other CV chondrites, the U,Th-He retention also thank Jean-Luc Margot, reviewers Gregory ages of five of the six grains are much shorter F. Herzog and Julia A. Cartwright, as well as than in any other CV chondrite, and imply that two anonymous reviewers of an earlier version Khatyrka experienced a strong shock event, ap- of the manuscript, for their comments. proximately <600 Ma ago (assuming chondritic U,Th concentrations in our grains). The shock Author contributions: MM carried out the noble pressures needed to reset the U,Th-He clock to gas analysis and the orbital evolution modeling, the extent observed (at least ~5-18 GPa), and wrote the manuscript, and together with RW, the corresponding shock temperatures (up to LB, PJS, and H. Busemann, conceived and de- ~1500 °C) expected in an oxidized CV chon- veloped the project. PRH measured chemical drite of typical porosity, are consistent with the compositions of the grains, and together with minimum pressure and temperature required to AIN carried out the CT scan. NHS transferred explain some mineralogical observations in all grains with the micro-manipulator, MEIR de- Khatyrka (Hollister et al., 2014; Lin et al., termined NeH interferences, CM maintained 2017). While several large K-type asteroids near and set-up the unique noble gas mass spectrom- the 3:1 and 5:2 resonances are potential parents eter ready for analysis, and together with H. for Khatyrka, the asteroid 89 Julia and its asso- Baur helped with the development of the new ciated collisional family stand out by uniting fitting routine. Page 14 of 27 Busemann, H., Baur, H., Wieler, R., 2000. Pri- REFERENCES mordial noble gases in “Phase Q” in car- Amelin, Y., Krot, A., 2007. Pb isotopic age of bonaceous and ordinary chondrites stud- the Allende chondrules. Meteorit. Planet. ied by closed system stepped etching. Sci. 42, 1321–1335. Meteorit. Planet. Sci. 35, 949–973. Asimow, P.D., Lin, C., Bindi, L., Ma, C., Choi, J., Nagao, K., Baek, J.M., Bartoschewitz, Tschauner, O., Hollister, L.S., Stein- R., Park, C., Lee, J.I., Lee, M.J., 2017. hardt, P.J., 2016. Shock synthesis of Noble Gas Compositions of Seven CV, quasicrystals with implications for their CO, and CK Chondrites from Deserts. origin in asteroid collisions. Proc. Natl. Presented at the Lunar and Planetary Acad. Sci. 201600321. Science Conference, p. 1852. Baur, H., 1999. A noble-gas mass spectrometer Cloutis, E.A., 2002. Pyroxene reflectance spec- compressor source with two orders of tra: Minor absorption bands and effects magnitude improvement in sensitivity, of elemental substitutions. J. Geophys. in: EOS Transactions, AGU 46. Res. Planets 107, 6–1. Bell, J.F., 1988. A probable asteroidal parent Cloutis, E.A., Hudon, P., Hiroi, T., Gaffey, body for the CO or CV chondrites. Me- M.J., Mann, P., Bell, J.F., 2012. Spectral teoritics 23, 256–257. reflectance properties of carbonaceous Bindi, L., Eiler, J.M., Guan, Y., Hollister, L.S., chondrites: 6. CV chondrites. Icarus 221, MacPherson, G., Steinhardt, P.J., Yao, 328–358. N., 2012. From the Cover: Evidence for Consolmagno, G.J., Britt, D.T., Macke, R.J., the extraterrestrial origin of a natural 2008. The significance of meteorite den- quasicrystal. Proc. Natl. Acad. Sci. 109, sity and porosity. Chem. Erde - 1396–1401. Geochem. 68, 1–29. Bindi, L., Steinhardt, P.J., 2014. The quest for di Martino, M., Migliorini, F., Cellino, A., Zap- forbidden crystals. Mineral. Mag. 78, pala, V., 1997. Can CO/CV Meteorites 467–482. Come from the EOS Family? Meteorit. Bindi, L., Steinhardt, P.J., Yao, N., Lu, P.J., Planet. Sci. Suppl. 32. 2009. Natural Quasicrystals. Science Farinella, P., Vokrouhlický, D., Hartmann, 324, 1306. W.K., 1998. Meteorite Delivery via Birlan, M., Barucci, M.A., Vernazza, P., Yarkovsky Orbital Drift. Icarus 132, Fulchignoni, M., Binzel, R.P., Bus, S.J., 378–387. Belskaya, I., Fornasier, S., 2004. Near- Gaffey, M.J., Bell, J.F., Brown, R.H., Burbine, IR spectroscopy of asteroids 21 Lutetia, T.H., Piatek, J.L., Reed, K.L., Chaky, 89 Julia, 140 Siwa, 2181 Fogelin and D.A., 1993. Mineralogical Variations 5480 (1989YK8), potential targets for within the S-Type Asteroid Class. Icarus the Rosetta mission; remote observations 106, 573–602. campaign on IRTF. New Astron. 9, 343– Gladman, B.J., 1997. Dynamical Lifetimes of Objects Injected into Asteroid Belt Res- Bland, P.A., Artemieva, N.A., 2006. The rate of onances. Science 277, 197–201. small impacts on Earth. Meteorit. Planet. Hamann, C., Stöffler, D., Reimold, W.U., 2016. Sci. 41, 607–631. Interaction of aluminum projectiles with Burbine, T.H., Binzel, R.P., Bus, S.J., Clark, quartz sand in impact experiments: For- B.E., 2001. K asteroids and CO3/CV3 mation of khatyrkite (CuAl2) and reduc- chondrites. Meteorit. Planet. Sci. 36, tion of SiO2 to Si. Geochim. Cos- 245–253. mochim. Acta 192, 295–317. Bus, S.J., Binzel, R.P., 2002. Phase II of the Heber, V.S., Wieler, R., Baur, H., Olinger, C., Small Main-Belt Asteroid Spectroscopic Friedmann, T.A., Burnett, D.S., 2009. Survey. A Feature-Based Taxonomy. Noble gas composition of the solar wind Icarus 158, 146–177. as collected by the Genesis mission. Page 15 of 27 Geochim. Cosmochim. Acta 73, 7414– Kryachko, V., Steinhardt, W.M., Yu- 7432. dovskaya, M., Steinhardt, P.J., 2013. Heck, P.R., Marhas, K.K., Hoppe, P., Gallino, Khatyrka, a new CV3 find from the Ko- R., Baur, H., Wieler, R., 2007. Presolar ryak Mountains, Eastern Russia. Mete- He and Ne Isotopes in Single Circum- orit. Planet. Sci. 48, 1499–1514. stellar SiC Grains. Astrophys. J. 656, MacPherson, G.J., Lin, C., Hollister, L.S., 1208–1222. Bindi, L., Andronicos, C.L., Steinhardt, Hollister, L.S., Bindi, L., Yao, N., Poirier, G.R., P.J., 2016. The Khatyrka Meteorite: A Andronicos, C.L., MacPherson, G.J., Summary of Evidence for a Natural Ori- Lin, C., Distler, V.V., Eddy, M.P., gin of Its Remarkable Cu-Al Metal Al- Kostin, A., Kryachko, V., Steinhardt, loys. Presented at the Lunar and Plane- W.M., Yudovskaya, M., Eiler, J.M., tary Science Conference, p. 2655. Guan, Y., Clarke, J.J., Steinhardt, P.J., Meier, M.M.M., Alwmark, C., Bajt, S., Böttger, 2014. Impact-induced shock and the for- U., Busemann, H., Fujiya, W., Gilmour, mation of natural quasicrystals in the J., Heitmann, U., Hoppe, P., Hübers, H.- early solar system. Nat. Commun. 5. W., Marone, F., Ott, U., Pavlov, S., Ivanova, M.A., Lorenz, C.A., Borisovskiy, S.E., Schade, U., Spring, N., Stampanoni, M., Burmistrov, A.A., Korost, D.V., Ko- Weber, I., 2014. A Precise Cosmic-Ray rochantsev, A.V., Logunova, M.N., Exposure Age for an Olivine Grain from Shornikov, S.I., Petaev, M.I., 2017. the Surface of Near-Earth Asteroid Composition and origin of holotype Al- (25143) Itokawa. Presented at the Lunar Cu-Zn minerals in relation to quasicrys- and Planetary Science Conference, p. tals in the Khatyrka meteorite. Meteorit. 1247. Planet. Sci. 52, 869-883. Meier, M.M.M., Welten, K.C., Riebe, M.E.I., Kööp, L., Davis, D.W., 2012. Classification and Caffee, M.W., Gritsevich, M., Maden, U-Pb Isotopic Study of Northwest Africa C., Busemann, H., 2017. Park Forest 6514. Presented at the Lunar and Plane- (L5) and the asteroidal source of tary Science Conference, p. 2066. shocked L chondrites. Meteorit. Planet. Levine, D., Steinhardt, P.J., 1984. Quasicrystals Nesvorny, D., Broz, M., Carruba, V., 2015. - A new class of ordered structures. Identification and Dynamical Properties Phys. Rev. Lett. 53, 2477–2480. of Asteroid Families. ArXiv150201628 Leya, I., Ammon, K., Cosarinsky, M., Dalcher, Opeil, C.P., Consolmagno, G.J., Safarik, D.J., N., Gnos, E., Hofmann, B., Huber, L., Britt, D.T., 2012. Stony meteorite ther- 2013. Light noble gases in 12 meteorites mal properties and their relationship from the Omani desert, Australia, Mauri- with meteorite chemical and physical tania, Canada, and Sweden. Meteorit. states: Meteorite thermal properties. Me- Planet. Sci. 48, 1401–1414. teorit. Planet. Sci. 47, 319–329. Leya, I., Masarik, J., 2009. Cosmogenic nu- Ott, U., 2014. Planetary and pre-solar noble clides in stony meteorites revisited. Me- gases in meteorites. Chem. Erde - teorit. Planet. Sci. 44, 1061–1086. Geochem. 74, 519–544. Lin, C., Hollister, L.S., MacPherson, G.J., Rubincam, D.P., 1995. Asteroid orbit evolution Bindi, L., Ma, C., Andronicos, C.L., due to thermal drag. J. Geophys. Res. Steinhardt, P.J., 2017. Evidence of 100, 1585–1594. cross-cutting and redox reaction in Ruzicka, A., Grossman, J.N., Garvie, L., 2014. Khatyrka meteorite reveals metallic-Al The Meteoritical Bulletin, No. 100, 2014 minerals formed in outer space. Sci. June. Meteorit. Planet. Sci. 49, E1– Rep. 7. E101. MacPherson, G.J., Andronicos, C.L., Bindi, L., Scherer, P., Schultz, L., 2000. Noble gas record, Distler, V.V., Eddy, M.P., Eiler, J.M., collisional history, and pairing of CV, Guan, Y., Hollister, L.S., Kostin, A., CO, CK, and other carbonaceous chon- Page 16 of 27 drites. Meteorit. Planet. Sci. 35, 145– Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.-Y., White, D.J., Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: an open- source platform for biological-image analysis. Nat. Methods 9, 676–682. Schmitt, R.T., 2000. Shock experiments with the H6 chondrite Kernouvé: Pressure calibration of microscopic shock effects. Meteorit. Planet. Sci. 35, 545–560. Schultz, L., Franke, L., 2004. Helium, neon, and argon in meteorites: A data collection. Meteorit. Planet. Sci. 39, 1889–1890. Sharp, T.G., de Carli, P.S., 2006. Shock Effects in Meteorites, in: Meteorites and the Early Solar System II. pp. 653–677. Shechtman, D., Blech, I., Gratias, D., Cahn, J.W., 1984. Metallic phase with long- range orientational order and no transla- tional symmetry. Phys. Rev. Lett. 53, 1951–1953. Stagno, V., Bindi, L., Park, C., Tkachev, S., Prakapenka, V.B., Mao, H.-K., Hemley, R.J., Steinhardt, P.J., Fei, Y., 2015. Qua- sicrystals at extreme conditions: The role of pressure in stabilizing icosahedral Al Cu Fe at high temperature. Am. 63 24 13 Mineral. 100, 2412–2418. Stoeffler, D., Keil, K., Scott, E.R.D., 1991. Shock metamorphism of ordinary chon- drites. Geochim. Cosmochim. Acta 55, 3845–3867. Trull, T.W., Kurz, M.D., Jenkins, W.J., 1991. Diffusion of cosmogenic3He in olivine and quartz: implications for surface ex- posure dating. Earth Planet. Sci. Lett. 103, 241–256. Vokrouhlicky, D., Bottke, W.F., Chesley, S.R., Scheeres, D.J., Statler, T.S., 2015. The Yarkovsky and YORP Effects. ArXiv150201249 Wasson, J.T., Kallemeyn, G.W., 1988. Compo- sitions of Chondrites. Philos. Trans. R. Soc. Lond. Ser. Math. Phys. Sci. 325, 535–544. Page 17 of 27 Tables Table 1: Volumes, masses, noble gases and U,Th-He ages of six Khatyrka olivine grains #126-01 #126-02 #126-03 #126-04 #126-05 #126-07 Voxels* 219±14 190±13 228±15 292±18 274±23 181±15 Volume, μm 35800±2300 31200±2100 37400±2400 47900±3000 44900±3800 29700±2500 Diameter, μm** 40.9±0.9 39.0±0.9 41.5±0.9 45.1±0.9 44.1±1.2 38.4±1.1 #Fo*** 0.96 0.96 0.98 0.91 0.97 0.96 Calculated mass, μg 0.119±0.008 0.104±0.007 0.125±0.008 0.160±0.010 0.150±0.013 0.099±0.009 3 21 3,4 20,21,22 3 20,21,22 3,4 21,22 3,4 20,21,22 3 21 Above DL**** He, Ne He, Ne He, Ne He, Ne He, Ne He, Ne 3 4 He/ He (>0.025) 0.0035±0.0006 (>0.026) 0.025±0.009 0.026±0.020 (>0.018) 3 3 He = He 5.41±0.51 5.33±0.46 5.30±0.41 6.13±0.46 5.60±0.55 4.50±0.49 meas cos 4 4 He = He (<190) 1510±250 (<180) 220±90 190±170 (<240) rad non-cos 20 22 Ne/ Ne -- 11.5±1.7 17.7±3.9 (<6.58) 10.7±1.6 -- 21 22 Ne/ Ne (>0.87) 0.086±0.016 0.20±0.06 0.64±0.18 0.14±0.03 (>0.26) Ne 1.83±0.21 1.36±0.22 0.91±0.19 1.80±0.18 1.33±0.22 0.65±0.17 meas Ne 1.83±0.22 0.92±0.22 0.80±0.20 1.78±0.19 1.08±0.22 (>0.60) cos 3 21 He/ Ne 2.95±0.35 5.80±1.38 6.67±1.65 3.45±0.33 5.18±0.96 7.20±2.10 cos 20 4 Ne / He (>0.0079) 0.0006±0.0002 (>0.0036) 0.0067±0.0025 0.0046±0.0036 (>0.0023) cos R4, Ma (U=14 ppb) (<590) 3520 (<560) 680 590 (<740) R4, Ma (U=23 ppb) (<370) 2460 (<350) 420 370 (<460) *Note that the number of voxels are rounded, i.e., are not integers, due to the surface-effect corrections. **The diameter is for a spherical grain with the same volume. ***#Fo is the (nominal) forsterite number. #Fo = 0.95±0.05 was used to calculate the mass. ****“Above DL” lists the isotopes which have been detected above their respective detection limits as -8 3 discussed in the methods section and the Supplementary Text. Concentrations are given in units of 10 cm STP/g, uncertainties are 1σ. Values in brackets are upper or lower limits resulting from only one of the two involved isotopes being above DL. “--” is given for ratios where neither of the two involved isotopes was above DL. “R4” are the nominal radio - genic ages determined from Herad, assuming a bulk U concentration of either 14 or 23 ppb (and Th/U = 3.5). Page 18 of 27 measured in the post-extraction cycles are ex- SUPPLEMENTARY TEXT trapolated backward in time to the same mo- ment. The signal contribution from the sample An improved noble gas data fitting routine gas is then defined as the difference between the for very low gas amounts values of the two extrapolations at time zero. In the unique (inhouse-built) “compressor- The uncertainty of the sample signal is the source” noble gas mass spectrometer at ETH quadratically added error of the two extrapola- Zurich (Baur, 1999), used for the analysis of tions at time zero. A wide variety of functions very small amounts of He and Ne (>~10’000 can be chosen to do these extrapolations, e.g., atoms), the “memory” signal from He and Ne linear, quadratic, exponential or double expo- constantly increases over the course of an analy- nential (a superposition of two exponential) sis. While this phenomenon exists in many functions. Whenever possible, the same type of static vacuum noble gas mass spectrometers, it function is used for the same ion species is exacerbated by the use of the compressor throughout an analysis session, unless the ex- source, which concentrates gas from the entire trapolation obviously fails to provide a physi- volume of the mass spectrometer into the vol- cally meaningful fit to the data. To determine ume of the ion source. For samples with gas the “detection limit”, i.e., the smallest gas amounts close to the detection limit, the change amount that can be reliably resolved from the in signal from this “memory” over the duration background noise in this type of analysis, a se- of an analysis is much larger than the signal ries of “blank runs” is measured. These are from the sample. Heck et al. (2007) developed a identical to the normal sample runs, except that protocol in which the increase of the memory the laser is either not operated at all (a so-called signals in the static vacuum of the mass spec- “cold blank”) or aimed at an empty spot on the trometer and the extraction line is first moni- Al sample holder (a so-called “hot blank”). The tored over four or five cycles (after separation scatter of the gas amounts derived from the from the turbo-molecular pumps). A cycle is a blank measurements, which typically scatter sequence in which the ion beam currents of all + 3 + 4 + around zero regardless of their type, is then used species of interest, typically HD , He , He , + 20 + 21 + 22 + 40 + + to determine the detection limit, defined as two H O , Ne , Ne , Ne , Ar , and CO , are 2 2 standard deviations of the blank scatter above measured in peak jumping mode on a single de- the average blank value. A “detection” of gas tector (either a channel electron multiplier or a for a specific species is only claimed for sam- Faraday cup). Then, without opening or closing ples where the 1σ lower limit of the measured any valves (which might release additional gas value is above the detection limit. This approach into the spectrometer), the laser is fired at the has allowed to reliably measure very small gas sample, melting or in some cases completely va- amounts in a variety of samples. porizing it, and releasing all He and Ne in the process. After extraction (referred to as “time Nevertheless, this approach has its shortcom- zero”), another five to ten cycles are measured. ings. This can be exemplified using blank mea- In the cycles before the extraction, the gas in the surements, in which generally no gas is released spectrometer consists exclusively of memory at time zero. A single fit running through all cy- gas, but after the extraction, the sample gas has cles (both from before and after time zero) pro- been added. In the protocol developed by Heck vides a better (i.e., more precise) estimate of the et al. (2007), the signals measured in the four or absolute signal intensity at time zero than the five pre-extraction cycles are extrapolated for- “classical” approach used above. First, this is ward in time to “time zero”, while the signals Page 19 of 27 because the available data set is larger than in the quadratic and exponential fits, and from the case of the two individual fits. Second, a fit- eight to five in the double-exponential fit. Fur- ting function is usually constrained best near the thermore, time zero is now close to the center of center of the data distribution. However, we are the distribution of values, which can be opti- not interested in the centers of the pre-run and mized by measuring the same number of cycles the post-run sections, and only really care about before and after time zero. The uncertainty on P time zero, a point which is at the edge of each of (determined by standard statistical methods and the two time domains, i.e., where the fit func- not to be confused with the error of the fit func- tions are likely to provide a poor estimate. It tion at time zero) is the uncertainty of the signal would be better if the fitting function had “time change due to gas release at time zero. Since zero” close to the center of its data distribution. this value will still scatter around zero for cold This is possible when using a fitting function and hot blanks, we still have to determine a de- with a “step” added at time zero. By using a sin- tection limit using a suite of hot and cold gle function for the entire analysis run, we as- blanks, but since their scatter (and their individ- sume that the rate of signal change does not ual uncertainties) is reduced in the new fitting change at time zero, for which there is no physi- approach, this detection limit will, typically, be cal justification if significant gas amounts are lower than in the “classical” approach. released by laser heating. We discuss this issue The new fitting routine also has its limitations. further below. In addition, and in spite of the For most species of interest, the rate of signal concept of negative gas amounts released at change at any time depends on the total signal time zero not being a physical one, the fitting al- of that species. Consequently, if the signal gorithm is allowed to assume negative values jumps due to the sample gas added at time zero for the step height. This is to allow for the statis- is too large, the slope of the memory increase tical nature of the approach, e.g., due to count- before and after time zero will be different, in ing statistics on individual data points, and not which case the step function fit will be system- to bias the blanks to positive values. For a linear atically wrong compared to the classical ap- fit, the step function has the following form: proach, which uses two independently fitted s(t) = P + P × 0.5 × (1 + sign(t)) + P × t slopes. Therefore, the step function fit can only 0 1 2 be used for very low gas amounts, which do not Here, s is the signal at time-index t, while P , P , 0 1 change the slope of the memory increase signifi- and P are fitting parameters (y-axis intercept, cantly. In particular, the step function cannot be step-size, and slope, respectively), sign(t) is the used to fit a calibration run (an analysis of a “sign” function, which is -1 if t (time) is <0 known quantity of gas used to determine the (prior to time zero), and +1 if t is >0 (after time isotope-specific sensitivity of the mass spec- zero). Therefore, the value of the P -term (P × 1 1 trometer) because the slope after time zero is 0.5 × (1 + sign(t))) is zero before time zero, and negative for the large gas amounts introduced in P after that. Hence, P corresponds to the signal 1 1 a calibration run. Finally, and as a higher order increase due to the sample gas released at time effect, if laser heating releases a large amount of zero (when the laser was fired). For non-linear gas species other than a particular one of interest functions, the P1-term is added analogously. (possibly one that is not monitored), the slope of Compared to the “classical” approach, this re- the memory signal increase of the species of in- duces the number of fitting parameters in the terest (which also depends on the total pressure linear fit from four to three, from six to four in in the mass spectrometer) might also change at Page 20 of 27 time zero. Therefore, the step function should only be used for relatively “clean” samples which release very small gas amounts. As a gen- eral rule, the signal jump at time zero should be smaller than the total memory increase of a species of interest over the course of an analy- sis. In summary, the step function cannot re- place the classical approach, but it can comple- ment and improve it for very low gas amounts. Page 21 of 27 SUPPLEMENTARY TABLES Supplementary Table S1 3 4 20 21 22 36 38 40 4 21 Meteorite (CV) He He Ne Ne Ne Ar Ar Ar Ref. He Ne R4, 14 R4, 23 T21 nc c Acfer 082 1.41 152 2.97 0.49 0.77 13.9 2.71 1070 1 144 0.48 453 279 1.5 Acfer 086 10.7 1690 2.75 1.53 1.98 11.0 2.34 2910 1 1680 1.52 3770 2670 4.7 Acfer 272 0.25 440 3.12 0.66 1.03 29.6 5.76 383 1 439 0.65 1300 820 2.0 ALH84028 41.6 2700 9.98 7.11 8.52 16.6 4.14 2360 1 2450 7.09 4620 3500 21.8 ALH85006 5.95 3160 9.00 1.45 2.42 24.0 4.68 1050 1 3120 1.42 *4570 4050 4.4 ALHA81003 26.7 2269 5.05 4.13 4.88 8.07 1.96 2700 1 2110 4.12 4290 3160 12.6 Allende 7.63 2830 5.48 1.76 2.28 16.9 3.30 2310 1 2780 1.75 *4570 3790 5.4 Arch 112 273000 785 8.50 71.5 95.3 19.1 350 1 273000 6.57 *4570 *4570 20.2 Axtell 22.4 2290 8.56 5.41 6.23 10.2 2.59 211 1 2150 5.41 4330 3200 16.6 Bali 26.2 2740 18.7 9.58 11.3 30.7 6.67 1530 1 2580 9.55 *4570 3610 29.3 Denman 002 20.0 2260 5.71 3.01 3.77 6.95 1.81 752 1 2140 3.00 4320 3190 9.2 Efremovka 9.93 1050 12.3 3.03 4.10 246 45.8 1390 1 990 3.01 2600 1720 9.2 Grosnaja 1.15 1730 6.93 0.56 1.30 25.3 4.95 1040 1 1720 0.54 3820 2720 1.7 Kaba 4.15 2310 11.9 3.70 5.25 39.4 8.39 597 1 2280 3.66 4460 3330 11.2 Leoville 14.1 2280 11.7 2.84 4.05 226 41.8 1790 1 2200 2.81 4380 3250 8.6 Mokoia 57.9 97600 334 3.91 29.6 23.5 4.80 1410 1 97500 3.12 *4570 *4570 9.6 Mundrabilla 012 15.4 1350 2.22 2.14 2.56 1.67 0.72 941 1 1330 2.13 3240 2220 6.5 Nova 002 1.86 899 9.53 2.42 3.45 183 35.1 617 1 888 2.40 2390 1570 7.4 NWA 10670 6.01 1730 3.28 1.15 1.61 2.86 0.65 167 2 1700 1.14 3800 2690 3.5 NWA 3304 298 604000 1110 10.0 95.5 59.2 12.0 283 2 604000 7.41 *4570 *4570 22.7 NWA 6743 59.2 1510 10.4 8.87 10.8 102 20.6 954 2 1150 8.84 2920 1960 27.1 NWA 6746 5.59 1010 4.67 1.36 1.88 0.72 0.14 73.1 2 976 1.35 2570 1700 4.1 QUE 93429 50.3 2460 10.5 8.53 9.93 17.8 4.79 1850 1 2160 8.52 4340 3210 26.1 QUE 93744 52.1 1930 10.9 8.67 10.2 30.8 7.24 3400 1 1610 8.65 3670 2580 26.5 RaS 221 25.1 1570 8.72 5.99 7.39 15.2 3.51 667 3 1410 5.97 3370 2320 18.3 RaS 251 9.83 6420 6.29 1.59 2.37 4.94 1.34 2450 3 6360 1.57 *4570 *4570 4.8 Tibooburra 42.2 2630 11.3 7.54 9.11 88.5 17.4 171 1 2380 7.52 4560 3430 23.1 Vigarano 8.75 4940 17.1 1.85 3.32 43.3 8.39 2120 1 4880 1.81 *4570 *4570 5.5 Page 22 of 27 Y 86009 12.9 23500 147 2.16 13.6 58.0 11.2 799 1 23500 1.81 *4570 *4570 5.6 Y 86751 27.7 21300 138 7.92 19.0 169 33.4 1210 1 21200 7.60 *4570 *4570 23.3 -8 3 Compilation of noble gas data used to create Fig. 4. All concentrations are given in 10 cm STP/g. References (Ref.) are (1) the compilation by Schultz and Franke (2004), and re- 21 21 search papers by (2) Choi et al. (2017) and (3) Leya et al. (2013). The cosmogenic Ne ( Ne ) was calculated by a two-component deconvolution as described in section 3.1. The 4 4 4 4 3 3 21 4 non-cosmogenic He ( He ) was calculated as follows: He = He – 6 × He , unless He was larger than 7 × Ne , in which case that value was subtracted from He in- nc nc meas meas meas c meas stead (based on Leya & Masarik, 2009). The U,Th-He ages (R4; in Ma) are given for U concentrations of 14 ppb and 23 ppb, in the third- and second-to last column, respectively 21 21 -8 3 (Th/U = 3.5; R4 ages are caped at 4570 Ma). The Ne-based CRE age (T21; in Ma) in the last column assumes a uniform Ne production rate of 0.326 × 10 cm STP/gMa. Supplementary Table S2 Grain O (%) Na (%) Mg (%) Al (%) Si (%) Ca (%) Fe (%) Total (%) Mg# Description #126-01 45.91 n.d. 33.41 n.d. 19.14 n.d. 1.54 100 0.96 Forsterite #126-02 42.37 n.d. 35.29 n.d. 20.83 n.d. 1.51 100 0.96 Forsterite #126-03 52.44 n.d. 30.81 n.d. 16.1 n.d. 0.66 100 0.98 Forsterite #126-04 45.32 n.d. 32.05 n.d. 19.6 n.d. 3.03 100 0.91 Forsterite #126-05 51.01 n.d. 30.58 n.d. 17.48 n.d. 0.93 100 0.97 Forsterite #126-06 46.46 n.d. 30.08 n.d. 18.13 n.d. 5.34 100 0.85 Forsterite #126-07 55.21 n.d. 29.36 n.d. 14.34 n.d. 1.09 100 0.96 Forsterite Fo 95 44.49 - 32.10 - 19.52 - 3.88 100 0.95 (theoretical) Qualitative chemical composition of the seven original grains determined by SEM-EDS. Grain #126-06 was not selected for analysis. All abundances have been renormalized to 100%. „n.d.“= not detected. Page 23 of 27 SUPPLEMENTARY FIGURES Fig. S1: Grain BSE images. All six Khatyrka olivine grains imaged by SEM (BSE) after nano-CT scanning. The white scale bar is 20 μm long in each image (Field Museum of Natural History, Chicago). Page 24 of 27 Fig. S2: Khatyrka olivine grain on micro-manipulator needle. Grain #126-07 on the tip of a micro-manipulator nee- dle, during transfer from the carbon tape to the sample holder which was later used for noble gas analysis (SEAES, University of Manchester). Page 25 of 27 20 22 21 22 Fig. S3: Three isotope diagram for Ne. Neon isotope ratios ( Ne/ Ne vs. Ne/ Ne) for all six Khatyrka olivine grains. For grains #126-01 and -07, only Ne was detected above the detection 21 22 limit, and therefore, only a lower bound (2σ) on the Ne/ Ne ratio can be given – the data points 20 22 20 (right-pointing triangles) are placed arbitrarily at Ne/ Ne = 0. For grain #126-04, Ne was not 20 22 detected, therefore only an upper bound (2σ) on the Ne/ Ne ratio can be given (down-pointing triangle). Grain #126-03 plots nominally outside the field in which we would expect it (surrounded by dashed lines), although it is still less than 2σ away. Also shown are four noble gas components (black diamonds): the solar wind as measured by Genesis (SW; Heber et al., 2009), Q gases and the Earth’s atmosphere (Ott, 2014), and cosmogenic Ne (GCR; Leya and Masarik, 2009). Page 26 of 27 Fig. S4: U,Th-He ages as a function of U concentration. The U,Th-He age of all Khatyrka grains (black curves) is shown as a function of the U concentration in the range 5-150 ppb, assuming Th/U = 3.5. Also shown, for comparison, are the other CV chondrites (open circles; for 15 ppb U) as given in Fig. 4 and Supplementary Table S1. Page 27 of 27
Journal
Astrophysics – arXiv (Cornell University)
Published: Mar 9, 2018