High-spatial resolution U Pb dating of phosphate minerals in Martian meteorite Allan Hills 84001

Transcription

1 Geochemical Journal, Vol. 48, pp. 423 to 431, 2014 doi: /geochemj High-spatial resolution U Pb dating of phosphate minerals in Martian meteorite Allan Hills MIZUHO KOIKE, 1 ** YOSHIHIRO OTA, 1 YUJI SANO, 1 * NAOTO TAKAHATA 1 and NAOJI SUGIURA 2 1 Atmosphere and Ocean Research Institute, The University of Tokyo, Kashiwa, Chiba , Japan 2 Department of Earth and Planetary Science, School of Science, The University of Tokyo, Bunkyo-ku, Tokyo , Japan (Received February 15, 2013; Accepted June 4, 2014) Phosphate minerals, which are ubiquitous in terrestrial and extra-terrestrial rocks, are important carriers of trace elements, including U and Th, which provide chronological information because of their radioactive decay. Abundance and grain sizes of the phosphates are limited in extra-terrestrial materials. Therefore, high resolution and minimally or nondestructive analytical methods must be used for age determination. For this study, we conducted U Pb dating using a NanoSIMS for three phosphate grains from an old Martian meteorite: ALH The 238 U 206 Pb and 207 Pb 206 Pb isochron ages were found to be 3850 ± 170 Ma and 4002 ± 52 Ma, respectively, suggesting a concordant signature at approx. 4.0 Ga. A total Pb/U isochron age of 3990 ± 160 Ma is consistent with a previous SHRIMP U Pb age of 4018 ± 81 Ma (Terada et al., 2003). Moreover, heterogeneous distributions of U are observed in these grains, which might have been preserved since igneous crystallization of the phosphates because the diffusion of U in the mineral is considerably slow. Keywords: U Pb dating, phosphates, NanoSIMS, Martian meteorites, ALH INTRODUCTION Martian meteorites, which are known to show extremely wide variations in chemistry, chronology, mineralogy, and petrology (Meyer, 2012), are regarded as providing important clues to elucidate the evolution of the Martian mantle crust, and of surface environments. Radiometric ages of these meteorites might provide fundamental and valuable information about the Martian environment. Phosphate minerals such as apatite and merrillite are common in extraterrestrial materials. They are important carriers of incompatible elements including REE, Th and U, and were used for leaching experiment (e.g., Chen and Wasserburg, 1986). However, because the phosphates are volumetrically minor and because their grain sizes are limited, minimally or non-destructive analytical methods must be used for accurate dating, even though enriched shergottites contain phosphate mineral grains up to several hundred micrometers. Currently, in-situ U Th Pb dating by SHRIMP is a major analytical method for the radiogenic system in terrestrial (Sano et al., 1999, 2006a; Lan et al., 2013) and extraterrestrial phosphates (Sano et al., 2000; Terada et al., 2003; Terada and Sano, *Corresponding author ( ysano@aori.u-tokyo.ac.jp) **Present address: Central Research Laboratory, Hitachi, Ltd., Kokubunji, Tokyo , Japan. Copyright 2014 by The Geochemical Society of Japan. 2004; Li et al., 2012; Zhou et al., 2013; Yin et al., 2014). This report describes our study of U Pb dating of single phosphate grains in an extremely old Martian meteorite, Allan Hills (ALH) 84001, using a NanoSIMS 50 (Ametek Inc.). SAMPLE AND ANALYTICAL METHODS In 1984, Martian meteorite ALH was collected from the Far Western Icefield of Allan Hills. It is unique because of its extremely old crystallization and metamorphic age (Previous works are presented in Table 1 except for SHRIMP data, which are presented in Table 2.) and complicated thermal history (Treiman, 1995, 1998). The crystallization age is Ga (Jagoutz et al., 1994, 2009; Nyquist et al., 1995, 2001; Nyquist and Shih, 2013; Bouvier et al., 2009; Lapen et al., 2010), although Ar Ar results of studies suggest several shock heating events at Ga (Ash et al., 1996; Knott et al., 1996; Turner et al., 1997; Ilg et al., 1997; Bogard and Garrison, 1999; Cassata et al., 2010). Moreover, this meteorite experienced other chemical/physical processes including deposition of secondary carbonate minerals and possibly lowtemperature aqueous alteration (Wadhwa and Lugmair, 1996; Borg et al., 1999; Halevy et al., 2011; Beard et al., 2013). Because ALH is a rare sample of ancient Martian material with age of more than 4 billion years, elucidating its history is crucially important to understand the early days of Martian evolution, even though ex- 423

2 Table 1. Summary of ALH geochronometry Method Age (Ga) Mineral Reference 147 Sm 143 Nd ~4.56 WR, Opx, Msk Jagoutz et al., Sm 143 Nd 4.5 ± 0.13 WR, Opx Nyquist et al., Sm 143 Nd 4.41 ± 0.03 WR, Opx Lapen et al., Sm 143 Nd 4.57 ± 0.09 WR, Opx Nyquist and Shih, Sm 142 Nd 4.47 ± 0.04 WR, Opx Nyquist and Shih, Rb 87 Sr 4.58 ± 0.3 WR, Opx Nyquist et al., 1995, 2001# 87 Rb 87 Sr 3.87 ± 0.05 WR, Opx Wadhwa and Lugmair, Rb 87 Sr 1.4 ± 0.1 Crb, Msk Wadhwa and Lugmair, Rb 87 Sr 3.92 ± 0.02 Crb Borg et al., 1999 Borg and Drake, Rb 87 Sr 4.34 ± 0.23 WR, Opx, Msk Beard et al., 2013* 87 Rb 87 Sr 3.95 ± 0.02 WR, Opx, Msk Beard et al., 2013* Chr, Crb 176 Lu 176 Hf 4.09 ± 0.03 WR, Opx, Chr Lapen et al., Ar 40 Ar ~3.6 Crb Knott et al., Ar 40 Ar 4.0 ± 0.1 Msk Ash et al., Ar 40 Ar 4.07 ± 0.04 Msk Ilg et al., Ar 40 Ar 4.1 ± 0.2 Msk Bogard and Garrison, Ar 40 Ar 3.93 ± 0.12 Msk Turner et al., Ar 40 Ar 4.16 ± 0.04 Msk Cassata et al., Ar 40 Ar 1.16 ± 0.11 Opx Cassata et al., Th 208 Pb 2.93 ± 0.41 WR, Opx, Phs Jagoutz et al., Pb 206 Pb 4.05 ± 0.09 Crb Borg et al., 1999 Borg and Drake, Pb 206 Pb 4.07 ± 0.10 WR, Opx Bouvier et al., Pb 206 Pb 4.04 ± 0.14 First leachate steps Bouvier et al., Pb 206 Pb 4.12 WR, Opx Jagoutz et al., 2009 Simplified notations are: WR = whole rock, Opx = orthopyroxene, Crb = carbonates, Msk = maskelynite, Chr = chromite, Phs = phosphates. *:Remark. The carbonate isochron for ALH has been recalculated using the new decay constant; previously this isochron was calculated using an 87 Rb decay constant of yr 1, as proposed by Minster et al. (1982) and produced an age of / 19 Ma using an Isoplot regression (Borg and Drake, 2005). The Nyquist et al. (1995) and Wadhwa and Lugmair (1996, 1997) isochron ages have not been recalculated using this new decay constant because data, analytical details, and standard analyses were not reported in these abstracts, preventing a detailed assessment of the impact of different decay constants. #:Remark. Nyquist et al. (2001) described: Here, as elsewhere in this paper, we use the value of the 87 Rb decay constant recommended by Minster et al. (1982); i.e., λ 87 = yr 1. tremely old brecciated Martian meteorites (NWA 7034/ 7475/7533) have been found recently (Agee et al., 2013; Humayun et al., 2013). The samples used for this study were two polished thick sections of ALH (Fig. 1a), which were studied previously for hydrogen isotopes in carbonates and maskelynite (Sugiura and Hoshino, 2000). Back-scattered electron images of the samples were obtained preliminarily using a secondary electron microprobe with an energy dispersive spectrometer (SEM-EDS) to locate phosphate grains. Three phosphate grains with sizes of µm were found (Fig. 1b). It is well documented that orthopyroxene is the dominant mineral in ALH (Meyer, 2012). Phosphate grains appear to be surrounded by orthopyroxene (Opx) in the BSE images. They are also in close association with plagioclase glass (Pg). These signatures resemble those described in Greenwood et al. (2003), although carbonate cannot be found close to the phosphate. Chemical compositions of the phosphates, orthopyroxene, and plagioclase glass were measured using SEM-EDS. They are listed in Supplementary Table S1. The atomic Na/Ca ratio of phosphates is approximately ten, which is consistent with the chemical formula of merrillite (Mer), (Ca 18 (Mg,Fe) 2 Na 2 (PO 4 ) 14 ). Their crystallographic features were analyzed using electron backscatter diffraction (EBSD) and were identified as merrillite. The sections were polished again, gold-coated, and then baked at approx. 100 C overnight in the air-lock to reduce absorbed water, a potential contributing factor for hydride interference. 424 M. Koike et al.

3 Table U 206 Pb ages and 207 Pb 206 Pb ages for single grain dating and multi grain dating 238 U 206 Pb 207 Pb 206 Pb Total U/Pb Age (Ma) Age (Ma) Age (Ma) Grain # ± ± ± 470 Grain # ± ± ± 180 Grain # ± ± ± 540 Multi-grain 3850 ± ± ± 160 SHRIMP* 3700 ± ± ± 81 *Terada et al. (2003). Error assinged to the U Pb and Pb Pb ages is two sigma, while total U/Pb is 95% confidential level. An apatite extracted from an alkaline rock of Prairie Lake circular complex in Ontario, Canadian Shield, PRAP, with age of 1155 ± 20 Ma (Sano et al., 2006a) was used as the standard apatite. The average U concentration of PRAP is 196 ppm (Sasada et al., 1997). The U Pb dating was conducted using a NanoSIMS 50 installed at the Atmosphere and Ocean Research Institute of The University of Tokyo. An approx. 10 na 16 O primary ion beam with a spot diameter of approx. 15 µm is focused on the sample surface. Each spot is sputtered preliminarily for 5 min to eliminate surface residual contaminants. Positive secondary ions are extracted with an accelerating voltage of 8 kv. For 238 U 206 Pb dating, 43 Ca +, 204 Pb +, 206 Pb +, 238 U 16 O + and 238 U 16 O 2 + are collected simultaneously with a dual-collector-combined multi collection system. The observed background and 204 Pb count rates are, respectively, cps and cps. Although 204 Pb abundance of the phosphates is so low that identification of the 204 Pb peak on mass spectra images is difficult, no isobaric interference was found in this mass range or the mass range over 206 Pb and 207 Pb with mass resolution of approx at 1% peak height (Sano et al., 2006b). One analysis session takes 10 min to yield statistically sufficient counts. For SIMS measurements, secondary U ions are extracted mainly as monoxide and dioxide, although Pb is emitted almost entirely as atomic ions. Calibrations must derive true U/Pb ratios of samples from the secondary ion counts. For SHRIMP measurements, 238 U/ 206 Pb ratios are calculated from an empirical power law relation between 206 Pb + / 238 U + and 238 U 16 O + / 238 U + ratios (Hinthorne et al., 1979; Williams, 1998; Sano et al., 1999). Regarding NanoSIMS, the relative ratios of secondary UO 2 + :UO + :U + for phosphates are around 10:10:1 (Sano et al., 2006b). Empirically, the relation between 206 Pb + / 238 U 16 O + and 238 U 16 O 2 + / 238 U +16 O + ratios is useful for calibration. We applied a quadratic relation that was derived originally for zircon dating (Takahata et al., 2008). ( 206 Pb + / 238 U 16 O + ) = a ( 238 U 16 O 2 + / 238 U 16 O + ) 2 + b.(1) In that equation, a and b are constants determined by correlation between 206 Pb + / 238 U 16 O + and 238 U 16 O 2 + / 238 U 16 O + of the standard. More details about the calculation are given elsewhere (Takahata et al., 2008). Application of this method to terrestrial phosphate is currently underway (Sano et al., 2014). Uranium concentrations of analyzed spots are obtained by comparing the measured 238 UO + / 43 Ca + ratios of the sample against those of the standard (statistical error obtained by repeated analyses is approx. 30% at the 2-sigma level). After 238 U 206 Pb measurements, 207 Pb 206 Pb ages on the same spots were determined using single-collector mode, where the magnet was cyclically peak-stepped through 204 Pb +, 206 Pb +, and 207 Pb +. That process required approx. 1 h, the sum of 100 cycles and waiting time, to attain statistically sufficient counts. The pit depth was possibly less than 3 µm after the measurement, which is markedly smaller than the spot diameter (approx. 15 µm). The most stable and best analytical data were obtained in the condition (Ireland, 2004). RESULTS Eleven spots on Grain #1 were analyzed, as were seven spots each on Grains #2 and #3 (Fig. 1b and Supplementary Table S2). Results obtained for U concentrations 238 U/ 206 Pb, 207 Pb/ 206 Pb, and 204 Pb/ 206 Pb ratios are presented in Table S2. The 238 U/ 206 Pb ratios of the samples were calculated through the described quadratic equation (1) with the best-fit values of constants a and b of 0.14 ± 0.02 and 0.04 ± 0.03 (Supplementary Fig. S1). A few data (1-1 and 2-3) were discarded because the pit overlapped the cracks of the merrillite grains. 238 U 206 Pb, 207 Pb 206 Pb, and total U/Pb isochron ages of multi-grain We first calculated the multi-grain age, a conven- NanoSIMS U Pb dating of ALH

4 (a) (b) Fig. 1. (a) Backscattered electron images of two sections of ALH with phosphate locations as circles. (b) Analyzed merrillite grains with spot numbers. Rough contour images of U concentrations (ppm) are also overprinted. Pg, plagioclase glass; Opx, orthopyroxene; Mer, merrillite. tional way where all grains data are combined to determine a single isochron age. This age estimate is based on the idea that all grains were formed simultaneously. The 23 data points obtained from the three grains show correlations in both 238U 206Pb and 207Pb 206 Pb inverse plots (Figs. 2a and 2b). Based on least-squares fitting using the York method (York, 1969), both 238U 206Pb and 207Pb 206 Pb isochron ages were determined (Table 2). The calculations were made using isoplot3 (Ludwig, 2003). The 238 U 206Pb and 207Pb 206Pb ages were, respectively, 3850 ± 170 Ma (2 σ, MSWD = 2.1) and 4002 ± 52 Ma (2σ, MSWD = 1.3). The two ages are concordant at approx. 4.0 Ga within uncertainties. Calculation of three426 M. Koike et al. dimensional (3D) total U/Pb isochron age requires 238U/ 206 Pb, 207Pb/206Pb, and 204Pb/ 206Pb ratios (Wendt, 1989). Two datasets of 204Pb/ 206Pb ratios were used (Table S2). One is derived from the multi-collector mode. The other derives from the single mode with magnet scanning. We calculated the weighted mean average of 204Pb/ 206Pb ratios when considering the error and used the value for 3D total U/Pb isochron. Concordia-constrained linear regression shows the age of 3990 ± 160 Ma (95% confidence limit, MSWD = 1.9 in Table 2 and Supplementary Fig. S2), although the common Pb plane intercept gives no precise number because the error is large.

5 a large error, our age results agree around 4.0 Ga, indicating that the U Pb system in the Grain #1 is probably concordant. The 3D total U/Pb isochron ages are, respectively, 3830 ± 470 Ma (95% confidence limit, MSWD = 1.2), 4220 ± 180 Ma (95% confidence limit, MSWD = 1.2), and 3770 ± 540 Ma (95% confidence limit, MSWD = 0.21) for Grains #1, #2, and #3. They are mutually consistent within the allocated error. Concentrations of uranium Variations of U concentrations are ppm for three grains. The individual grains have U heterogeneity of a factor of approx. 2 (see Fig. 1b and Table S1). Compared to terrestrial apatite, for which typical U concentrations vary from a few parts per million to approx. 200 ppm (Sano et al., 1999), the concentrations of U in the observed merrillite grains are significantly lower and show smaller variations. The variations of U concentrations observed in individual grains are even smaller than those of grain-to-grain differences reported previously for apatite and merrillite in ALH (Terada et al., 2003). Although the variation is rather small, the observed distributions of U contents are not random but instead show some trend. The U-rich regions are, respectively, on the right side in Grain #1, the lower area in Grain #2, and the upper side in Grain #3. No simple relation exists between the U content and mineral close to the phosphate. For example, a higher U concentration is located adjacent to plagioclase glass in Grain #1, although it is lower in Grain #2. Fig. 2. Correlations between 204 Pb/ 206 Pb and 238 U/ 206 Pb ratios (a), and 204 Pb/ 206 Pb and 207 Pb/ 206 Pb ratios (b) of the three merrillite grains. Solid lines are regression lines calculated with Isoplot3 (Ludwig, 2003). Dashed curves are error envelopes. Previous U Pb study of phosphates (Terada et al., 2003) and the isotopic ratios of terrestrial common Pb (Stacey and Kramers, 1975) and initial Pb estimated for ALH (Bouvier et al., 2009) are also shown. 238 U 206 Pb, 207 Pb 206 Pb and total U/Pb isochron ages of single grains The 238 U 206 Pb and 207 Pb 206 Pb isochrons for the three individual grains are calculated separately (Supplementary Fig. S3, Table 2). The calculated 238 U 206 Pb ages are, respectively, 3970 ± 460 Ma (2σ, MSWD = 3.7), 3610 ± 840 Ma (2σ, MSWD = 0.63), and 3550 ± 860 Ma (2σ, MSWD = 0.32) for Grains #1, #2, and #3. The 207 Pb 206 Pb ages of three grains are also determined respectively as 4010 ± 180 Ma (2σ, MSWD = 2.2), 3760 ± 530 Ma (2σ, MSWD = 0.35), and 3920 ± 160 Ma (2σ, MSWD = 0.16). Although 238 U 206 Pb ages of Grains #2 and #3 have DISCUSSION Evaluations of NanoSIMS U Pb dating Observed multi-grain 238 U 206 Pb and 207 Pb 206 Pb ages by NanoSIMS are concordant: approx. 4.0 Ga within uncertainty, which agrees well with the previous SHRIMP study: 3700 ± 440 Ma for 238 U 206 Pb and 4022 ± 96 Ma for 207 Pb 206 Pb in Table 2 (Terada et al., 2003). Furthermore, the errors of our results (170 Ma and 52 Ma for the 238 U 206 Pb and 207 Pb 206 Pb ages, respectively) are smaller than those of the previous work (Table 2), suggesting that NanoSIMS U Pb dating of phosphates with low U abundance can provide accurate and reasonably precise age information. A few differences exist in 204 Pb/ 206 Pb ratios between datasets of 238 U 206 Pb and 207 Pb 206 Pb measurements. For example, in Grain #1, the ratios of spots 1 4 are discrepant, although the other spots are mutually consistent within experimental error. Similar features are observed in Grains #2 and #3. If a discrepancy results from considerable contaminations of terrestrial common Pb from the surface, all spots might show a similar tendency, i.e., 204 Pb/ 206 Pb ratios of multi-collector mode (shallow) are NanoSIMS U Pb dating of ALH

6 Fig. 3. Calculated relations between closure temperatures of U and Pb in phosphate and cooling rates of the system. Diffusion radii are 10 µm (solid curve) and 50 µm (dashed curve), which respectively present intra-grain diffusion and grain-tograin diffusion. Peak metamorphic temperature between 900 C and 1200 C is shown as shaded area. larger than those of single (deep). However, such is not the case. They differ spot by spot (Table S2), although a general tendency is apparent: the 204 Pb/ 206 Pb ratios are somewhat higher in 207 Pb 206 Pb measurements than in 238 U 206 Pb measurements (see Fig. S3). The origin of common Pb is more complicated, requiring some other explanation than simple terrestrial contamination on the sample surface. We again calculated the 3D total U/Pb isochron age based on the results obtained in this work (Table S1), together with those reported by Terada et al. (2003). Concordia-constrained linear regression using 35 data gives the isochron age of 4015 ± 60 Ma (95% confidence limit, MSWD = 2.2). This ion microprobe age is consistent with the Pb Pb silicate mineral age of 4074 ± 99 Ma obtained using the MC-ICP-MS measurement of strong leachates and orthopyroxene residue of ALH (Bouvier et al., 2009). It is noteworthy that threecomponent mixing of a terrestrial contaminant, common Martian Pb, and radiogenic Pb are not expected to form a linear array in the 238 U/ 206 Pb 207 Pb/ 206 Pb 204 Pb/ 206 Pb diagram (Wendt, 1989). A simple mixing relation exists between radiogenic and common Pb in this work. Common-Pb plane intercept in the 3D diagram provides the end member with the 206 Pb/ 204 Pb = 13.1 ± 6.2 (2σ) and 207 Pb/ 204 Pb = 13.4 ± 5.0 (2σ) with the 207 Pb/ 206 Pb ratio of These initial ratios are less radiogenic than those of the present terrestrial common Pb (Stacey and Kramers, 1975), even though the errors are large. Borg et al. (2005) reported that the Martian meteorites appear to be contaminated by Martian surface Pb characterized by a 207 Pb/ 206 Pb ratio of at least one. They further suggested that the surface Pb is prevalent in impact glassrich mineral fractions. Borg and Drake (2005) reported that the surface Pb was derived from the sulfur-rich soil in most aqueous systems. Observed common Pb in this work might be related to the surface Pb. Isochrons calculated for individual grains have larger errors because of low U concentrations and small variations as well as the limited number of spots analyzed. However, their mean ages are consistent with the multigrain 238 U 206 Pb and 207 Pb 206 Pb ages, indicating the U Pb system within an individual grain is probably concordant at approx. 4.0 Ga. These results suggest that analyses of single grains provide useful U Pb age information. Regarding the age precision, it can be regarded as reasonable considering the limited sample size ( µm diameter) and small variations of U concentrations (Table S2). For all cases in this study, the 207 Pb 206 Pb ages can be determined more precisely than the 238 U 206 Pb age because of the old age of approx. 4.0 Ga and the systematics of U Pb Concordia curve where the time variation of 207 Pb/ 206 Pb ratio is large at 3 4 Ga. Based on these results, it is noteworthy that the U Pb dating of the single phosphate grains by NanoSIMS might provide chronological information of much smaller spatial resolutions (Terada et al., 2014) than previously studied. Interpretation of U Pb ages Although the U Pb system shows concordance for multi-grain isochron, one can not simply regard them as igneous age of the host rock. From a number of textural, chemical and isotopic studies, ALH is regarded as having suffered severe impact shock, consequent annealing and aqueous alterations for several times (summarized in Treiman (1995, 1998)). The metamorphic temperature of the ancient impact heating at Ga (Ash et al., 1996; Turner et al., 1997; Bogard and Garrison, 1999; Cassata et al., 2010) is estimated as >875 C using mineral geothermometers (Treiman, 1995), but it is not above 1200 C because no evidence exists of wholesale melting (Treiman, 1998). Therefore, the peak metamorphic temperature might be between about 900 C and 1200 C. Phosphates in ALH are known as igneous by a textural, chemical and oxygen isotopic study (Greenwood et al., 2003). We examined our samples by the texture (Fig. 1b). Numerous cracks in merrillite minerals are apparently connected with those in orthopyroxene, although plagioclase glasses have no cracks. The phosphate might be generated simultaneously of orthopyroxene or at least before the shock heating formation of plagioclase glass, suggesting its igneous origin. Burger et al. (2012) reported that a positive relation exists between Mg# and Na contents in igneous Martian merrillite. Our chemistry data (Table S1) are consistent with the trend. The rare earth element (REE) pattern of whole-rock ALH M. Koike et al.

7 is similar to those of other igneous Martian meteorites (Meyer, 2012), again suggesting merrillite s igneous origin because the phosphate is a major carrier of REE elements. The 207 Pb 206 Pb isochron age of 4002 ± 52 Ma in this work might show an igneous formation event. Several studies have identified later shock events for ALH (e.g., Treiman, 1998; Cassata et al., 2010). The 207 Pb 206 Pb system in the phosphates might have not been disturbed during those later events, although some 238 U 206 Pb isochrons of the individual grains show a slightly younger age with large uncertainty (Table 2). An experimental shock and heating study (Gaffney et al., 2011) reported the weakness of U Pb system to impact heating compared to Rb Sr and Sm Nd systems. Further discussion related to the younger 238 U 206 Pb ages, however, requires additional analyses of phosphate grains in the meteorite. Distributions of U in the individual grains are also informative, although the variations observed in the grain are not so large (Fig. 1b and Table S2). The success of the single grain dating is because of these U heterogeneities and consequent production of radiogenic Pb. From diffusion behaviors of U and Pb, the thermal history of the host rock can be discussed. More specifically, the relation between peak temperatures and cooling rates for heating events of ALH can be regulated by distributions of U and Pb in the observed grains. The closure temperature (T c ) of a certain ion species in a mineral crystal can be approximated from the following equation (Dodson, 1973). E / R Tc =. 2 2 ln ART D / a / ET ( c ( 0 ) ) ( 2) Therein, R is the gas constant, A stands for the geometrical constant of 55 for spherical shape, T denotes the cooling rate, D 0 represents the pre-exponential factor of diffusion coefficient, E is the activation energy, and a is the diffusion radius. Previous study of Shergotty phosphates used this relation for Pb diffusion and discussed the closure temperature of U Pb age (Sano et al., 2000). Diffusion parameters of U or Pb in merrillite are not well known. However, it might be reasonable to apply those of apatite in this case because previous data of U Th Pb analyses of apatite and merrillite plot on the same regression lines (Sano et al., 2000; Terada et al., 2003), which indicates diffusion behaviors of U Th Pb system are not significantly different between the two. Therefore, we apply apatite parameters to our results here. Experimentally acquired D 0 and E values in apatite crystal are reported, respectively, as D 0 = cm 2 /s and E = 231 kj/mol for Pb and D 0 = cm 2 /s and E = 394 kj/ mol for U (Cherniak et al., 1991; Cherniak, 2005). For diffusion scale a, we used 10 µm (approximate spot diameter) and 50 µm (grain size), respectively, for intragrain diffusion and grain-to-grain diffusion. Figure 3 shows the calculated closure temperatures with various cooling rates of the system. The peak temperature of metamorphic event ( C) is shown in the diagram. Based on the heterogeneous U content distributions in each grain, one might say that U in the phosphate has never been lost or homogenized since its igneous crystallization. These observations suggest that the cooling rate is lower than 0.4 C/year if the maximum peak temperature of 1200 C is adopted in the event. The rate is consistent with the rapid cooling system of C/year in Shergotty phosphates (Sano et al., 2000). No evidence exists of partial melt of phosphate minerals in texture (Fig. 2b). Therefore, the peak temperature might be less than 1000 C. However even in this condition, Pb might be partly lost by diffusion from the grain. Then 238 U 206 Pb system can be reset somehow. After the approx. 4.0 Ga event, the temperature might not have exceeded approx. 600 C except for localized heating (inferred from the Ar Ar study; Cassata et al., 2010) because the 207 Pb 206 Pb system has been closed for approx. 4.0 Ga, even with the rapid cooling rate of Shergotty. If the cooling rate is on the order of 10 5 C/year, similar to the terrestrial sample assumed by Cherniak et al. (1991), which is much smaller than Shergotty, the sample might not have been heated more than 500 C at any event since 4.0 Ga because Pb is not homogenized at the 50 µm scale (see Fig. 3). CONCLUSIONS The U Pb dating of merrillite grains in ALH was conducted using a NanoSIMS. For both multi-grain dating and single-grain dating, obtained U Pb ages and Pb Pb ages are approx. 4.0 Ga. They also agree with the previous SHRIMP U Th Pb work, indicating that our dating method reveals a local history of the sample. Furthermore, the observed heterogeneous distributions of U content within the grains might have been preserved since its igneous formation. Calculation of the closure temperatures of U and Pb in phosphate crystal indicates that the cooling rate might be consistent with the rapid value of C/year. The U Pb system remained almost closed since approx. 4.0 Ga, suggesting that the temperature has not exceeded 500 C since then. Acknowledgments Useful advice related to experimental techniques and chronological studies from Prof. H. Hiyagon, Dr. T. Iizuka, and Dr. K Ichimura is greatly appreciated. We are also thankful to Dr. K R. Ludwig for providing computer software: Isoplot/Ex. We are most grateful to Dr. K. Misawa and Dr. D. J. Cherniak for valuable comments, and to Dr. Terada for kindly handling the manuscript from previous versions. This NanoSIMS U Pb dating of ALH

8 work was partly supported by Grant-in-Aid for Scientific Research Program No from the Ministry of Education, Science and Culture of Japan. REFERENCES Agee, C. B., Wilson, N. V., McCubbin, F. M., Ziegler, K., Polyak, V. J., Sharp, Z. D., Asmerom, Y., Nunn, M. H., Shaheen, R., Thiemens, M. H., Steele, A., Fogel, M. L., Bowden, R., Glamoclija, M., Zhang, Z. and Elardo, S. M. (2013) Breccia Northwest Africa 7034 Unique Meteorite from Early Amazonian Mars: Water-Rich Basaltic. Science 339, Ash, R. D., Knott, S. F. and Turner, G. (1996) A 4-Gyr shock age for a martian meteorite and implications for the cratering history of Mars. Nature 380, Beard, B. L., Ludois, J. M., Lapen, T. J. and Johnson, C. M. (2013) Pre-4.0 billion year weathering on Mars constrained by Rb Sr geochronology on meteorite ALH Earth Planet. Sci. Lett. 361, Bogard, D. D. and Garrison, D. H. (1999) Argon-39 - argon-40 ages and trapped argon in Matian shergottites, Chassigny, and Allan Hills Meteorit. Planet. Sci. 34, Borg, L. E. and Drake, M. J. (2005) A review of meteorite evidence for the timing of magmatism and surface or nearsurface liquid water on Mars. J. Geophys. Res. 110, E12S03. Borg, L. E., Connelly, J. N., Nyquist, L. E., Shih, C. Y., Weismann, H. and Reese, Y. (1999) The age of the carbonates in Martian meteorite ALH Science 286, Borg, L. E., Edmunson, J. E. and Asmerom, Y. (2005) Constraints on the U Pb isotopic systematics of Mars inferred from a combined U Pb, Rb Sr, and Sm Nd isotopic study of the Martian meteorite Zagami. Geochim. Cosmochim. Acta 69, Bouvier, A., Blichert-Toft, J. and Albarede, F. (2009) Martian meteorite chronology and the evolution of the interior of Mars. Earth Planet. Sci. Lett. 280, Burger, P. V., Shearer, C. K., Papike, J. J. and McCubbin, F. M. (2012) Crystal chemistry of merrillite in martian basalts and its significance to interpreting basalt petrogenesis. Abstract 1178, LPSC. Cassata, W. S., Shuster, D. L., Renne, P. R. and Weiss, B. P. (2010) Evidence for shock heating and constraints on Martian surface temperatures revealed by 40 Ar/ 39 Ar thermochronometry of Martian meteorites. Geochim. Cosmochim. Acta 74, Chen, J. H. and Wasserburg, G. J. (1986) Formation ages and evolution of Shergotty and its parent planet from U Th Pb systematics. Geochim. Cosmochim. Acta 50, Cherniak, D. J. (2005) Uranium and manganese diffusion in apatite. Chem. Geol. 219, Cherniak, D. J., Lanford, W. A. and Ryerson, F. J. (1991) Lead diffusion in apatite and zircon using implantation and Rutherford Backscattering techniques. Geochim. Cosmochim. Acta 55, Dodson, M. H. (1973) Closure temperature in cooling geochronological and petrological systems. Contrib. Mineral. Petrol. 40, Gaffney, A. M., Borg, L. E., Asmerom, Y., Shearer, C. K. and Burger, P. V. (2011) Disturbance of isotope systematics during experimental shock and thermal metamorphism of a lunar basalt with implications for Martian meteorite chronology. Meteorit. Planet. Sci. 46, Greenwood, J. P., Blake, R. E. and Coath, C. D. (2003) Ion microprobe measurements of 18 O/ 16 O ratios of phosphate mineral in the Martian meteorites ALH and Los Angeles. Geochim. Cosmochim. Acta 67, Halevy, I., Fischer, W. W. and Eiler, J. M. (2011) Carbonates in the Martian meteorite Allan Hills formed at 18 ± 4 C in a near-surface aqueous environment. Proc. Nat. Acad. Sci. 108, Hinthorne, J. R., Andersen, C. A., Conrad, R. L. and Lovering, J. F. (1979) Single grain 207 Pb/ 206 Pb and U/Pb age determinations with a 10 µm spatial resolution using the ion microprobe mass analyzer (IMMA). Chem. Geol Humayun, M., Nemchin, A., Zanda, B., Hewins, R. H., Grange, M., Kennedy, A., Lorand, J.-P., Göpel, C., Fieni, C., Pont, S. and Deldicque, D. (2013) Origin and age of the earliest Martian crust from meteorite NWA7533. Nature 503, Ilg, S., Jessberger, E. K. and El Goresy, A. (1997) Argon-40/ argon-39 laser extraction dating of individual maskelynites in SNC pyroxenite Allan Hills (abstract). Meteorit. Planet. Sci. 32, A65. Ireland, T. R. (2004) SIMS measurement of stable isotopes. Handbook of Stable Isotope Analytical Techniques (de Groot, P. A., ed.), Jagoutz, E., Sorowka, A., Vogel, J. D. and Wanke, H. (1994) ALH 84001: Alien or progenitor of the SNC family? Meteorit. Planet. Sci. 29, Jagoutz, E., Bowring, S., Jotter, R. and Dreibus, G. (2009) New U Th Pb data on SNC meteorite ALH (abstract). 40th Lunar Planet. Sci. Conf Knott, S. F., Ash, R. D. and Turner, G., (1996) 40 Ar 39 Ar dating of ALH84001: evidence for the early bombardment on Mars (abstract). 26th Lunar Planet. Sci. Conf Lan, Z. W., Guo, C. L., Yang, Y. N., Liu, Y. and Tang, G. Q. (2013) Monazaite and xenotime U Th Pb geochronology by ion microprobe dating highly fractionated granites at Xihuashan tungsten mine, SE China. Contrib. Mineral. Petrol. 166, Lapen, T. J., Righter, M., Brandon, A. D., Debaille, V., Beard, B. L., Shafer, J. T. and Peslier, A. H. (2010) A younger age for ALH84001 and its geochemical link to shergottite Sources in Mars. Science 328, Li, Q. L., Li, X. H., Wu, F.-Y., Yin, Q.-Z., Ye, H.-M., Liu, Y., Tang, G.-Q. and Zhang, C.-L. (2012) In-situ SIMS U Pb dating of phanerozoic apatite with low U and high common Pb. Gondwana Res. 21, Ludwig, K. R. (2003) Users manual for Isoplot/Ex: A geochronological toolkit for Microsoft Excel. Geochronology Center, Berkley, Special Publication No. 4. Meyer, C. (2012) ALH84001 in Mars meteorite compendium, NASA-ARES JSC. Available at antmet/mmc/ Minster, J. F., Birck, J. L. and Allegre, C. J. (1982) Absolute age of formation of chondrites studied by the Rb-87 Sr M. Koike et al.

9 method. Nature 300, Nyquist, L. E. and Shih, C. Y. (2013) Peering through a martian veil: ALH Sm Nd age revisited (abstract). 44th Lunar Planet. Sci. Conf Nyquist, L. E., Bansal, B., Wiesmann, H. and Shih, C. Y. (1995) Martians young and old: Zagami and ALH (abstract). Proceedings, 26th Lunar Planet. Sci. Conf Nyquist, L. E., Bogard, D. D., Shih, C., Greshake, A., Stoffler, D. and Eugster, O. (2001) Ages and geologic histories of martian meteorites. Space Sci. Rev. 96, Sano, Y., Oyama, T., Terada, K. and Hidaka, H. (1999) Ion microprobe U Pb dating of apatite. Chem. Geol. 153, Sano, Y., Terada, K., Takeno, S., Taylor, L. A. and McSween, H. Y., Jr. (2000) Ion microprobe uranium thorium-lead dating of Shergotty phosphates. Meteorit. Planet. Sci. 35, Sano, Y., Terada, K., Ly, C. V. and Park, E. J. (2006a) Ion microprobe U Pb dating of a dinosaur tooth. Geochem. J. 40, Sano, Y., Takahata, N., Tsutsumi, Y. and Miyamoto, T. (2006b) Ion microprobe U Pb dating of monazite with five micrometer spatial resolution. Geochem. J. 40, Sano, Y., Toyoshima, K., Ishida, A., Shiai, K., Takahata, N., Sato, T. and Komiya, T. (2014) Ion microprobe U Pb dating and Sr isotope measurement of a protoconodont. J. Asian Earth Sci. (in press). Sasada, T., Hiyagon, H., Bell, K. and Ebihara, M. (1997) Mantle-derived noble gases in carbonatites. Geochim. Cosmochim. Acta 61, Stacey, J. S. and Kramers, J. D. (1975) Approximation of terrestrial lead isotope evolution by a two-stage model. Earth Planet. Sci. 26, Sugiura, N. and Hoshino, H. (2000) Hydrogen-isotopic compositions in Allan Hills and the evolution of the Martian atmosphere. Meteorit. Planet. Sci. 35, Takahata, N., Tsutsumi, Y. and Sano, Y. (2008) Ion microprobe U Pb dating of zircon with a 15 micrometer spatial resolution using NanoSIMS. Gondwana Res. 14, Terada, K. and Sano, Y. (2004) Ion microprobe U Th Pb dating and REE analyses of phosphates in the nakhlites Lafayette and Yamayo / Meteorit. Planet. Sci. 39, Terada, K., Monde, T. and Sano, Y. (2003) Ion microprobe U Th Pb dating of phosphates in martian meteorite ALH Meteorit. Planet. Sci. 38, Terada, K., Sano, Y., Takahata, N., Ishida, A., Tsuchiyama, A., Nakamura, T., Noguchi, T., Karouji, Y., Uesugi, M., Yada, T., Nakabayashi, M., Kamioka, M., Fukuda, K. and Nagahara, H. (2014) Thermal and impact histories of Itokawa recorded in Hayabusa particles. Nature Communications (in revision). Treiman, A. H. (1995) A petrographic history of Martian meteorite ALH84001: two shocks and an ancient age. Meteorit. Planet. Sci. 30, Treiman, A. H. (1998) The history of Allan Hills revised: Multiple shock events. Meteorit. Planet. Sci. 33, Turner, G., Knott, S. F., Ash, R. D. and Gilmour, J. D. (1997) Ar Ar chronology of the Martian meteorite ALH84001: Evidence for the timing of the early bombardment of Mars. Geochim. Cosmochim. Acta 61, Wadhwa, M. and Lugmair, G. W. (1996) The formation age of carbonates in ALH Meteorit. Planet. Sci. 31, A145. Wadhwa, M. and Lugmair, G. W. (1997) The controversy of young vs. old age of formation of carbonates in Allan Hills (abs). Conference on Early Mars: Geologic and Hydrologic Evolution, Physical and Chemical Environments, and the Implications for Life, 79 81, Lunar and Planetary Institute, Houston, TX. Wendt, I. (1989) Geometric considerations of the threedimensional U/Pb data presentation. Earth Planet. Sci. Lett. 94, Williams, I. S. (1998) U Th Pb geochronology by ion microprobe. Rev. Economic Geol. 7 (McKibben, M. A., Shanks, W. C., III and Ridley, W. I., eds.), Yin, Q.-Z., Herd, C. D. K., Zhou, Q., Li, X.-H., Wu, F.-Y., Li, Q.-L., Tang, G.-Q. and McCoy, T. J. (2014) Reply to comment on Geochronology of the Martian meteorite Zagami revealed by U Pb ion probe dating of accessory minerals. Earth Planet. Sci. Lett. 385, York, D. (1969) Least squares fitting of a straight line with correlated errors. Earth Planet. Sci. Lett. 5, Zhou, Q., Herd, C. D. K., Yin, Q.-Z., Li, X.-H., Wu, F.-Y., Li, Q.-L., Liu, Y., Tang, G.-Q. and McCoy, T. J. (2013) Geochronology of the Martian meteorite Zagami revealed by U Pb ion probe dating of accessory minerals. Earth Planet. Sci. Lett. 374, SUPPLEMENTARY MATERIALS URL ( data/48/ms319.pdf) Figures S1 to S3 Tables S1 and S2 NanoSIMS U Pb dating of ALH