1 Noble Gas Chronology Of

Noble Gas Chronology of Meteorites: Brachinites, Ureilites, and Chelyabinsk

Item Type text; Electronic Dissertation

Authors Beard, Sky

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 09/10/2021 07:41:47

Link to Item http://hdl.handle.net/10150/631308 1

NOBLE GAS CHRONOLOGY OF METEORITES: BRACHINITES, UREILITES, AND CHELYABINSK

Sky Beard

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF PLANETARY SCIENCES

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2018 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by SkyBeard, titled Noble Gas Chronology of Meteorites:Brachinites, Ureilites, and Chelyabinsk and recommend that it be accepted as fulfillingthe dissertation requirement for the Degree of Doctor of Philosophy.

Date: (9/24/2018)

Date: (9/2412018)

d 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of the requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that an accurate acknowledgement of the source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.

SIGNED: Sky Beard

ACKNOWLEDGMENTS

Mom, you always support whatever I want to do and I cannot tell you how much I appreciate you. You want everyone to be happy and sacrifice so much. Please remember to take care of yourself, too! I miss making you laugh. Lee, thanks for always being willing to do whatever is needed. I am so happy you have a career you love and am proud of how hard you are working. Thanks for helping the family when needed and thank you for our fun talks that help make the distance from home seem shorter. Misty, I know you are working so hard to make things better for your family. Do not give up. I admire that and know your effort will be rewarded. You are so fun and sweet, I miss you! Meadow, I always appreciate your kindness and open invitations to visit you. You are so accommodating and always make my visits feel like home. Brooke, I am happy to see all the great changes you have made in your life and to see you achieve your goals. Thank you for helping others as much as you do and for being a positive influence. Prarie, thank you for always welcoming me to stay at your home when I visit, coloring amazing pictures, and ensuring that I’m fed all our homemade favorites! I enjoy our laughs together and look forward to visiting soon. To all of you, thank you for your encouragement! I love you.

Tim Swindle: thank you for being the most wonderful advisor I could have imagined. I am very thankful for the opportunity to learn from you as a scientist and friend. I have learned so much under your guidance and will always be grateful. Tom Zega: thank you for always asking such good questions and making each conversation a fun learning experience. Mike Nolan: thank you for filling in last minute to serve on my committee before even knowing me. Your positivity always is always encouraging. Vic Baker: I always enjoy our chats, work related or not. Your anecdotes and philosophy are always enlightening and refreshing. Barbara Cohen: thank you for your helpful discussions about experimental techniques, helping with analysis, and taking the extra responsibility of being on my committee. Ken Domanik: You are always happy to help me understand microprobe related matters. Your enthusiasm is inspiring. Clark Isachsen: Thank you for taking care of whatever needed to be tested or fixed in the lab. I would be still be in the lab changing gaskets and leak testing if it was not for your help and expertise. You have always been enthusiastic about solving problems and I have learned a lot from you. Thank you all for your patience and understanding.

Agli miei amici, la famiglia Vimercati: Michele, Mary Cruz, Lara, y Ylena, grazie per tutto. Avete avuto un profondo impatto sulla mia vita e ti sono molto grato. Spero che stiate tutti bene, mi mancate a caro prezzo. Vi voglio bene! Asia, grazie mille pet il tuo disegno dello spettrometro, ma piu importante la tua amicizia e cura. Sawadi nong Ploy. Khap khun thuk yang. Khun nang lae narak mak, lae guan teen mak mak.

TABLE OF CONTENTS LIST OF FIGURES...... 7

LIST OF TABLES...... 10

ABSTRACT...... 11

CHAPTER 1: INTRODUCTION...... 13 1.1 Ordinary Chondrites………………………………………..………………………………….……….. 13 1.2 Brachinites……………………….…………………………………………………………………………..18 1.3 Ureilites………………………..……………………………………………………………………..………. 22 1.4 Thesis Structure and Logistics………………………..………………..………………………………24

CHAPTER 2: TECHNIQUES AND SAMPLES...... 26 2.1 Argon-Argon Background...... 26 2.2 Ar-Ar Experimental Design and Data Aquisition...... 36 2.3 Cosmic-Ray Exposure Background...... 41 2.4 CRE Experimental Design and Data Aquisition...... 46 2.5 Oxygen Isotopes ……………………………………………………………………………………………52 2.6 Samples………………………………………………………………..…………………………….…………54

CHAPTER 3: AR-AR DATA REDUCTION EXAMPLE...... 63 3.1 Gas Extraction...... 63 3.2 Data Corrections...... 64

CHAPTER 4: AR-AR OF CHELYABINSK...... 73 4.1 Introduction...... 73 4.2 Ar-Ar Analysis...... 74 4.3 Ar-Ar Results of Chelyabinsk...... 76 4.4 Impact Age Discussion...... 87 4.5 Conclusions...... 97

CHAPTER 5: AR-AR DIFFICULTIES WITH BRACHINITES...... 99 5.1 Introduction...... 99 5.2 Analysis...... 99 5.3 Brachinite Ar-Ar Results...... 101 5.4 Ar-Ar Discussion...... 105 5.5 Brachinite Argon Conclusion...... 109

CHAPTER 6: COSMIC-RAY EXPOSURE AGES OF UREILITES...... 110 6.1 Introduction...... 110 6.2 Methods...... 114 6.3 Ureilite Exposure Age Results...... 118 6.4 Discussion...... 122 6.5 Conclusion...... 125 6

CHAPTER 7: COSMIC-RAY EXPOSURE AGES OF BRACHINITES AND OTHER ACHONDRITES...... 127 7.1 Introduction...... 127 7.2 Brachinite, Brachinite-like, and Ungrouped Classifications...... 132 7.3 UMA Noble Gases...... 137 7.4 Experiment and Methods...... 139 7.5 Cosmogenic Results...... 141 7.6 Discussion...... 145 7.7 Conclusions...... 154

CHAPTER 8: CONCLUSIONS...... 156

SUPPLEMENTARY MATERIALS...... 158

REFERENCES...... 173

LIST OF FIGURES

1.1 Schematic of meteorite classification...... 14 1.2 Ar-Ar ages of LL chondrites……………...... 15 1.3 Neon cosmic-ray exposure ages of LL chondrites ...... 15 1.4 Chelyabinsk hand sample, entry, isotopic, and elemental characteristics …...... 16 1.5 Sm-Nd isochron of Chelyabinsk...... 17 1.6 Pb-Pb isochron of Chelyabinsk...... 17 1.7 Crossed polar photomicrographs of select samples……………...... 19 1.8 Highly siderophile element abundances of select samples ...... 19 1.9 Oxygen isotopes of brachinites and UMA…...... 21 1.10 Histogram of CRE ages of acapulcoites and lodranites...... 21 1.11 Olivine compositions in Fo for ureilites...... 22 1.12 Oxygen isotopes of achondrites...... 22 1.13 Neon CRE ages of ureilites and AhS...... 23 1.14 Oxygen isotopes vs Fo of ureilites ...... 24 2.1 Decay schematic of 40K...... 28 2.2 Mass spectrometer schematic ...... 30 2.3 Isochron and reverse isochron examples ...... 34 2.4 Extraction line schematic for Ar-Ar system ...... 37 2.5 Example of SAES getter used for purification...... 38 2.6 Extraction line for laser and furnace ...... 39 2.7 Production rate depends on nuclei……...... 43 2.8 Production rate depends on target chemistry ...... 44 2.9 Production rate models compared to empirical measurements ...... 44 2.10 Sample holder and extraction line for CRE experiments...... 50 2.11 Albatros mass spectrometer and extraction line for CRE experiments...... 52 2.12 Oxygen isotopes of main chondrite groups and schematic of behavior ...... 53 2.13 Oxygen isotope plot showing 17O of achondrite groups...... 54 2.14 NWA 595 Si map…………………..……………………...... 56 2.15 NWA 595 BSE image………………..………………...... 56 2.16 NWA 1500 X-ray map………….……..…………………...... 57 2.17 NWA 1500 BSE image………….……………………...... 58 2.18 NWA 6077 Mg-X-ray map…....……..…………………...... 59 2.19 NWA 7297 X-ray map……..…….…..…………………...... 60 2.20 NWA 7297 BSE image………………………………...... 60 2.21 RaS 309 BSE image…….……………………………...... 62 2.22 RaS 309 BSE image…………….……………………...... 62 3.1 Pumpout of isotopes in mass spectrometer during Ar-Ar analysis ...... 64 3.2 Magnetic centering and 40Ar values during hot blank procedure…….……...... 65 3.3 Blank values compared to blank-corrected sample values……..…….……...... 66 3.4 Example of uncorrected plateau ……………………………………………………………. 67 3.5 Isochron plot for all steps of SB5……...………………...…………………………………. 68 3.6 Isochron plot for 1000-1500 °C steps of SB5……...………………………………….……. 68 3.7 Reverse isochron plot for 300-500 °C steps of SB5……...….…………………..…………. 69 8

3.8 Corrected isochron plot for 300-500 °C steps of SB5, including trapped correction and age ...………………………………………………………………………………………………… 71 3.9 Corrected isochron plot for 950-1175°C steps of SB5, including trapped correction and age …………………………………………………………………………………………………... 71 3.10 Corrected plateau plot for Chelyabinsk SB5.…………..………....………………………. 72 4.1 Dark and light lithologies of Chelyabinsk with melt veins…………………………………. 73 4.2 Regression of 40Ar to t0………………………………...... 76 4.3 Isochron plots for 6 splits of Chelyabinsk, trapped correction, and age………...………….. 77 4.4 Raw plateau plot and K/Ca release profile for Chelyabinsk SB4 ………………………….. 78 4.5 Raw plateau plot and K/Ca release profile for Chelyabinsk SB1 ………….………………. 78 4.6 Fully corrected plateau plot for Chelyabinsk SB1 ……………………………...………….. 82 4.7 Fully corrected plateau plot for Chelyabinsk SB2 …………………...…………………….. 83 4.8 Fully corrected plateau plot for Chelyabinsk SB3 ……………………….……..………….. 83 4.9 Fully corrected plateau plot for Chelyabinsk CH1 ………..……………………………….. 84 4.10 Fully corrected plateau plots for Chelyabinsk MB,2 ...…………………………………… 84 4.11 Fully corrected plateau plots for Chelyabinsk MB,5 ………………………………...…… 86 4.12 Combined age distribution of Chelyabinsk compared with all LL impact distribution…... 93 4.13 Revised impact age distribution as a result of this work…………………...………...…… 94 4.14 Impact history of Chelyabinsk…………………….……………………………….……… 94 4.15 Ar-Ar Plateau plot for Itokawa..…………. ……………………………………….……… 95 4.16 Chelyabinsk CH2 apparent age model……………………………………………………. 98 5.1 Sample holder for laser samples..….………… …………………………………….……...100 5.2 Sample viewport attached to gas extraction line..……… ………… ……………….…..… 100 5.3 Corrected plateau plot for NWA 6077………………………..………………………….... 103 5.4 Corrected plateau plot for NWA 595……………………………………………………… 104 5.5 Corrected plateau plot for NWA 7297……….……….………………………………….... 104 5.6 Corrected plateau plot for NWA 1500…………….….…………………………………….105 5.7 Sample holder design and cracked sapphire cover……..…………….…………………… 106 5.8 Modified sample wrap designed for less sputtering………………………………………. 107 5.9 Modified wrap and holder designed for less sputtering………..…….……………………. 109 6.1 Pre-and-post-impact on monomict and polymict structured ureilite with possible ejection scenarios………………………………………………………………………………………...113 6.2 Relative probability plot of CRE ages of ureilites…………………...……………………. 119 6.3 Histogram of CRE ages of ureilites………………………………….……………………. 119 6.4 Ureilite Fo vs CRE age……………………………………………………….…...………. 121 6.5 Ureilite 17O vs CRE age………………………………………………………...………. 122 7.1 17O of achondrites with highlighted samples included in this work……………………. 129 7.2 Fe-Mn-Mg systematics for achondrites and schematic of igneous effects……..…………. 131 7.3 High siderophile element abundance patterns for brachinites and UMA……..……..……. 132 7.4 Fe-Mn-Mg systematics indicating a distinction between relatively oxidized and reduced samples………………………………………………………………………………………… 133 7.5 17O vs Fe/Mg of brachinite and brachinite-like samples show distinction…….……..… 135 7.6 Ne 3-Isotope plot of brachinites and UMA…………………………………………..…… 142 7.7 21Ne CRE ages of UMA………………………………………………………..………… 145 7.8 3He vs 21Ne CRE ages of UMA………………………………………………..………… 146 7.9 Relative probability plot of CRE ages of UMA…………………………………………… 147 9

7.10 17O vs 21Ne exposure ages of UMA………….…….……………………….………… 148 7.11 Fe/Mg vs 21Ne exposure ages of UMA………………….……………………………… 151

LIST OF TABLES

1.1 Chelyabinsk ages determined by various isotopic dating methods...... 17 2.1 Potassium decay constants... ……………...... 28 2.2 Important interfering reactions on Ca and K ...... 33 2.3 Production rates of cosmogenic He, Ne, and Ar...... 42 2.4 Average composition of olivine in wt.% used for shielding model…...... 48 2.5 Average composition of clinopyroxene in wt.% used for shielding model...... 48 2.6 Average composition of orthopyroxene in wt.% used for shielding model...... 49 2.7 Average composition of plagioclase in wt.% used for shielding model...... 49 2.8 Average composition of chromite in wt.% used for shielding model...... 49 2.9 Suppliers of brachinite and other ultramafic achondrite samples for this work...... 50 4.1 Chelyabinsk age summary from multiple dating methods...... 74 4.2 Trapped corrections and isochron ages for Chelyabinsk...... 80 4.3 SB1 Argon values and age by extraction step…………...... 80 4.4 Chelyabinsk argon-argon summary……………………...... 86 4.5 Age summary of Chelyabinsk…….……………………...... 96 6.1 Key data for the 39 ureilites used in this work………...... 116 6.2 Possible ureilite clusters……………………………………...... 121 7.1 ΔIW of brachinites and other achondrites.……….…...... 130 7.2 Conditions for brachinite, brachinite-like, and ungrouped achondrites……...... 136 7.3 Additional conditions for brachinite classification………………………...... 137 7.4 Compositional measurements from this work, in wt.%….…..……………...... 140 7.5 Cosmogenic abundances of samples in this work……..….………..………...... 143 7.6 Shielding parameters and 3He, 21Ne, and 38Ar CRE ages.…….………...... 144 7.7 Impact groupings for brachinites and brachinite-like samples…………………..…...…… 149 7.8 21Ne, 17O, and classification...... 153

ABSTRACT

Noble gas studies offer a glimpse into the history experienced by meteorites, their parent bodies, and the solar system. This thesis reports measurements and interpretations of noble gases of achondrites and the well-known chondrite, Chelyabinsk. Several geochronology dating methods have been used to measure the formation age and impact history of Chelyabinsk. The data from these measurements is generally not well defined and often contradictory. Ar-Ar measurements from this work show well defined impact ages ~30 Ma and ~2700 Ma recorded in different lithologies of Chelyabinsk, which help provide clarity to the complex history proposed by other studies. Chelyabinsk experienced 3-4 impacts after formation at ~4450 Ma, ~2700 Ma, ~30 Ma, and was exposed to cosmic rays ~1.5 Ma.

Cosmic-ray exposure (CRE) ages of ureilites are combined with magnesium numbers of olivine and oxygen isotopes to search for common impact events. This technique can also be used to investigate the heterogeneity of the body from which the samples originated. There are 39 ureilites included in this work, and although there is weak evidence of possible clusters, it is clear that most ureilites did not originate in one or two events on a homogeneous parent body.

A suite of ultra-mafic achondrites make up the brachinite and brachinite-like achondrites.

Assigning an origin to one or multiple parent bodies has proven to be difficult and confusing. This work includes a subset of brachinites, brachinite-like achondrites, and ungrouped achondrites that have some brachinite affinity in some way (oxygen isotopes or mineral abundances). The motivation of this work is to determine and compare their cosmic-ray exposure age distribution, which may indicate common impact and parent body relationships. Noble gases He, Ne, Ar, Kr, and Xe are analyzed to determine their cosmic-ray exposure ages, nominal gas retention ages and trapped compositions. The measured 3He and 21Ne are purely cosmogenic. The 38Ar has a 12 significant trapped component and calculation of the exposure age based on 38Ar is also sensitive to the amount of calcium in the sample, which is contained in minor phases. Brachinite-like samples in this study have three clusters of 21Ne exposure ages ~10Ma, ~26 Ma, and ~ 50 Ma, while the brachinites have possible groupings at ~10 Ma, ~26 Ma, ~40 Ma, and ~50 Ma. Three of the four brachinite clusters coincide with brachinite-like 21Ne ages and Δ17O values. The Δ17O values of the three older (~26 Ma, ~40 Ma, ~50 Ma) clusters are similar, while the younger cluster

(~10 Ma) is slightly more depleted in 17O and might suggest a different parent body or reflect heterogeneity in the same parent body. Three of the four clusters contain brachinite-like samples, and suggests that the brachinite and brachinite-like achondrite clusters may originate from the same primitive parent body. Some samples show significant abundances of trapped Kr and Xe of

Q composition (36Ar/132Kr ~ 90-150, 84Kr/132Xe ~ 1-8, and 129Xe/132Xe ~ 1.1-1.8) that must have survived thermal metamorphism, illustrating that the brachinites and brachinite-like achondrites are intermediate between chondrites and fully/more degassed differentiated achondrites.

Chapter 1: Introduction

1.1 Ordinary Chondrites

Meteorites can be classified in three broad groups: chondrites, primitive achondrites, and differentiated achondrites. Designation of a class type is determined by petrology, mineralogy, oxygen isotopes, and whole-rock chemistry. These groups are then separated into several clans and subgroups based on common affinities (Fig. 1.1).

Chondrites (carbonaceous, ordinary, enstatite) are generally considered among the most primitive solar system material and can contain chondrules and Ca-Al-rich inclusions in a fine- grained matrix (Weisberg et al., 2006) that has not experienced differentiation. Ordinary chondrites

(subgroups H, L, and LL depending on iron content from high to low-low) represent ~85% of observed falls, are the most common meteorite found in meteorite collections, and show a relatively wide range of compositions and textures (Kallemeyn et al., 1989; Weisberg et al., 2006).

Their dominance in our meteorite collection suggests that impacts are relatively common in the inner Main Asteroid Belt or near resonances that deliver these meteorites to the earth.

Using cosmic-ray exposure and argon-argon techniques together (Figs. 1.2 and 1.3) further develops our understanding of the history of a sample and its parent body. The 40Ar-39Ar dating technique has commonly been used to discuss quantitatively the impact and closure age chronology of meteorites. Argon diffuses out of a system at relatively low temperatures and is radiogenically produced by potassium, a common mineral component, making it an ideal isotopic dating system to measure impact events. Argon chronology of ~100 ordinary chondrites shows evidence of impact events in the first ~100 Ma of the solar system, an increase in impacts during the ‘lunar cataclysm’ (Turner et al., 1973) between 3500-4100 Ma, an impact at ~500 Ma in L 14

chondrites signaling disruption from their parent body, and at least one younger event in the H

chondrites at <1000 Ma (Swindle et al., 2014).

In comparison, cosmic-ray exposure (CRE) ages measure the duration of exposure to energetic

particles after a collision that liberated a meteorite from its parent body (Eugster et al. 2006). The

CRE ages of the H chondrites show a peak at 6-8 Ma (Graf and Marti, 1995), that is not reflected

in the Ar-Ar data (Swindle et al., 2014). The LL chondrites show evidence of degassing (i.e. a

collision) in the last ~1000 Ma (Takagami and Kaneoka 1987, Swindle et al., 2014), while

Samples included in this work

Figure 1.1. Meteorite classification by Weisberg et al., 2006. Samples from this work include ordinary chondrites and primitive achondrites. Others classify ureilites and brachinites as differentiated achondrites (Mittlefehldt et al., 2003). 15

LL Chondrite Age Distribution

e R

0 1000 2000 3000 4000 Figure 1.3. Neon cosmic-ray exposure (CRE) ages of LL Ar-Ar Age (Ma) chondrites (Graf and Marti, 1991). CRE data shows duration of space exposure after parent body ejection; Figure 1.2. Ar-Ar ages of 16 LL chondrites. information that is not obtained by Ar-Ar dating. Data from Swindle et al., 2014.

distinguishing themselves from impacts in the H and L chondrites with multiple ages ranging

~4200-4350 Ma (Fig.1.2). CRE ages of LL chondrites (Fig.1.3) show peaks ~15 Ma (Marti and

Graf 1992). Low energy impacts can cause break up from a parent body, exposing it to cosmic rays, without degassing argon. Absolute impact ages measured by Ar-Ar combined with relative transit times from a sample’s parent body to the earth measured by CRE ages provides a more complete picture of a sample’s history. More details on the background of these techniques will be discussed in Chapter 2.

One of the focuses of this work is the Ar-Ar chronology of one of the most famous and well- studied ordinary chondrites; Chelyabinsk (Fig.1.4). This meteorite is classified as an LL chondrite

(Popova et al., 2013; Galimov et al., 2013) and exploded in an airburst over the Chelyabinsk region of Russia in February of 2013. The airburst was estimated to be the equivalent of ~500 kilotons of

TNT (Brown et al., 2013). Thousands of stones were collected (Ivanova et al., 2013) with the largest weighing ~650 kg. Cosmic-ray exposure ages of Chelyabinsk indicate exposure for 1.2 ±

0.2 million years (Nishiizumi et al., 2013, Povinec et al., 2015).

Ages from absolute chronometers of Chelyabinsk are not well resolved. Samarium-neodymium

(Sm-Nd) isotopic analyses of Chelyabinsk (Fig. 1.5) were interpreted to show an impact event at 16

290 Ma that may represent separation from the parent body of Chelyabinsk. However, this age is

speculative because the isochron data is quite scattered and it seems a wide range of ages are

equally likely. No clear isochron is given by rubidium-strontium (Rb-Sr) isotopes (Fig. 1.6) either

(Galimov et al, 2013; Righter et al., 2015). Bouvier (2013) studied a dark melt fragment from

Chelyabinsk and obtained a lead-lead (Pb-Pb) age of 4538.3 ± 2.1, which is similar to the uranium-

Figure 1.4. A). Four-centimeter piece of Chelyabinsk showing shock veins (top left). B). Stills taken from video of Chelyabinsk fireball (by A. Ivanov), showing two main fragmentation stages (bottom left). C) 17O versus molar percent of olivine [Fa=Fe/(Fe+Mg)] of Chelyabinsk shows agreement with other LL chondrites (top right). D) Fe/Mg/Si three-element plot of Chelyabinsk compared with the other main chondrite groups (middle right). E) Rare earth elemental pattern of Chelyabinsk (normalized to CI) in red, indicates behavior similar to L and LL chondrites. Figure is adapted from Popova et al., 2013.

Figure 1.5. Sm-Nd measurements of Figure 1.6. Rb-Sr measurements of Chelyabinsk are Chelyabinsk are complicated and show many complicated and show many possible ages. Righter et possible ages. Righter et al., 2015. al., 2015.

lead (U-Pb) age of 4457± 35 (Lapen et al., 2014). Argon ages of Chelyabinsk show multiple possible impact ages: 1184 ± 40 Ma, 1014 ± 24 Ma, 716 ± 30 Ma, 312 ± 6 Ma (Righter et al.,

2015), 2083 ± 5 Ma, 264 ± 2 Ma (integrated ages from Lindsay et al., 2015), and 1700 ± 100 Ma

(Trieloff et al., 2018).

Table 1.1. Chelyabinsk ages determined by various isotopic dating methods. All ages are given in Ma. U- Method Ar-Ar K-Ar Rb-Sr Sm-Nd U-PB Re-Os He Age ~271 ~303 865 ± 971 153 ± 585 ~2906 585 ± 3908 455812 1,2 4 1 1 9 312 ± 6 1000 880 ± 120 2900 ± 500 834 ± 7 1 1 1 7 9 716 ± 30 1945 ± 168 1400 ± 300 3733 ± 110 2744 ± 13 1,2 1 5 9 1014 ± 24 1952 ± 169 4567 2861 ± 15 1,2 1 10 1184 ± 40 2736 ±199 4433 ± 110 11 1700 ± 1003 4452 ± 21 4454 ± 678 1. Righter et al. (2015), 2. Lindsay et al. (2015), 3. Trieloff et al. (2015), 4. Haba et al. (2014),

5. Nakamura et al. (2015), 6. Galimov et al. (2013), 7. Bogomolov et al. (2015), 8. Lapen et al. (2014), 9. Skublov et al. (2015), 10. Kamioka et al. (2014), 11. Popova et al. (2013), 12. Day et al. (2014)

Results from different geochronometers for Chelyabinsk are summarized in Table 1.1, which shows multiple different ages even within an isotopic system. Chapter 3 of this thesis demonstrates 18 an example of Ar-Ar measurement and data reduction, followed by new Ar-Ar results of

Chelyabinsk and the context of multiple dating results in Chapter 4.

The remaining focus of this thesis is on brachinites and ureilites, which are achondrites. Some consider them to be primitive achondrites, that is, meteorites that have experienced metamorphism that sometimes resulted in partial melting (Nehru et al., 1992; Weisberg et al., 2006) while others prefer the term differentiated achondrites based on high fractionation of lithophile and siderophile elements and evidence of an igneous origin (Mittlefehldt 2004). Generally speaking, achondrites have experienced extensive heating and differentiation (McCoy et al., 2006) and primitive achondrites are a bridge between the pristine chondrites and evolved achondrites. Further discussion of brachinites and ureilites proceeds below.

1.2 Brachinites

Only a few types of meteorite classes are dominated by olivine. One such type is the brachinites, a small group of achondrites that have a poorly understood history. In fact the first member of this group, Brachina, was thought to be a chassignite (Nehru et al., 1979) until further studies of oxygen and chemical compositions demonstrated that it was from a unique source (Nehru et al 1983;

Clayton and Mayeda, 1983). Since this time, there have been ~40 samples classified as brachinite or ‘brachinite-like’ ungrouped achondrites. The majority of studies on brachinites have focused on petrology and bulk chemistry (Fig. 1.7, 1.8) with little focus on chronology (Mittlefehldt et al.,

2003; Goodrich et al., 2010; Day et al., 2012).

Chapter 5 and 7 of this work include brachinites, brachinite-like achondrites, and ungrouped ultramafic achondrites that may have links to the brachinites (based on mineralogy, oxygen isotopes, and/or textures comparisons). Brachinites are mostly considered to be dunitic melt 19

Figure 1.7. Crossed polar transmitted light Figure 1.8. CI normalized highly siderophile element abundance patterns for brachinites show a wide range of photomicrographs of brachinites show dominant variability (data from Day et al., 2012). They have near olivine with minor clinopyroxene. Samples imaged are chondritic ratios of Re, Os, and Ru, and show enhanced a.) EET 99402, b.) Hughes 026, c.) NWA 3151, and d.) depletion of Pt, and Pd. NWA 4969. Images have a 4.1 mm horizontal axis (Keil, 2014 with images from K. Gardner-Vandy).

residues from a partial melt of chondritic-like precursor based on near-chondritic abundances of lithophile elements and relatively high abundances of siderophile elements (e.g. Johnson et al.,

1977; Ott et al. 1985, Nehru et al., 1992; Keil 2014). However, there is also evidence that suggest a cumulate formation based on petrofabric analyses and fractionated rare earth element (REE) patterns (Mittlefehldt et al., 2003). Brachinite-like achondrites share a similar composition to brachinites, though many consider them to have a more magnesian olivine and a higher volume % of orthopyroxene (Day et al., 2012). It is not clear if they are part of the brachinite group, giving rise to the term ‘brachinite-like’, while others are classified as ungrouped achondrites (e.g. Day et al., 2012; Singerling et al., 2013).

Brachinites are thought to have formed in the first ~5 Ma of the solar system, based on 53Mn-

53Cr age of 4563.7 ± 0.9 Ma for Brachina (Wadhwa et al. 1998), which is considered the best estimate of the crystallization age of the brachinite group. Brachinite NWA 4882 yields a 53Mn-

53Cr age of 4550.2 ± 0.8 Ma (Dunlap et al., 2016), which is considerably younger and suggests 20 either a different parent body than Brachina or a shared parent body large enough to maintain significant heat over ~14 Ma. Other dating methods used on this group include 4He, K-Ar, and

40Ar-39Ar (Patzer et al., 2003; Ott et al., 1985; Swindle et al., 1993; Mittlefehldt et al., 2003). The only reliable Ar-Ar age shows a resetting event at 4130 ± 60 Ma for brachinite EET 99402

(Mittlefheldt et al., 2003). Literature studies of CRE ages of brachinites are limited in sample size

(Patzer et al. 2003, Ott et al., 1985, Swindle et al., 1998, Bogard et al., 1983). The combined literature lists 21Ne measurements of only seven brachinites, which spread between ~3-50 Ma, with a group of three 21Ne CRE ages at ~50Ma (Eagles Nest, EET 99402, Hughes 026).

Elemental abundance and oxygen isotope measurements are required for better understanding of the ages and relationships of these meteorites. Oxygen isotopes are subject to mass fractionation effects that can trace the interactions of different reservoirs in the heterogenous solar nebula and have been used to help identify related meteorites from a shared parent body (Clayton et al., 1973;

1993; 1996). Achondrites have undergone at least some differentiation which can cause oxygen isotopes to be homogenized, which in theory would amplify the ability of using oxygen isotopes as a parent body signature. The brachinite oxygen isotopes are heterogenous (Fig. 1.9) which can be explained by inefficient heating of a single parent body or different oxygen reservoirs from individual parent bodies (Eugster and Michel, 1995) that might be distinguishable (common history) in the CRE and Ar-Ar data. Oxygen isotopes were not measured independently in this work primarily because many of the samples have already had these measurements performed.

Mineral abundances have been measured in some brachinite, brachinite-like, and ungrouped ultramafic achondrites (Goodrich et al., 2006; 2010; Day et al., 2012; Gardner-Vandy et al., 2013) and are homogenous within a sample, are similar to one another, and have been used to distinguish 21

these samples from other meteorite groups (e.g. Fo and wt.% CaO and Cr2O3 of olivine Goodrich and Righter, 2000; Goodrich et al., 2006; 2010).

Figure 1.9. Oxygen isotopes of brachinites and Figure 1.10. Histogram of CRE ages for acapulcoites brachinite -like achondrites. Measurements marked with and lodranites, showing a tight clustering of ages blue arrows are samples that are included in this work. interpreted as common impact events. Black regions The light grey box shows the 2 variation for the correspond to acapulcoites while blue regions are from oxygen measurements on 18 samples. The dark grey box lodranites. Data from Weigel et al. 1999. excludes Brachina (2). Adapted (addition of blue arrows) from Greenwood et al., 2017. One of the goals of this work is to use noble gas mass spectrometry to expand the currently limited data set on Ar-Ar ages (Chapter 5) and CRE ages (Chapter 7) in order to better understand the history of 16 brachinite, brachinite-like, and ungrouped achondrites. These measurements will significantly augment the existing data on impact and exposure ages, which combined with literature data (noble gas, elemental and mineral abundances, and oxygen isotopic measurements) may help address the classification and relation of brachinites and brachinite-like achondrites from a different perspective.

Furthermore, I hoped to provide a systematic study of 4He and 40Ar gas retention ages and of the potentially present trapped heavy noble gases in these samples for the first time. However, a combination of low sample gas abundance, high uncertainties, or missing/insufficient elemental data has made it difficult to present significant results. 22

1.3 Ureilites

Ureilites are ultramafic achondrites that might represent the mantle of a large single asteroid that was stratified in olivine composition; the ureilite parent body (UPB) (Goodrich, 1992; Warren and Kallemeyn, 1992; Goodrich et al., 2004). As in the case of the brachinites, it is difficult to determine the petrogenesis of the UPB due to conflicting igneous and primitive characteristics

(Goodrich, 1992; Mittlefehldt et al., 1998).

Figure 1.11. Olivine compositions (in Fo) of main group Figure 1.12. Oxygen isotopes of achondrites showing ureilites, polymict ureilite clasts, and ureilite Almahata strong heterogeneity in the ureilites (red circles- Sitta. There is a prominent peak at Fo ~ 79. Goodrich et Clayton and Mayeda (1996), blue/white squares- al., 2015. Greenwood et al., 2017) including Almahata Sitta (quartered squares-Greenwood et al., 2017).

The collection of ureilites is dominated by the main group (or monomict) ureilites that are believed to be mantle residues. A subclass of ureilites consists of polymict fragmental breccias that span the same range of magnesium to iron ratios (Fig. 1.11) in olivine (given as Fo; 100*molar

Mg/(Mg+Fe)) as the monomict ureilites, suggesting both classes derive from the same source

(Goodrich et al., 2004; Downes et al., 2008) despite having extremely heterogenous oxygen isotopes (Fig. 1.12). 23

CRE studies for the ureilite group (Rai et al., 2003) suggest possible groupings around 1 and

10 Ma, but Rai et al. (2003) based their analysis on CRE data alone. Recent work by Goodrich et al. (2015) suggests a common history that links all ureilites to a single Ureilite Daughter Body

(UDB) that formed after the initial UPB breakup (rather than multiple daughter sources). They suggest either a single breakup of the UDB at > 46 Ma with subsequent collisions producing smaller fragments, or a cratering event at 46 Ma followed by UDB break up at ~20 Ma based on cosmic-ray exposure ages of recovered samples from the fall of Amahata Sitta among other ureilites (Fig. 1.13). If it is assumed that the ureilites come from a single daughter body, an impact could break apart material from different heterogeneous or homogeneous regions (Warren and

Kallemeyn, 1989; Goodrich et al., 2002), which might be evident in their CRE data.

Figure 1.13. Neon exposure ages of main group (monomict) ureilites, polymict ureilites, ureilites from Almahata Sitta (AhS), and ordinary chondrites from AhS. Goodrich et al., 2015.

Chapter 6 of this thesis examines CRE ages for evidence of clusters that may improve our understanding of the origin of ureilites. Possible relationships of CRE ages with other parameters are examined for implication about how they might relate to the structure and heterogeneity of the proximate ureilite source body. This will be done by comparing oxygen isotopic ratios (Δ17O) and the ratios of magnesium to iron in olivine (Fo) with their corresponding CRE ages to look for evidence of possible common impact events. Although Δ17O and Fo (Fig. 1.14) tend to correlate

(Clayton and Mayeda 1988; 1996; Mittlefehldt et al., 1998; Rumble et al., 2010), an impact into a homogeneous area (Fig. 1.13) would be expected to produce a very tight cluster in both parameters.

Figure. 1.14. Δ17O and Fo for ureilites (diamonds- Clayton and Mayeda 1988) and polymict ureilite Almahata Sitta (circles-Rumble et al., 2010). In general, samples show stronger depletion in 17O with higher Fo number, listed here as mg# of olivine cores.

1.4 Thesis Structure and Logistics

• Chapter 1: Introduction

• Chapter 2: Techniques and Samples

• Chapter 3: Ar-Ar Data Reduction Example

• Chapter 4: Ar-Ar Ages of Chelyabinsk 25

• Chapter 5: Ar-Ar Ages of Brachinites and Brachinite-Like Achondrites

• Chapter 6: CRE Ages of Ureilites

• Chapter 7: CRE Ages of Brachinites, Brachnite-Like Achondrites, and other Achondrites

• Chapter 8: Conclusions

Background information of the techniques used and details of the samples measured in this thesis will be covered in Chapter 2, followed by an example of data collection and reduction of

Ar-Ar in Chapter 3. The results of my Ar-Ar studies are in Chapters 4 (Chelyabinsk) and 5

(Brachinites). I will briefly discuss the multiple geochronometric ages of Chalyabinsk including my results and discuss their significance. The Ar-Ar measurements of brachinites proved difficult and were unsuccessful, but my work is discussed to provide context for future measurements. CRE ages, oxygen isotopes, and Fo of ureilites are discussed in Chapter 6. Though no new measurements on ureilites are provided, the motivation of this chapter was to evaluate the heterogeneity in oxygen isotopes in CRE age groupings. Chapter 7 discusses new measurements and results for CRE ages of brachinites, brachinite-like, and ultramafic achondrites (together referred to as UMA- ultramafic achondrites). This work increases the CRE data of brachinites by a factor of ~ three and increases our understanding of impacts on the parent body(ies). Sample classifications are readdressed and more clearly defined based on oxygen isotopes and mineral data. Finally, the classification and proposed CRE age groupings are discussed. Figures, equations, and tables in this thesis will be numbered by the chapter they are in followed by the chronological order of the figure. For example, the third figure in chapter four will be identified as

Figure 4.3. Samples measured in this work include North West Africa (NWA) 595, 1500, 3151,

4518, 4874, 4876, 4882, 4969, 6077, 6474, 6962, 7297, 7605, 8777, 10637, Ramlat as Sahmah

(RaS) 309, and Chelyabinsk. 26

Chapter 2: Techniques and Samples

2.1 Argon-Argon Background

Some naturally occurring nuclides are radioactive, i.e. they spontaneously decay into nuclides of other elements at a rate that can be accurately measured (eq. 2.1). This discovery (Becquerel

1896) led to the development of isotopic dating, which measures the age of a rock based on precise measurements of the daughter isotope and parent element. More accurately, what is being measured is the amount of time it would take to produce the measured amount of the radiogenic product (eq. 2.2). Radiogenic noble gas isotopes are particularly useful because noble gases are relatively rare in geologic materials and are chemically unreactive. Therefore, the production of even a small amount of noble gas through radioactive decay can cause a significant, i.e. measurable, change.

푑푁 퐸푞. 2.1: ⁄푑푡 = −휆푁

--Where the number of radioactive atoms at a time t is represented by N, and the decay constant is given by lambda. The decay constant represents the fraction of parent nuclides that decay in a given time.

Integration of eq. 1, rearranging and taking the natural log, yields the basic equation used for geochronology:

1 퐸푞. 2.2: 푡 = ln (1 + 퐷⁄ ) 휆 푁

-- Where the radioactive parent is N, the daughter product is D, and t is the age.

The 40Ar-39Ar (referred to as Ar-Ar) dating technique is a modification of the 40K-40Ar dating system (referred to as K-Ar), one of the earliest isotopic dating techniques developed. Even though

40K only constitutes about 0.01% of naturally occurring potassium, and only ~10% of it decays to 27

40Ar (the rest decays to 40Ca), it is a very useful geochronometer. 40K decays to 40Ar with a relatively long half-life (1.25 Ga) and is at least a minor element in many common rock minerals

(e.g. feldspars, micas, pyroxene) , making 40K-40Ar an appropriate isotopic dating system that can reliably date materials from >1 Ma (McDougall and Harrison, 1999). Because of the dual decay of 40K (Fig. 2.1), eq. 2.2 needs to be modified to include the fraction of 40K that decays to 40Ar

(Table 2.1) (McDougall and Harrison, 1999).

Eq. 2.3:

40 40 40 -- Where FAr is the fractional decay of K that decays to Ar,  is the total decay constant for K,

40 e and e’ are the decay constants of K by two electron-capture pathways (Fig. 2.1).

Now equation 2.2 becomes

Eq. 2.4:

The K-Ar system has a solid (K) and gas (Ar) component which are measured by different methods. Typically, a portion of the sample has its K measured by chemical dissolution, while another portion has its argon extracted via heating and measured with a mass spectrometer. Having two separate measurements is not ideal, especially for samples that are heterogenous or that may have experienced partial resetting (partial loss of argon from a thermal event that did not degas the minerals completely). Merrihue and Turner (1966) analyzed irradiated meteorite data and noticed that neutron irradiation of 39K produced 39Ar, i.e. 39K(n,p) 39Ar. This eventually led to the development of the Ar-Ar dating technique (this takes advantage of the known40K/39K in nature and the measured 40Ar/39Ar to determine the 40K/40Ar and thus its age).

Figure 2.1. Decay of 40K to 40Ca (~90%) and to 40Ar (~10%), which is dominated by electron capture. Figure adopted from Beckinsale and Gale (1969), Garner et al., (1975), Dalrymple and Lanphere (1969) from MacDougal and Harrison, 1999.

Table 2.1. Potassium (K) decay constants. Quantity Value

-10 -1 e+e’ (0.5757 ± 0.016) a - -10 -1  (4.9548 ± 0.0134) e a - -10 -1  = e + e’ +  (5.5305 ± 0.0279) e a - Branching ratio = (e + e’)/  0.116 ± 0.001 푙푛 2 (1.253 ± 0.005) e9 a-1 푡 = 1/2 

Constants from Renne et al., 2011

Some advantages of Ar-Ar over K-Ar include getting the K/Ar ratio from a single and more precise measurement (relying on 40Ar/39Ar ratio rather than absolute abundances), the ability to 29 measure smaller samples, and a multiple step-heating approach rather than a single fusion age. The step heating approach allows the same sample to be heated at successively higher temperatures until melting occurs, allowing multiple 40Ar/39Ar ratios to be measured. Multiple measurements at different temperatures makes it possible to identify partial resetting and release from multiple potassium-bearing minerals that are relatively more/less retentive of argon. As with all dating methods, many assumptions are made including a constant decay rate of the parent nuclide that is independent of temperature and pressure, the 40K/39K ratio in nature is constant, all radiogenic argon was produced by decay of 40K and not incorporated extraneously, and that the system remained closed from loss/gain of 40K and 40Ar since crystallization or the event in question that is being dated. The first two assumptions are reasonable to make, while the latter two may require corrections during data reduction (discussed later).

Ar-Ar ages are typically presented in the form of an age spectra or ‘plateau plot’. After the necessary corrections are made to the data (discussed later), the plateau plot compares the apparent age vs. the fraction of the total 39Ar release from the sample from multiple temperature steps. If there is series of concordant ages (ages calculated from different steps that agree within uncertainty) that contains >~50% of the total released 39Ar, then this age is interpreted to represent the formation age or resetting event and is commonly referred to as the plateau age. This work uses the term ‘partial plateau’ for situations where this definition is not strictly met, but the data still seem somewhat reliable.

The samples’ argon (or other noble gases for CRE analysis) isotopes are measured with a mass spectrometer, with a brief description provided here (Fig. 2.2). The main components include a flight tube, electron bombardment ion source, magnetic analyzer, and a detector. A high vacuum 30 is required so that ions traveling in the flight tube are not affected by collisions with other gas molecules on the way to the detector. The source creates a beam of positive ions from the

Figure 2.2. Schematic of general mass spectrometry. Sample gas is released into the source, ionized, collimated into an ion beam, and accelerated toward the detector through ~ 4 kilovolt potential. The ion beam interacts with a magnetic field, which separates the ions based on the mass to charge ratio. Schematic kindly provided by Asia Marianelli. sample gas and is collimated and accelerated through the flight tube to the detector. Electrons produced by thermal emission from the filament are accelerated by a potential difference and collide with sample gas. This causes removal of one or more electrons from the gas, resulting in positive ions that are dominantly of single charge. The positive ions are accelerated across the flight tube by a potential difference of 4 kilovolts. A homogenous magnetic field (typically of

~100 Gauss) is produced in the gap between the sector magnet, which is perpendicular to the direction of the ion beam. The magnetic field applies a force on the ions, which cause them to follow a circular path of a radius that is determined by the mass to charge ratio (m/e) on the way 31 to the detector. The detector is set to measures the current of specific isotope based on the discrete m/e ratios, as the radius of the path varies according to the square root of the mass to charge ratio.

Energy of a particle going through an electric potential:

퐸푞. 2.5: 푒푉 = 1 푚푣2 2 Expressed in terms of velocity: 2푒푉 퐸푞. 2.6: 푣 = √ 푚 Setting centripetal force = magnetic force:

푚푣2 퐸푞. 2.7: = 퐵푒푣 푟 Expressed in terms of the radius: 1 2푉푚 퐸푞. 2.8: 푟 = √ 퐵 푒

Expressed in terms of the mass to charge ratio:

푚 퐵2 푟2 퐸푞. 2.9: = 푒 2푉

--Where m is the mass of the ion, e is the charge of the ion (produced in the source), which is accelerated through a potential of V with a velocity v though a magnetic field B in a circular path of radius r.

Because argon nomenclature can be confusing, some common terms and some considerations for Ar-Ar analysis are described below.

Radiogenic argon (40Ar) is produced in situ by decay from 40K and forms the basis of K-Ar dating. Atmospheric argon refers to argon that has the isotopic composition of earth’s atmosphere

(40Ar/36Ar ~ 298 and 38Ar/36Ar ~ 0.1869 Renne et al., 2011; Wieler, 2002). It is very important to consider and correct samples for atmospheric contamination, as atmosphere can be absorbed onto mineral grains, particularly in hot desert environments. Atmospheric argon is also used for 32 determining the mass discrimination of the mass spectrometer. To do this, atmospheric air introduced through a pipette is measured and compared to the known standard argon ratios in air to correct the mass discrimination, which usually changes on the time scale of ~months.

Trapped argon is the argon (of any isotope) incorporated into a mineral or rock at the time of formation or subsequent heating event (which could be atmospheric argon, for example). The use of a three-isotope plot, or isochron (Fig. 2.3), has the potential to determine the composition of the trapped component. This can be done by plotting the incremental data of radiogenic 40Ar (the daughter product) and the neutron induced 39Ar (representative of the parent nuclide), each normalized to the non-radiogenic 36Ar (i.e. 40Ar/36Ar vs 39Ar/36Ar). The y-intercept indicates the trapped composition and the slope of the isochron is related to the sample’s age. However, 36Ar is usually not as well determined as 40Ar and is not as abundant, which can introduce large errors.

Using 40Ar as the reference isotope can help avoid this problem as it is the most abundant and thus is measured more precisely. This inverse isochron plot (36Ar/40Ar vs 39Ar/40Ar) is used in addition to the standard isochron plot to identify the trapped component (given by the inverse of the y- intercept) and estimate the age (determined from the inverse of the x-intercept) when possible

(Roddick et al., 1980). It is not uncommon for identification of more than one trapped component

(Korochantseva et al., 2007), particularly in samples that have been partially reset.

Cosmogenic argon is produced from cosmic-ray spallation of target nuclei (Ca, K, Ti, and Fe) and from secondary neutrons produced from cosmic-ray interaction with other target nuclei

(Hohenberg et al., 1978). Purely cosmogenic 38Ar/36Ar ~1.54 (Wieler, 2002), and a two- component deconvolution of the cosmogenic and trapped components can be applied, and the trapped component is then subtracted from the measured 36Ar. 33

Neutron bombardment is necessary for the production of 39Ar which is needed to determine a sample’s age. Argon is also produced during bombardment of Cl, K, and Ca and is referred to as neutron-induced argon. Samples for this work were irradiated at the Cadmium-Lined In-Core

Irradiation Tube (CLICIT) facility at Oregon State University. A significant advantage of using the CLICIT facility is that the Cd lining significantly reduces the number of thermal neutrons, which in turn significantly reduces the production of 38Ar by neutron capture on 37Cl. This means that 36Ar and 38Ar are both predominantly mixtures of only trapped Ar and Ar produced by cosmic- ray spallation.

Table 2.2. Important interfering reactions on Ca and K. Isotope K Ca 40Ar 40K(n,p) 40Ar 39Ar 39K(n,p) 39Ar 42Ca(n,)39Ar 38Ar 42Ca(n,n)38Ar 37Ar 40Ca(n,)37Ar 36Ar 40Ca(n,n)36Ar

Neutron-induced argon from K and Ca (Table 2.2) are corrected for by using irradiation monitors (recently melted and re-crystallized K2SO4 and CaF2) that are included in the sample irradiation package. Irradiation of K does not produce 36Ar, so corrections can be made for its

40 production of Ar by using the known atmospheric argon ratio and correcting the K2SO4 data.

Similarly, irradiation of Ca does not produce 40Ar, therefore corrections can be made for its

36 production of Ar based on data from CaF2 (in both cases, the amount of reactor produced argon is calculated and then subtracted). Two reactor-induced isotopes are also radioactive (39Ar, 37Ar) and are corrected for decay by using the time difference between irradiation and measurement

(39Ar has a half-life of ~270 years and 37Ar has a much shorter half-life of ~35 days, e.g. Fireman et al., 1960; Stoenner et al., 1965). Another effect of neutron irradiation to consider is called recoil, which can artificially increase the apparent age at low temperature steps and decrease the age at 34 high temperature steps by processes that can cause 39Ar to move from a relatively low retentive mineral to a more retentive mineral. A mineral that is less retentive will degas at lower temperatures and the loss of 39Ar from recoil will artificially increase the apparent age. The increase of 39Ar in more retentive minerals such as olivine, will lower the apparent age at high temperatures.

Fig 2.3. Example of isochron and reverse isochron. (Left) Full data from each temperature step of sample SB1. The circled data might represent the trapped component and is isolated (right). The trapped component in this case is 40Ar/36Ar ~ 300.

The following derivation can be found in more detail in McDougal and Harrison (1999). The amount of 39Ar produced from neutron bombardment is proportional to the amount of 40K (via the constant 40K/39K ratio in nature) in the sample, but also depends on the irradiation time, the neutron flux, and the energy of the neutrons. This is expressed (Mitchell, 1968) as: 35

Eq. 2.10:

39 39 --Where Ar is the amount of argon produced from irradiation of K in the sample,  is the irradiation duration, (E) is the neutron flux integrated over energy E, (E) is the cross section of the neutron capture at energy E. Note that in practice the production of 39Ar is not well determined in this equation because the actual neutron dose is not well measured.

Rearranging equation 2.4 to solve for 40Ar, we have:

Eq. 2.11:

Now we can use these equations (2.10 and 2.11) to express the 40Ar/39Ar ratio:

Eq. 2.12:

It is useful to take advantage of a dimensionless number called “J” (Grasty and Mitchell, 1966), known as the irradiation parameter.

Eq. 2.13:

Which now simplifies the 40Ar/39Ar ratio to:

Eq. 2.14:

How is J determined? As mentioned earlier, the neutron dose that a sample receives is not well known. This is resolved by including flux monitors with a known age along with the samples during irradiation (Merrihue and Turner 1966), which act as dosimeters. J can be solved using the age of the flux monitor and its measured 40Ar/39Ar. 36

Eq. 2.15:

--In this form, the t refers to the known age of the flux monitor. For this work, a hornblende

Hb3gr was used as a flux monitor (1081.0 ± 1.2 Ma, Renne et al., 2011). Hb3gr is the oldest monitor, making it the most ideal choice for ancient samples (meteorites).

Now that J is known, the age of the sample (t) can finally be determined!

Eq. 2.16:

2.2 Ar-Ar Experimental Design and Data Acquisition

Argon is extracted via heating of whole rock samples under ultra-high vacuum and analyzed with a mass spectrometer to determine the relative abundances of its isotopes. Samples are heated either by a furnace or laser that heats the sample in a controlled manner, allowing for successively higher temperature extractions. A laser system typically allows direct viewing of sample heating and has lower blanks than the furnace due to a smaller volume. Lower laser blanks allow for analysis of smaller samples than the furnace (more details on blanks will be discussed later). Gases that are released from heating are corralled through a series of isolation valves and purified by

‘getter’ systems. The schematic representing the extraction system used for Ar-Ar is shown in

Figure 2.4.

The experiments for this work used an all metal resistance heated furnace with two separately pumped sections. The outer portion of the furnace is enveloped by a copper coil that contains chilled running water and is kept under a low vacuum (~10-3 torr) to reduce the amount of oxidation of the crucible. The inner portion of the tungsten furnace contains a tantalum crucible that has a hole in its base for inserting a tungsten-rhenium (W-Re) thermocouple, and is in direct contact 37 with the sample holder and the volume of the rest of the argon extraction line (~10-8 torr). Power to the furnace is supplied by a low-voltage, high-current Electronic Instruments (model TCR

10T500-1-D) transformer that is controlled by a WatLow PID (Proportional Integral Derivative) temperature controller. The temperature set point, opening and closing of valves, and mass spectrometer data collection are all automated and controlled by software developed by Alan

Deino (Deino 2001).

Furnace samples

Laser samples

Figure 2.4. Extraction line schematic for the mass spectrometer extraction line used for this work . This set up includes two laser ports (Laser Left and Laser Right) and a furnace system. MS Turbo, HiCube-F and HiCube- L are turbo pumps. PA1-PC2 refer to air pipettes used for calibrations.

For a typical experiment heating step, a sample is dropped into the furnace crucible from the sample holder. The sample is dropped by manipulating pieces of ferromagnetic metal inside a glass

‘sample tree’ by using magnets from outside of the glass. The volume of the tree and furnace is isolated by closing the appropriate valve in the extraction line during heating. Power to the crucible is linearly increased over a period of ~3 minutes before reaching the target temperature. The 38 sample is maintained at the target temperature for 12 minutes which releases gas from the sample.

This gas must be cleaned of active species before isotopic analysis can be performed to avoid oxidation and overloading of the ion source filament (tungsten) required for ionization before entry into the mass spectrometer. This is done by a two-stage expansion and ‘gettering’, once at room temperature (cold gettering) and once at ~400°C (hot gettering) while the mass spectrometer volume is kept in isolation. Getters use an alloy (such as titanium, zirconium, aluminum) that adsorb active gases (CO, CO2, H2, H2O, CH4), while leaving argon and noble gases unaffected

(Fig. 2.5). Once the crucible has cooled to ~500°C, the isolation valves are now opened to expose the gas to cold gettering for ~3 minutes. Next, the gas is advanced through the next valve for hot gettering for an additional 3 minutes (Fig. 2.6). The sample gas is now ready for analysis and is introduced to the mass spectrometer. After 30 seconds of equilibration time, the mass spec is again isolated from the rest of the extraction line and gas analysis begins.

Figure 2.5. Example of SAES getter used for gas purification. The alloy is heated inside the extraction line and absorbs gases (CO, CO2, H2, H2O, CH4) while leaving noble gases unaffected.

This extraction procedure is repeated for each temperature in a multi-step experiment (usually

~10-15 steps ranging from ~300°C-1500°C). Gas analysis in the mass spectrometer takes ~20 minutes, so it is ideal to overlap the long heating and gettering time in order to maximize efficiency. It has been determined through prior experiments on this system that waiting ~5 minutes after gas analysis begins to start the next temperature step will provide the sufficient ~5 minutes to pump out the gas in the mass spectrometer after the analysis is completed.

Figure 2.6. (Left) Part of the actual extraction line. (Right) Labeled version of the figure on the left. Purple indicates the path of gas from the furnace, blue indicates possible paths of gas from the laser. The numbers refer to valves used to isolate the gas for cleaning (see figure 2.4). M stands for Manual valves, G stands for Getter, and IG stands for Ion Gauge. Note gas from the furnace and laser (blue and purple lines) extend into the extraction volume up to valve 9 and 10 but it is not shown for clarity. Data was collected and analyzed from the VG5400 mass spectrometer following automated procedures from the Mass Spec Ver. 7.884 computer program (Deino, 2001). The mass to charge ratio (referred to just as ‘mass’ for purposes of this section since virtually all gas is singly ionized) between 36-40 and the baseline (electronic signal measured between mass 40 and 39) are measured 40 by the mass spectrometer based on magnetic field strengths from a calibration table. The walls of the mass spectrometer can be a source or a sink for various argon isotopes, so by the time the last isotope in a cycle is measured, it has already experienced some source/sink effect. This behavior and the time difference need to be considered in order to obtain the most accurate isotopic ratios.

To do this, multiple measurements are repeated and extrapolated as follows. The current

(proportional to the number of ions striking the detector) of each relevant isotope is measured by a multiplier with the respective time (in seconds) starting at high mass (40) and ending at low mass

(36), completing one cycle. The measured signal is the average amount of charges per second recorded over a specific integration time, and each cycle usually measures 8 integrations. This measurement process is repeated for 10 cycles and the data for each isotope is extrapolated to t0

(time that the gas was first introduced to the spectrometer) using an appropriate linear or parabolic fit.

It is important to minimize the ‘blank’ of the extraction line. The blank is the amount of gas

(specifically the argon isotopes) measured in the mass spectrometer after following the normal operation procedures without the sample and is typically of atmospheric composition. The blank data are needed in order to apply proper blank corrections to the sample data. Different types of blanks can be performed, including a ‘hot blank’ that involves heating an empty crucible, or a

‘cold blank’ that does not include any heating. Multiple blanks were analyzed or ‘run’ before and after changes to the extraction line (sample loading, valve changes, getter replacements, other maintenance). More importantly, a set of hot blanks was measured before and after a sample analysis. These hot blanks were typically in two sets of four (500°C, 1000°C, 1250, and 1505 °C), interpolated, and applied to the appropriate sample steps (determined by the temperature) to 41 subtract the contribution of argon of the system from the sample. More details of blank corrections can be found in Chapter 3.

An advantage for the extraction system used in this work in particular, is that higher temperatures can be stably maintained with the furnace than with the laser. The furnace system has much better calibration of temperature than the laser, based on melting of pure metals, which was unable to be successfully determined in the laser experiments.

2.3 Cosmic-Ray Exposure Background

Energetic particles called cosmic rays permeate the interplanetary medium and produce nuclides through nuclear interactions with the elements contained in asteroids (Table 2.3). These cosmic-ray produced isotopes can provide useful information about a meteorite’s exposure age and its pre-atmospheric size. Cosmic rays are effectively shielded from interacting with a sample that is buried by ~ a few meters of rock on its parent body and also by the earth’s atmosphere after the meteorite impacts the earth. Therefore, the cosmic-ray exposure (CRE) age represents the length of time that a meteorite was exposed to cosmic rays after excavation on its parent body until impacting the earth. There are two types of cosmic rays, solar cosmic rays (SCR) and the more energetic galactic cosmic rays (GCR). Galactic cosmic rays originate from outside of the solar system and are thought to be synthesized by similar processes as nuclei in the sun. Their high energy (GeV) suggests that they may be linked to shock acceleration from Wolf Rayet stars or supernovae environments (Lingenfelter et al., 2000; Binns et al., 2001). The more energy a cosmic ray has, the more likely it will induce a reaction during bombardment, typically around the several MeV range (Eugster 2006). Most of the high-energy reactions are due to GCR (~12%

α-particles, ~87% protons) primaries, while the lower-energy reactions are due 42 to GCR secondaries, particularly neutron reactions (Niedermann 2002).

The most commonly reported exposure ages are from the stable nuclides 3He, 21Ne, and 38Ar, which have the largest number of measurements and best quality data (Schultz and Franke,

2002). Each age is based on the production rate of the nuclide, Pstable, and the measured abundance, Cstable (Wieler 2002).

퐶푠푡푎푏푙푒 Eq. 2.17: 푡푒푥푝 = ⁄ 푃푠푡푎푏푙푒

Table 2.3. Production rates of cosmogenic He, Ne, and Ar. 3He 21 Ne 22 Ne/ 21 Ne 38 Ar radius/ depth O Mg Al Si Fe Mg Al Si Ca Fe Mg Al Si Ca Fe Ni 5,s 148 113 124 138 86.6 53.2 32.6 25.2 8.57 2.03 1.290 1.188 1.197 1.256 7.47 6.54 5,c 166 124 130 143 89.0 69.0 37.8 28.8 8.77 1.99 1.200 1.224 1.197 1.236 7.43 6.58 10,s 163 125 130 142 88.2 70.2 37.4 28.8 8.69 1.97 1.191 1.228 1.190 1.234 7.39 6.50 10,c 190 144 139 149 90.8 98.7 45.9 34.9 8.89 1.86 1.091 1.265 1.187 1.197 7.23 6.41 15,s 170 132 131 143 87.3 80.4 39.6 30.6 8.53 1.88 1.152 1.260 1.187 1.223 7.19 6.33 15,c 202 158 139 150 87.2 121.0 49.9 38.0 8.40 1.66 1.054 1.325 1.185 1.175 6.66 5.89 25,s 186 148 135 147 86.8 102.7 44.7 34.6 8.36 1.75 1.095 1.309 1.184 1.200 6.78 6.01 25,c 232 188 149 158 88.2 159.2 59.3 45.5 8.32 1.46 1.020 1.377 1.188 1.127 6.17 5.52 40,s 188 152 132 144 82.8 111.7 45.5 35.6 7.88 1.60 1.076 1.339 1.176 1.185 6.33 5.60 40,c 246 205 148 156 82.7 184.3 65.0 51.2 7.63 1.14 1.002 1.413 1.151 1.075 5.32 4.79 65,s 173 144 120 133 74.5 108.0 41.8 33.2 7.02 1.41 1.071 1.369 1.174 1.179 5.60 4.95 65,c 210 180 119 127 62.0 171.7 53.6 43.4 5.44 0.69 0.969 1.470 1.159 1.022 3.63 3.22 100,s 161 135 112 123 68.1 104.3 39.2 30.9 6.33 1.26 1.066 1.388 1.189 1.180 5.08 4.47 100,c 138 130 78.5 86.2 38.7 120.6 36.5 30.2 3.20 0.37 1.020 1.537 1.150 0.994 2.00 1.80 Modele d production rates of He, Ne, and Ar are listed in 10-10 ccSTP/(gMa), based on the main target elements (Mg, Al, Si, Ca for Ne, for example) where 1 ccSTP = 2.69e19 atoms. The radius of the meteoroid is given in centimeters, the subscript ‘s’indicates values for the surface, ‘c’ indicates values at the center. This table is from Wieler 2002 and is a summary of work from Leya et al., 2000; Leya et al., 2001; Hohenberg et al., 1978, and others.

In practice, determining the age is not as straight forward as the equation 2.17 appears. The production rate for each nuclide depends on multiple factors that are unique to each meteorite, such as its chemical composition, its pre-atmospheric size (the size of the object during exposure), and the depth from the surface of the measured sample (Figs. 2.7-2.8). The greater the size of the object and the greater the depth (after a certain point), the greater the shielding from cosmic rays.

Therefore, these two variables are referred to as ‘shielding’ and must be determined for a proper determination of the production rate, and thus its age. 43

Production rates have been empirically determined (Bogard and Cressy, 1973; Eugster, 1988;

Graf et al., 1990) and modeled (Masarik and Reedy 1994; Masarik et al., 2001; Leya et al., 2001,

Leya and Maraik, 2009). Figure 2.9 shows a comparison of model 21Ne production with the measured concentrations from the chondrite Knyahinya.

Figure 2.7. Production rates of 60Co and 21Ne in chondrites and iron meteorites for two different radii (30 and 100 cm for 21Ne), illustrating the depth and size dependence of production rates vary considerably for different nuclei. From Wieler 2002.

21 Figure 2.8. Production rates of Ne for H (left axis) and L (right axis) ordinary chondrites, for various radii between 5 cm and 120 cm. LL production rates are approximately 2% higher than the L chondrite rates. This demonstrates that the production rate depends on the target material (i.e. the target chemistry).

Figure 2.9. Measured (squares) 21Ne concentrations in the L/LL chondrite Knyahinya compared to modelled (solid lines) production rates of 21Ne (Leya et al., 2000; Masarik et al., 2001). The models show good agreement to the measured data (Graf et al., 1990). The contributions from primary protons, secondary protons, and secondary neutrons to the production rate (dashed lines) show that secondary neutrons dominate production of cosmogenic nuclides. Figure from Wieler 2002. 45

CRE ages in this work are calculated based on production rates determined from shielding models by Leya and Masarik (2009), in conjunction with the measured cosmogenic 21Ne/22Ne of the respective sample. Their model determines the appropriate combinations of pre-atmospheric radius of the meteorite along with the depth from the surface that our representative sample could have been located. The model is an improvement over prior versions because it includes a better understanding of the relevant nuclear reactions (Koning et al., 2005), better capability of determining the appropriate neutron cross sections (Ammon et al., 2008), and is the first model that accurately predicts 3He production rates. The basic equation used for the production rate in this model is given by:

Eq. 2.18:

--Where Pj is the production rate of a cosmogenic nuclide j as a function of spherical radius R, depth d, and solar modulation parameter M- a measure of GCR particle flux reduction by magnetic fields from the sun. The first summation goes through all target elements i where Ai is the mass number of the target element i, NA is Avogadro’s number, ci is the abundance of i. The second summation represents the reaction of particle type k (primary proton, secondary proton, secondary neutron, see Figure 2.9). The excitation function of the production of j from element i by particle type k is represented by σj,i,k(E), and Jk(E,R,d,M) is the differential flux density of particle type k, which both depend on energy E (Leya and Masarik, 2009). Excel files for the calculations of elemental production rates using this model can be found at http://www.noblegas.unibe.ch or from the authors, who note that this model is only reliable for objects that have a pre-atmospheric size of less than ~65 cm. 46

2.4 CRE Experimental Design and Data Acquisition

The 16 brachinite/brachinite-like achondrite samples were divided into two aliquots (as permitted by available mass; ranging from 8-95 mg; average ~46 mg) in order to have their noble gases (He, Ne, Ar, Kr, and Xe) measured. Gas extraction, cleaning, and measurement followed established methods (Busemann et al., 2000; Riebe et al., 2017). Noble gases were measured on the custom-built sector-field noble gas spectrometer at ETH Zurich, named “Albatros”, that uses a Faraday cup (for gas at high pressure) as well as a multiplier in counting mode (for gas at lower pressure).

Some basic background information on mass spectrometry, gettering, and furnace systems can be found in Ar-Ar Experimental Design and Data Acquisition section of this chapter, which applies to both the cosmogenic isotopes and Ar-Ar measurements. The notable differences in the cosmogenic isotope measurements will be further detailed here.

Samples were placed in an all metal sample holder maintained in vacuum above a furnace that was continuously preheated to ~300-400 °C (Fig. 2.10). Gas extraction involved dropping the sample into a Mo crucible increasingly heated to ~1700 °C for ~30 minutes. Complete degassing was confirmed by an additional re-extraction temperature step at ~1750 °C on select samples, which showed that the sample was thoroughly degassed in the 1700 °C extraction. The gas was purified of reactive gases such as hydrocarbons, water, etc. by a series of SAESTM getters that were maintained at various temperatures from 20 up to 350 °C. The gas was then measured in two or three stages, depending on the suitability of measuring all Kr/Xe isotopes for each sample 47

(determined by results of the first aliquot measured for all He and Ne isotopes in the first step, and all Ar and main isotopes of Kr and Xe in the second step).

Argon, Kr, and Xe were separated from He and Ne by freezing them onto activated charcoal at liquid nitrogen temperature. This allowed for the measurement of He and Ne without interference and loss of the other isotopes. The heavy noble gases were separated using two cold traps (-125

°C and -196 °C) to separate Ar from Kr and Xe when sufficient Kr and Xe were expected. Some samples contained enough argon to overload the multiplier. When this happened, an additional step was added to the procedure to dilute the amount of Ar in order to have reliable measurements of the 36Ar/38Ar measured in ion counting mode. The effect of doubly charged interferences (e.g.

40Ar++ and 20Ne+ have the same m/e ratio) was monitored by measuring the amount of 40Ar and

CO2 background in the He and Ne portion of analysis and were found to be negligible.

Table Table Table Table 7 1 O C

See references in Table 2.4 referencesSee Table in

. . 48 l l ) ) C A i C A i T S N G W v T S N n r M F C M K l r M F C M K l i i 2 2 i i o R a i 2 2 o e O a O 2 a e O a O 2 n O O a O O n g O p 2 n g O o e O O R 2 O O e r 2 2 O 2 2 3 O 3 O

O y 2.4. Average composition ofolivine Averagecomposition 2.4. 3 O 3 O d r f

e 2.5. Average composition of clinopyroxene ofclinopyroxene Averagecomposition 2.5. e . r r f o n i . c x

a h e n

e n d t e

K A a l a L . 1 l , H 2 3 3 A l

, 0 0 0 2 e 2 9 3 7 L

y . . . 0 8 H , 0 4 . 1 . . 2 1 5 1 m 3 3 4 1 0 0 6 0 0 0 1 4 5 3 0 0 5 3 , , 5 3 3 ...... 0 0 2 8 e 4 8 2 8 1 . 6 . . 5 9 9 4 . y 2 , 4 8 3 9 6 8 0

3 8 n 8 3 8 5 0 . 2 1 )

5 G 9

8 o B 9 o 3 r . 2 3 3 d , 0 0 0 a B 2 4 8 4 7 3 r

. . . c . i . r , 1 1 5

, 0 4 . 2 . .

c ) h 5 0 0 8 0 0 0 a

5 1 3 0

8 6 3 8 0 5 h M i , ...... c , 0 2 9 n 6 9 2 4 3 . 8 . . 5 6

h 7 3 5 e i a , 6 4 3 1 4 4 t i t 1 9 1 5 n t

a 0 l a e l . f , e E

2 h E 0 l E T d 3 3 3

0 0 0 0 E

t 2 1 1 6 6

. . . T 9 e 2 1 5 , 0 4 . 1 . . . 0 0 6 0 1 0

3 0 2 6 t 9 2 2 5 3

9 2 1 6 ......

9 8 5 0 a 4

, 7 1 3 1 . 0 . . 3

3 5 3 0 in wt.% used for the shielding model. the for shielding wt.% used in 9 ) l 0 6 6 2 3 4 8 . 4 1 1 K 4 , 2

e 2 0 n 0 2

H 0 D 3 u o H

g m 3 3 3 3 u 0 0 0 h

. in wt.% used for the shielding model. the for shielding used wt.% in 3 0 2 7 a ) g . . . e 2 1 5

, 0 4 . 1 . . n G 0 0 6 0 1 0 h 4 3 8 0 s 3 2 5 3 3 4 2 i ...... a e 2 8 7 , 7 1 1 1 . 0 . . 4 k 0 7 5 1 5 r s

4 6 4 4 2 1 p d 2 8 0 3 0 e n 6 2 r e s 6 r o - V n a a n l 3 3 3 R

0 0 0 c d 0 4 7 . . . e 2 1 5 o 4 y R 0 4 . 1 . . 0 0 6 0 0 0 2 2 0 i 2 5 3 m e 3 6 1 d ...... e 4 7 1 5 6 1 2 1 . 7 . . 3 t 0 7 5 i m

1 8 0 8 4 7 d a 0 0 0 u l . n ,

2 i c 0 a N 1 t i 3 W N

2 3 3 o 0 0 0

4 5 6 7 W n 1 2 5 . . . A 7 7 0 0 0 1 0 0 . 0 4 . 0 . .

6 6 4 2 ) 8 3 4

, ...... A

3 6 8 5 G 1 2 4 . 0 1 4 . . 2 0 1 7 1 3 7 2

9 1 1 0 7 1 1 2 1 o 5 1 3 8 5 8 o 9 5 . d

i 0 c N . h ) N W

a 3 3 3 S W 0 0 0 n m 1 2 6 A 2 1 5 . . . 8 0 0 6 0 0 0 d 0 4 . 0 . .

2 5 3 A i 2 8 8 1 ......

8 t 4 3 9 R 6 1 8 1 . 9 . . 3

0 0 0

5 0 6 5 1 i

6 8 0 3 7 5

g 0 0 0 0 5 e h 0 t 0 t a 0 e l r .

, N 2

1 N 0 W 9 3 3 3 0 W 0 0 8 1 1 6 A 2 5 0 . . 0 0 6 0 1 0 3 3 3 . 2 . . 2 3 A .

1 6 7 ...... 3 3 5 8 4 7 1 6 1 . 0 . 3

S 6 8 7 1 1 1 3 . 3 6 8 2 0 8 t ) 4 5 5 1 a J 1 5 n o 1 d h a n r N s d N o W

d n 3 3 3 W 0 0 0 e 2 1 5 0 3 6 A e 0 0 6 0 0 0 . . . v 7 2 5 3 A t 0 4 . 1 . .

...... i 0 3 8 7

5 7 1 2 1 . 9 . . 4 a a 4 2 3

7 3 8 0 0 0 5 1 t l 5 8 0 8 9 7 . 0 0 0 i 1 9 , o

9 1 1 n 1 9

f 7 o 7 N r N

. a W

6 l W 3 3 3 l 0 0 . 2 1 5

0 2 6 A m ) 0 0 6 0 0 0 . .

3 2 5 3 A N 4 . 1 . .

...... 3 2 3 6 e 4 7 1 1 1 . 9 . . 4

1 2 e 7 2 0 4 a 3 8 0 9 3 5 7 1 7 5 h 0 0 2 s 9 6 u r 6 u 9 r 9

e e m t N

a N e W l n . W , t 2 3 3

0 0 0 0 0 0 A 2 1 5 s 1 4 7 7 0 0 5 0 0 0 A

...... 1 6 3 i 9 9 0 4 . 0 0 . . 0 0 ...... s 1 9 2 0 5 8 2 7 2 . 9 . . 4 8 1

4 6 1 7 0 1 0 4 1 < 0 0 6 9 1 4 8 3 6 2 3 0

7 4 7

6 t . 6 h 3 3 a 7 7 n

w A A t 1 2 3 3 0 v 1 1 5 v 0 0 0 0 0 . 9 0 0 7 0 0 0 0 9 3 6 . e % 9 6 3 e . . . . 0 . 9 ...... 0 0 4 . 0 1 . . 0 r 7 2 4 1 . 8 . . 4 r . 2 6 9 0 a . 8 7 4 . 4 3 1 3 1 a 9 1 1 5 6 7 1 4 5 1 9 3 g 7 6 5 g e e

Table 2.6. Average composition of orthopyroxene in wt.% used for the shielding model. Orthopyroxene Brachina NWA 595 NWA 1500 NWA 5191 NWA 10637 Average

Na2O 0.08 0.35 0.01 0.02 0.02 0.10

K2O 0.01 0.01

SiO2 54.90 53.23 54.07 54.20 54.45 54.17 MgO 25.10 16.10 27.13 26.10 27.73 24.43

Al2O3 0.40 0.81 0.14 0.11 0.43 0.38 CaO 2.00 21.79 0.78 1.40 1.08 5.41

TiO2 0.12 0.16 0.03 0.02 0.10 0.09 FeO 16.30 6.14 17.20 17.90 14.56 14.42 MnO 0.42 0.19 0.43 0.41 0.45 0.38

Cr2O3 0.38 0.60 0.10 0.13 0.25 0.29 99.7 Ref. 6 7,11 7,8 7 9 See references listed in Table 2.4.

Table 2.7. Average composition of plagioclase in wt.% used for the shielding model. Plagioclase Brachina EET 99402 Hughes 026 Reid NWA 1500 Average

Na2O 7.99 7.01 7.08 7.60 8.00 7.53

K2O 0.64 0.04 0.05 0.10 0.06 0.18

SiO2 64.42 57.51 58.80 60.00 60.50 60.25 MgO 0.03 0.02 0.03 0.02

Al2O3 23.04 27.11 25.40 24.40 24.90 24.97 CaO 3.67 8.27 7.75 6.61 6.10 6.48

TiO2 FeO 0.38 0.14 0.19 0.11 0.29 0.22 MnO 0.15 0.03 0.02 0.01 0.05

Cr2O3 0.21 0.03 0.02 0.01 0.07 99.8 Ref. 3,6 2,3 7 7 8 See references listed in Table 2.4.

Table 2.8. Average composition of chromite in wt.% used for the shielding model. Chromite ALH 84025 Brachina EET 99402 Hughes 026 NWA 595 NWA 1500 NWA 3151 NWA 4969 Average

Na2O 0.01

K2O

SiO2 0.03 0.07 0.06 0.14 0.03 0.04 0.04 0.06 MgO 3.68 6.01 4.29 4.63 4.96 4.80 4.60 4.48 4.68

Al2O3 7.46 7.69 13.50 12.53 11.48 11.80 12.67 12.36 11.19 CaO

TiO2 1.32 2.71 0.97 0.95 1.19 0.99 0.89 0.74 1.22 FeO 28.41 28.23 28.25 27.45 26.26 27.10 27.59 27.35 27.58 MnO 0.39 0.41 0.31 0.37 0.34 0.37 0.38 0.37

Cr2O3 58.25 53.34 52.55 52.40 54.26 53.40 52.99 52.44 53.70 98.8 Ref. 1,2,3 3,5,6 2,3 8 11 8 3 3 See references listed in Table 2.4.

Table 2.9. Suppliers of brachinite and other ultramafic achondrite samples for this work. Supplier Sample Rainer Bartoschewitz NWA 1500, NWA 7297 Fredric Stephan NWA 6077, NWA 10637 Cyril Lorenz and Marina Ivanova NWA 4518 Larry and Twink Monrad NWA 595 Beda Hoffman RaS 309 Tony Irving NWA 3151, NWA 4874, NWA 4876, NWA 4882, NWA 4969, NWA 6474, NWA 6962, NWA 7605, NWA 8777

Mass fractionation was monitored by measuring a series of calibrated gas mixtures of known composition throughout the experimental duration. Blank corrections were applied to sample results by replicating the sample measurement procedure on Al foil originating from the same source of foil used on samples. Typical blank/sample ratios were less than 1% for 4He (max ~7%),

~1% for 20Ne (max ~5%), much smaller for 3He and 21Ne, and are more considerable and varied for 40Ar (~1-120%, average ~25%), 36Ar (~0.5-53%, average ~9%), and 38Ar (~0.5-11%, average

~3%). Xe and Kr blanks/sample ratios were typically ~5-10% (max. ~58%) for 84Kr and ~3%

(max. 13%) for 132Xe.

Figure 2.10. Samples were stored in a rotating sample tree Sample holder above the furnace. Sample gases were isolated as desired using manual valves and cold traps on the way toward the mass spectrometer (MS). Manual Valves

Furnace Cold Trap

CRE ages were calculated based on production rates determined by the shielding model developed by Leya and Masarik (2009), in conjunction with the measured cosmogenic shielding 51 indicator 22Ne/21Ne of the respective sample. This determines the appropriate combinations of pre- atmospheric radius of the meteoroid along with the depth from the surface that our representative sample could have been located. The production rate is dependent on the chemical composition of the meteoroid, and it is therefore necessary to consider the bulk chemistry of the samples to be properly determined. The average bulk and mineral chemistry of brachinites (Tables 2.4-2.8) was used for the target chemistry for samples that do not have this data available in the literature. One of the aliquots of NWA 4969 has missing neon measurements due to experimental error. When it was necessary to separate Kr and Xe from Ar, typically, ~ 3% of the Ar was detected in the Kr/Xe phase, 2% of Kr and negligibly 0-1% of Xe in the Ar phase. The Ar and Kr amounts were corrected for (equation 2.19). The cosmogenic and trapped portions were determined using a 2-equations, 2- unknown decomposition (equation 2.20, where Arcorr is the correction to the given isotope, Arccstp

40 40 is the measured value in ccSTP, ArKrXe is the amount of Ar measured in the krypton and xenon

40 40 portion, and ArAr is the amount of Ar in the argon portion). This assumes the pure trapped component is 36Ar/38Ar = 5.34 ± 0.02, and the pure cosmogenic component is 36Ar/38Ar = 0.65 ±

0.02 (Busemann et al., 2000). Once the cosmogenic portion is determined, the trapped component of the argon isotopes can be solved for.

Figure 2.11. Extraction line, magnet (blue), and electronics of the Albatros Mass Spectrometer.

Eq. 2.19 :

Eq. 2.20:

2.5 Oxygen Isotopes

Oxygen is an abundant element that has three stable isotopes that are subject to strong mass-

dependent fractionation. A three-isotope plot of oxygen shows the relationship between the

abundances of the minor isotopes (17O, 18O) relative to the most abundant isotope (16O). Oxygen

variations are usually quite small and therefore using delta ‘’ notation makes it much more

convenient to analyze these small differences. In this isotope plot, terrestrial materials plot along

a line of slope 0.52 (Matsuhisa et al., 1978), called the terrestrial fractionation line (TFL).

Departures from this line are common in meteorites and may reflect mixing between oxygen

reservoirs and aqueous alteration, which enhances the heavier isotopes (Clayton, 2003). 53

Eq. 2.21:

--Where ‘x’ refers to the isotope (17 or 18), SMOW (Standard Mean Ocean Water) is a reference value, and  values are in per mil (‰).

Figure 2.12 (Left) Oxygen isotope plot with range of oxygen compositions for main chondritic groups. (Right) The addition of the various oxygen isotopes shifts the fractionation line as shown. Figures from Clayton, 2003.

It is also common to use 17O to represent the departure from the TFL instead of 17O, conveniently plotting samples from the same oxygen reservoir (and perhaps same parent body) along horizontal lines (slope = 0).

Eq. 2.22:

For meteorites from differentiated parent bodies, 17O is a powerful tool for confirming a common source. For example, SNC meteorites have a variety of lithologies but have the same oxygen signature, 17O = 0.30 ± 0.07 ‰ (uncertainty of the oxygen values are the standard deviation). The HED meteorites, thought to be from asteroid Vesta, have 17O = -0.25 ± 0.08 ‰.

However, there are different groups that have approximately the same oxygen values

(mesosiderites 17O = -0.24 ± 0.09 ‰, pallasites 17O = -0.28 ± 0.06 ‰) so independent data 54 needs to be combined with oxygen isotopes for better understanding their relationships (Clayton and Mayeda, 1996). Primitive achondrites or achondrites that were not fully differentiated may retain chemical and isotopic heterogeneities acquired from their source material, making parent body distinctions in 17O more ambiguous (Fig. 2.13).

17 Figure 2.13.  O isotope composition of primitive achondrites. Individual data points are anomalous or ungrouped and the shaded regions represent the range of oxygen values represented by main primitive achondrite groups. The five samples circled in purple are included in this study. Red/white data points show brachinite-like values overlap with the brachinite group (gray shaded region). Figure modified from Greenwood et al., 2017.

2.6 Samples

This section does not include the ureilites or Chelyabinsk because no physical samples of ureilites were studied in this work and there was not enough sample of Chelyabinsk for additional measurements. Many studies have focused on the composition and details of ureilites (e.g.

Mittlefehldt et al., 1998; Cohen et al., 2004; Goodrich et al., 2004) and of Chelyabinsk (e.g. Popova 55 et al., 2013; Badyukov et al., 2015). Ultramafic achondrite (UMA) descriptions below are primarily from literature sources, with element maps and BSE images of samples that had enough material to do so. The impetus for the backscatter and element mapping was to search for host phases of potassium. Samples were analyzed with a CAMECA SX50 and SX100 microprobes under the guidance of K. Domanik. Analysis conditions were 15 kV, 20 nA for olivine and pyroxene, and a defocused 5 m beam at 15kV for potential sites of plagioclase. No potassium was found in the samples studied, more details can be found in Beard et al. (2015).

The problematic classification of UMA partially motivates this work, since similar ages may support an origin from the same parent body (more details on ages/parent bodies later in this section). A review of brachinites from Keil, 2014 helps to clarify or at least highlight some of the confusion. Descriptions of the sixteen UMA samples that had noble gases measured in this work

(NWA 595, NWA 1500, NWA 3151, NWA 4518, NWA 4874, NWA 4876, NWA 4882, NWA

4969, NWA 6077, NWA 6474, NWA 6962, NWA 7297, NWA 7605, NWA 8777, NWA 10637, and RaS 309) in addition to noble gas data from literature samples are below.

NWA 595 (Figs. 2.14 and 2.15) is generally considered a brachinite-like achondrite because of the relatively high abundance of orthopyroxene (10-15% En72Fs25Wo2.2) and more magnesian olivine (Irving et al., 2005; Goodrich et al., 2010, Day et al., 2012). Olivine (~80 ± 10 vol. %,

Fo71.7, FeO/MnO 52 ± 6) has fine grained intergrowths of orthopyroxene and metal along grain boundaries (Goodrich et al., 2006). These intergrowth assemblages are also seen in Reid 013,

Hughes 026, NWA 1500, NWA 5191, NWA 4882, NWA 4874, NWA 4969, and NWA 4872

(Goodrich et al., 2006, Rumble et al., 2008). However, the intergrowths are not seen in Brachina.

Grain sizes range from 0.5-1 mm, and it has a polygonal-granular texture with alignment in 56

elongated grains. Other minerals present include augite (En45Fs10Wo45), chromite (Cr/[Cr + Al] =

0.77), and traces of Ni-rich metal and kamacite.

Figure 2.14. NWA 595 Si map, equigranular olivine (dark grey), interstitial orthopyroxene (light grey), and chromite (black). Pyroxenes tend to group together, forming a series of equigranular regions that are up to 4mm across Scale bar is 5 mm. Blue square outlines the approximate area shown in Figure 2.15.

Figure 2.15. NWA 595 BSE image with equigranular olivine (100-400 microns, light grey), interstitial orthopyroxene (dark grey), and chromite (white). Fe-Ni metal is often found along grain boundaries. Scale bar is 1mm. 57

NWA 1500 (Figs. 2.16 and 2.17) was originally classified as a ureilite (Bartoschewitz et al.,

2003), and is listed in the Meteoritical Bulletin as an ungrouped achondrite, but has more recently been considered a brachinite. Goodrich et al. (2010) argue that the similarities in texture, mineral and REE abundances, and elemental and oxygen compositions with the range of brachinites suggest that it is a brachinite. However, they note that the main difference is the presence of reduction rims on olivine as shown in grain boundary darkening in transmitted light, a feature that was unique to ureilites and unlike other brachinites at this time (NWA 7605 was also shown to contain these reduced rims in a later study (Irving et al., 2013)). The reversed zoning and grain boundary assemblages are attributed to reduction processes (e.g. orthopyroxene in assemblages from Mg2SiO4 + Fe2SiO4 = 2MgSiO3 + 2Fe + O2, and decreasing fugacity conditions for the reversed zoning; Goodrich et al., 2006; 2011). Perhaps NWA 1500 and NWA 7605 are related to one another.

Figure 2.16. X-ray (colors relate to intensity in Z) map of NWA 1500: subhedral olivine (light green) with interstitial clinopyroxene (dark green), chromite (red), and metal (yellow). Black outline is the approximate area in Fig. 2.17. Scale bar is 5 mm.

Figure 2.17. BSE image of NWA 1500, centered on a ~250 micron diopside grain (dark grey) interstitial in olivine. Fine grain orthopyroxene and metal form fine-grained intergrowths in between olivine grains. Scale bar is 100 microns.

NWA 6077 (Fig. 2.18) has been classified as brachinite-like achondrite based on higher magnesian composition of its olivine (Fo69.3-69.8) and presence of orthopyroxene (Day et al., 2012).

Its olivine has a protogranular texture with triple-junction boundaries. Other minerals include clinopyroxene, kamacite, chromite, chlorapatite, and Ni-bearing troilite (Irving and Kuehner,

Meteoritical Bulletin Database 2018).

NWA 6962 is classified as a brachinite-like ungrouped achondrite (Meteoritical Bulletin) and has similar texture and modal abundances as brachinites. However, its olivine and pyroxene are enriched in Fe (Fo53 for olivine, Mg# 0.68 for pyroxene) compared to most brachinites (~Fo64-73,

~Mg# 0.79-0.82) (Keil, 2014, Dunlap et al., 2015). The Fe/Mg ratios of NWA 6962 are outside the range of primitive achondrites (Goodrich and Delaney, 2000) and its oxygen isotopes (Clayton and Mayeda, 1996; Meteoritical Bulletin Database 2018) are near the range of ureilites. Other minerals include plagioclase (An21-31Or0.5), merrillite, kamacite, and sporadic Ti-poor chromite. 59

NWA 7297 (Figs. 2.19 and 2.20) is a highly weathered meteorite that has grain sizes up to 1 mm, olivine (~80 vol.%, Fo69.5), relatively large interstitial chromite up to ~0.5mm in size, clinopyroxene (~15 vol.%), and opaques (Bartoschewitz et al., Meteoritical Bulletin Database

2018).

NWA 4518 is an ultramafic brecciated achondrite that has olivine composition consistent with the brachinites but overall silicate chemistry similar to the winonaites and HEDs (Lorenz et al.,

2011). However, the oxygen composition (Greenwood et al., 2012) is much closer to the brachinites than the winonaites. The relationship between NWA 4518 and the brachinites is not well known and therefore has been added to this study. This sample is described as recrystallized equigranular polymict breccia with olivine (Fo68), minor pyroxene, Fe-Ni-metal, and sulfide-rich veins (Lorenz et al., 2011).

Figure 2.18. Mg X-ray map of NWA 6077, dominated by olivine (yellow) with clinopyroxene (blue), and orthopyroxene (green). Scale bar is 5 mm.

Figure 2.19. NWA 7297 X-ray (Z) image dominated by olivine (green) with clinopyroxene (Cpx), orthopyroxene (Opx), and chromite (red). Black outline is the approximate area shown in Fig. 2.20.

Figure 2.20. NWA 7297 BSE image showing clinopyroxene grains (dark), ~200-400 microns in size, interstitial in olivine (light grey). Grains are bordered by metal veins. Grey- white areas are chromite. Scale bar is 200 microns.

NWA 3151 has coarse-grained equigranular olivine (0.6-1.2 mm, Fo65), and contains clinopyroxene, troilite, chromite, altered metal (Gardner-Vandy et al., 2013), orthopyroxene, and

K-poor glagioclase (Irving et al., 2005). Siderophile elemental abundances are similar to other brachinites (Day et al., 2012).

NWA 4874 has medium grain sized (~0.6 mm) olivine (90 vol.%, Fo66), with clinopyroxene, plagioclase, chromite, and merrillite (Rumble et al., 2008).

NWA 4882 has been paired with NWA 4969, and is fine to medium grained (0.1-1.2mm) moderately shocked sample (Keil, 2014). It has a protogranular texture, olivine (85-90 vol.%,

Fo65), clinopyroxene, K-poor plagioclase, chromite, and fine-grained orthopyroxene intergrowths

(Rumble et al., 2008). Siderophile abundances show patterns similar to other brachinites (Day et al., 2012). Although this sample is suggested to be paired with NWA 4969 (Keil, 2014), this work demonstrates that these samples are not paired, based on different exposure ages.

NWA 4969 is fine-to-medium grained (0.06-0.7 mm) with weak triple junctions of olivine (~90 vol.%, Fo65.5), interstitial clinopyroxene found with chromite, and also contains plagioclase and troilite (Rumble et al., 2008; Gardner-Vandy et al., 2013).

NWA 6474 closely resembles NWA 4882 in composition and mineralogy but has distinct oxygen isotopes (Irving and Kuehner, Meteoritical Bulletin Database 2018). Minerals include olivine (Fo64, with variable Fe/Mn = 71-76), clinopyroxene, chromite, troilite, chlorapatite, and minor plagioclase.

NWA 10637 has texturally equilibrated triple junctions of olivine (87 vol.%, Fo73), orthopyroxene (10 vol.%), clinopyroxene (3 vol.%), and minor chromite. Oxidized Fe-Ni metal veins are found along grain boundaries (Domanik and Stephan, Meteoritical Bulletin Database

2018). 62

RaS 309 (Fig. 2.21 and 2.22) is weakly shocked (S2) and has an equigranular texture with a grain size of ~0.5mm, is composed of olivine (~94 vol.%, Fo67), clinopyroxene, and chromite

(Hoffman, Zurfluh, and Gnos, Meteoritical Bulletin Database 2018). Classification of all the samples discussed in this work are listed in Table 7.2.

Figure 2.21. BSE image of RaS 309, dominated by olivine (grey) with clinopyroxene (black), and chromite (white). The blue square outlines the approximate area in Fig. 2.22.

Figure 2.22. BSE image of diopside in RaS 309 shows metal within the grain (white spots). Olivine (light grey) and diopside (dark grey) are bordered by metal veins. 63

Chapter 3: Ar-Ar Data Reduction Example

3.1 Gas Extraction

The process of calculating accurate Ar-Ar ages takes effort and can be confusing. Here I present a step-by-step data analysis and experimental procedure to better illustrate the corrections for a real sample; Chelyabinsk SB5, a melt-rich sample of an LL5 ordinary chondrite.

After the sample has been irradiated and loaded into the mass spectrometer, an ‘oven’ is constructed around the entire extraction line while under vacuum and heated to ~200°C. This reduces the vacuum and cleans off any minor contamination that was introduced while the extraction line was exposed to atmosphere during sample loading. After ‘baking’ the system, blank levels are measured to determine if the system is clean enough for sample analysis. A set of four hot blanks are measured before the sample (at 295,1000, 1250, and 1505°C), before starting a 24- step heating schedule (300-1500°C). This is followed by an additional set of hot blanks that covers the same temperature range as the first. The released gas from each step (from the hot blank or the sample) is analyzed by the mass spectrometer, typically for 10 cycles of 8 measurements and regressed to t0 (Fig. 3.1). At the beginning of each cycle of measurements, the magnet re-centers on 40Ar to ensure that the magnetic calibration is on target (Fig. 3.2).

The process for determining the irradiation parameter follows the same procedure as above, but with fewer heating extractions of the standard (Hb3gr hornblende). Because the standard has a known age (1081.0 ± 1.2 Ma), the irradiation parameter J, which is needed to determine the age of the sample, can be calculated.

Eq. 3.1:

--Where  is the 40K decay constant, t is the age of the standard, and the 40Ar/39Ar is measured

(see equations 2.10-2.15 for more details). 64

3.2 Data Corrections

At this point, all the data required to calculate the sample age has been acquired. The sample and blank data are separated into three ranges, 300-1000°C, 1000-1250°C, and 1250-1500°C, with the blank value of each isotope in the respective temperature range interpolated and removed from the sample value. This is referred to as blank-corrected data (Fig. 3.3).

Figure. 3.1. This isotope plot shows the pumpout effect of the mass spectrometer and represents one full cycle of isotopic measurements for one temperature step of SB5. Eight measurements are averaged at each time and are cycled through each isotope 10 times. The y-axis is in 1e-6 ‘counts’. Unfilled shapes are measurements excluded from the regression.

Figure. 3.2 (Top) At the beginning of each cycle the magnet re-centers on the 40Ar peak to ensure proper magnet calibration. (Bottom) 40Ar blank measurements range from ~8e4 counts at 300°C, up to ~2e5 counts at 1500°C.

The blank contribution of 40Ar for SB5 is 1.24% at 295°C, 1.23% at 1000°C, 6.0% at 1250°C, and ~100% at 1500°C. Although the latter may not seem acceptable, it is actually a confirmation that at 1500°C the sample has completely degassed. Figure 3.3 shows the relative amounts of blank-corrected sample gas released at each step between 300-1000°C compared to the blank.

Blank

39 Figure. 3.3. Example of blank corrected Ar data (the blank has already been subtracted from the data here). The blue/green data points at 295°C and 1050°C are the blank measurements and the red data points are the sample measurements.

The Mass Spec software can now generate the apparent age spectra (Fig. 3.4) which includes corrections for 39Ar and 37Ar decay, where each step is labeled with the extraction temperature with the associated Ca/K ratio plotted separately below the age. The Ca/K ratio can be useful in interpreting the released 39Ar, and in this case shows a relatively higher potassium mineral phase at low temperatures and a lower potassium (high calcium) phase at higher temperatures. 67

Figure. 3.4. Preliminary age spectra (plateau plot) of SB5, with the age plotted versus the amount of released 39Ar. Each step is labeled with the extraction temperature and includes the overall integrated age calculated (2949 ± 13 Ma). Below the age spectra is the Ca/K released as a function of 39Ar, where each step corresponds to the appropriate extraction temperature as seen in the plateau plot.

Isochron plots can now be used to get an idea of what trapped 40Ar/36Ar component(s) the sample may contain. It is not uncommon to see more than one trapped phase in the isochron

(Korochantseva et al., 2007), which often includes terrestrial atmosphere (40Ar/36Ar =298).

Isochron subsets are determined from linear portions of the data. The isochron plots for SB5 show a terrestrial trapped component both at high and low temperatures (Fig. 3.5, 3.6). Interestingly, there are two separate series of continuous temperature steps that both show an initial terrestrial trapped component (300-500°C, and 1000-1500°C—shown in figure 3.6). The intermediate steps

(~600-900°C) result in an exceptionally high trapped component of several thousand. Because these steps reflect diffusive loss and are difficult to interpret, they are not considered in the age determination. 68

Figure. 3.5. Isochron plot for all SB5 data with no single clear trapped component indicated. The errors for the data points are shown as error ellipses.

Figure. 3.6. Isochron plot for subset of SB5 data (1000-1500°C), indicates a trapped component with a 40Ar/36Ar ratio ~250 ± 40.

Viewing the reverse isochron can be useful if the isochron data are hard to interpret. Isotopes normalized to 40Ar (as is done in a reverse isochron plot) can lead to better uncertainties than when normalizing to 36Ar. This was not the case with SB5, but the reverse isochron is shown here (Fig.

3.7) in a lower temperature range (300-500°C) than shown in the standard isochron (Fig. 3.6) for comparison.

Figure. 3.7. Isochron plot for subset of SB5 data (300-500°C), indicates a trapped component with a 40Ar/36Ar ratio ~280 ± 30.

The isochrons shown for SB5 do not include the effects of cosmic-ray spallation. The meteoroid is exposed to galactic cosmic rays during transit to the earth, which interact with the surface material by creating a cascade of neutrons and other particles that penetrate further into the meteoroid. These neutrons produce 38Ar and 36Ar (cosmogenic 36Ar/38Ar = 0.65, compared to 70

36Ar/38Ar = 5.32 in the earth’s atmosphere) that need to be corrected for (Wieler, 2002). This can be solved by deconvolution of each component (spallation and earth atmosphere) with the determined spallogenic amounts subtracted from the appropriate isotope measurements.

Interference corrections for the reactor production of 40Ar and 39Ar still need to be considered.

Potassium and calcium monitors (see Chapter 2) that were included with the samples during irradiation are used to make the appropriate corrections. Correction for 40K(n,p) 40Ar is achieved by assuming all the 39Ar measured in the potassium monitor is from reactor production, which allows a correction for the amount of reactor produced 40Ar (using the measured 40Ar/39Ar). A similar process is done with the calcium monitor, which assumes all 37Ar measured in this sample is from reactor production, 40Ca(n,)37Ar. The measured 39Ar/37Ar, and 36Ar/37Ar ratios from the calcium monitor are then used to correct the amount of sample 39Ar and 36Ar by removing the effects of 40Ca(n,n)36Ar, and 42Ca(n,)39Ar.

Now that the proper corrections have been made, isochron plots can be appropriately used to determine the trapped components. The corrected isochrons are provided (Figs. 3.8, 3.9), which use a method of least squares to provide the best fit (using source code by Philip Kromer based on

Reed, 1989). This method determines the line that minimizes the distance between all the data weighed by the inverse square of the uncertainty. Low temperature steps (300-500°C) indicate an initial 40Ar/36Ar of ~285 ± 6, and higher temperature steps included in the partial plateau age (950-

1175°C) indicates an initial 40Ar/36Ar of ~550 ± 180. 71

Figure. 3.8. Isochron plot for SB5 300-500°C temperature steps. Trapped component of 40Ar/36Ar ~ 285, which is probably terrestrial air ~300.

Figure. 3.9. Isochron plot for SB5 950-1175°C temperature steps. Trapped component of 40Ar/36Ar ~ 550

The trapped amount is subtracted from the interference and spallation corrected 40Ar to get the final values believed to best represent the sample. This is done for each temperature step (24 steps from 300-1500°C). Finally, the apparent age of each step is calculated and plotted as a function of the released 39Ar (Fig. 3.10). Although it looks similar to the uncorrected age spectrum (Fig. 3.4), the first step has been reduced considerably and the maximum ages are ~150 Ma younger. This process was repeated for all Ar-Ar measurements. The other two splits from the same sample as

MB020f,5 behave very similarly and will be discussed in more detail in Chapter 4.

39 Figure. 3.10. Final apparent age spectra of SB5. The partial plateau at > 60% Ar is around 2700 Ma, and is discussed further in the next chapter.

Chapter 4: Ar-Ar of Chelyabinsk

4.1 Introduction

Chelyabinsk is a metal-rich LL5 chondrite (Popova et al., 2013) that fell over the Chelyabinsk region of Russia in early 2013 (~500 kg TNT equivalent, Brown et al., 2013). The recovered samples have light clast-rich and dark melt-rich lithologies (Fig. 4.1) (Galimov et al., 2013,

Badykov et al., 2015). The light clast-rich lithology is described as a typical chondritic texture of petrographic type 5, shock stage 4, and contains 1-2 mm thick melt veins (Badyukov et al., 2015).

The dark melt-rich lithology contains impact melt that sometimes appear as dikes up to a few centimeters wide (Badyukov et al., 2015). Silicate minerals in Chelyabinsk are olivine, orthopyroxene, plagioclase, and clinopyroxene (decreasing order of abundance), and minor phosphates, ilmenite, troilite, chromite, and pentlandite. Metal phases include kamacite and taenite

(Popova et al., 2013; Badyukov et al., 2015). Shock features include undulatory extinction and

Figure 4.1. (Left) Example of Chelaybinsk lithologies (Trieloff et al.., 2015), ruler is in cm. (Right) Light lithology with black lines highlighting a network of veins (Badykov et al., 2015)

74 moderate to strong mosacism in olivine, and transformation of plagioclase into glass (Badyukov et al., 2015; Righter et al., 2015). The dark lithology is suggested to be the result of ‘shock darkening’, where pore spaces in silicate grains are filled with metal and sulfide-rich melt

(Bischoff et al., 2013; Galimov et al., 2013, Righter et al., 2015), which may affect Ar-Ar dating results by altering minerals (Trieloff et al., 2017). The formation of the brecciated nature of

Chelyabinsk is attributed to a single energetic impact event based on petrology (Badyukov et al.,

2015; Petrova et al., 2016), however, geochronology studies of Chelyabisnk have determined multiple ages (Table 4.1) that have been interpreted to represent at least eight impact events

(Righter et al., 2015). This chapter adds additional Ar-Ar measurements of Chelyabinsk and will discuss the implications of these ages.

Table 4.1. Chelyabinsk ages determined by various isotopic dating methods. All ages are given in Ma. Method U-He Ar-Ar K-Ar Rb-Sr Sm-Nd U-PB Re-Os Age ~271 ~303 865 ± 971 153 ± 585 ~2906 585 ± 3908 455812 1,2 4 1 1 9 312 ± 6 1000 880 ± 120 2900 ± 500 834 ± 7 1 1 1 7 9 716 ± 30 1945 ± 168 1400 ± 300 3733 ± 110 2744 ± 13 1,2 1 5 9 1014 ± 24 1952 ± 169 4567 2861 ± 15 1,2 1 10 1184 ± 40 2736 ±199 4433 ± 110 3 11 1700 ± 100 4452 ± 21 4454 ± 678 1. Righter et al. (2015), 2. Lindsay et al. (2015), 3. Trieloff et al. (2017), 4. Haba et al. (2014),

4.2 Ar-Ar Analysis

An overview of the general Ar-Ar methods for this work can be found in the chapters 2 and 3, with specific and additional details for the Chelyabinsk analysis provided here.

Both Chelyabinsk samples, clast rich MB020f,2 and melt rich MB020f,5, were divided into multiple splits of ~10 to 14 mg for Ar-Ar analysis. The splits were irradiated at the Cadmium- 75

Lined In-Core Irradiation Tube (CLICIT) facility at Oregon State University along with PP-20

Hb3gr Hornblende GSC (1080.4 ± 1.1 Ma, Renne et al., 2011) as the primary irradiation monitor to determine the J-factor; roughly 1.05×10-3 among our samples

Samples were irradiated for four hours in the linear-with-height portion of the reactor to provide a more evenly distributed neutron flux. Samples were then allowed to cool for about two months to allow for short-lived isotopes to decay before being placed in a glass storage tree above a double-vacuum, resistance-heated furnace. Argon was extracted from each sample using a computer-controlled step-heating procedure. Typical temperature steps were from 300-1500 °C in

25-200 °C intervals depending on the amount of argon released and the accuracy desired. The total number of heating steps varied between 16-32 per split.

The decay constants used are those reported in Renne et al. (2011). Use of the decay constants given by Steiger and Jäger (1977) would yield apparent ages that are slightly younger, by up to 6

Ma for the oldest apparent ages, but within the quoted uncertainties. As detailed in Chapter 3, uncertainties and corrections were applied to account for blanks, machine discrimination, spallation-produced isotopes, and interfering isotopes produced in the reactor from Ca and K.

Isotopic measurements were regressed to a time of zero using linear regression techniques, after subtracting baseline values in order to get the most accurate measurements (Fig. 4.2). It is worth noting that the first 1-2 data points were taken out of the regression for 40Ar for some samples due to their peculiar behavior. For unknown reasons the first one or two cycles seemed to be off of the fit and do not match the trend of the ~8 subsequent measurements. An extra effort was made to ensure that the peak was being properly centered and no system errors could be found. For consistency, the first cycle was also removed from temperature steps that did not appear to have the first cycle complication. 76

40 Figure 4.2. Regression of Ar to t0. For some samples the first cycle did not measure properly and was not included. Residual plot at the bottom shows the variation of the data from the best fit.

4.3 Ar-Ar Results of Chelyabinsk

Isochron plots were evaluated for the presence of trapped argon, showing at least two trapped components in both samples. MB020f,2, isochrons imply trapped 40Ar/36Ar ratios of 300-400 for temperatures below ~600-700°C (Fig. 4.3, Table 4.2). The standard terrestrial air correction (~298) was not used because all splits show evidence of a higher trapped ratio, although using it does not significantly alter the results. A separate trapped component was identified at higher temperatures and varied from split to split, ranging from ~50-70 ± 20. The isochron for CH1 follows a different behavior than the other splits, implying terrestrial air in the first two steps, followed by several extractions 375-475°C that appear to have a trapped component of 40Ar/36Ar = 65 +/- 25. The reason for this difference is not well understood; the sample has many more heating steps but that should not explain the different result. 77

Figure 4.3. Isochron plots for 6 Chelyabinsk splits. The top 4 are from the clast rich sample MB020f,2 (MB,2), and the bottom 2 are from the melt rich sample MB020f,5 (MB,5). The trapped component was determined by using the 40Ar/36Ar intercept of a least-squares fitting. Split SB5 of MB,5 seems to have a different trapped component (~ 550, see chapter 3) but has essentially the same age.

Figure 4.4. Plateau plot of split SB4 from MB020f,5 (top) with the K/Ca release (bottom).

Figure 4.5. Plateau plot of split SB1 from MB020f,2 (top) with the K/Ca release (bottom).

Although the uncertainty in the isochron data is problematic (see supplementary material), an intercept of ~65 ± 25 is a solution to the least-squares fitting and is at least a suggestion for CH1.

Using a higher trapped correction that encompass air, results in ages ~ 10 Ma younger, which would bring the bottom of the plateau to ~10 Ma rather than ~20 ± 2 Ma. Because the other splits show a trapped component of roughly the same composition at higher temperatures and because the data seems to properly represent the split, this correction (60 ± 25) is used for CH1. Note that the resulting ages from different splits using their different respective trapped correction are nearly in agreement.

The isochrons of MB020f,5 showed similar behavior as those discussed for MB020f,2. Two different trapped components were used, 300 ± 10 (encompassing terrestrial air) was used for

MB020f,5 for low temperatures, and a ratio of ~100 ± 50 was applied to higher temperatures (>

800°C). A different trapped component (~ 550) for split SB5 was found at higher temperatures

(see chapter 3) that was not evident in the other splits. If a correction of ~100 is used instead of the air value for SB5, the partial plateau and summed age change by less than 100 Ma. The isochron ages of the splits are similar, despite the different trapped corrections (Table 4.2). The higher temperature trapped component from both samples may reflect excess 40Ar from incomplete degassing (see discussion).

Sample ages were determined based on 40Ar-39Ar apparent age spectra (“plateau plots”).

Examples of plateau plots with no trapped correction can be found in Figs. 4.4 and 4.5, which also show the K/Ca released as a function of temperature. If the definition of a plateau as a region that contains at least 50% of the 39Ar and at least three consecutive temperature steps that agree at a 95

% confidence level (Deino, 2001) is followed, then none of the splits have a plateau. However,

“partial plateaus” that incorporate a smaller fraction of the total 39Ar, but contain the majority of 80 the 39Ar from the portion that has high K/Ca ratios . Ages for partial plateaus are given at the 2σ level and are calculated using the mean of all the steps in the plateau region, weighted by the amount of gas contained in each step.

Table 4.2. Trapped corrections and Isochron Ages for Chelyabinsk. Sample Split ID Temp. (°C ) 40Ar/36Ar trapped Isochron Age (Ma) MB020F,2 Clast Rich SB1 < 700 321 ± 10 45 ± 12 > 700 50 ± 10 SB2 < 700 404 ± 25 27 ± 25 > 700 70 ± 20 SB3 < 600 401 ± 20 78 ± 10 > 600 60 ± 30 CH1 < 400 300 ± 10 19 ± 3 > 400 65 ± 25 MB020F,5 Melt Rich SB4 < 800 300 ± 10 2673 ± 195 > 800 101 ± 36 SB5 < 800 300 ± 10 2656 ± 392 > 800 550 ± 180 CH2 < 800 300 ± 10 2713 ± 348 > 800 99 ± 68

Table 4.3. SB 1 Argon values and age for by extraction step. Sample mass is 10.06 mg, J = 1.062e-3. Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 250 1274.3 ± 4.4 1.035 ± 0.012 0.794 ± 0.013 2.652 ± 0.037 3.952 ± 0.051 * * 2 300 371.3 ± 3.0 1.018 ± 0.012 0.225 ± 0.005 1.397 ± 0.027 0.992 ± 0.038 57.8 35.1 3 350 302.7 ± 3.0 1.591 ± 0.015 0.186 ± 0.005 1.690 ± 0.035 0.797 ± 0.035 34.2 20.4 4 400 218.3 ± 2.9 2.531 ± 0.018 0.143 ± 0.005 2.770 ± 0.101 0.331 ± 0.035 85.6 10.6 5 450 111.3 ± 2.7 2.758 ± 0.021 0.126 ± 0.007 2.975 ± 0.073 0.237 ± 0.036 28.4 10.2 6 500 210.3 ± 2.5 2.420 ± 0.017 0.182 ± 0.007 2.396 ± 0.063 0.588 ± 0.035 12.4 12.4 7 550 76.4 ± 2.5 1.973 ± 0.016 0.077 ± 0.005 2.328 ± 0.067 0.157 ± 0.031 26.7 12.3 8 600 66.0 ± 2.5 1.505 ± 0.014 0.060 ± 0.006 2.014 ± 0.057 0.110 ± 0.031 41.7 15.9 9 650 82.9 ± 2.4 1.102 ± 0.012 0.050 ± 0.006 1.793 ± 0.053 0.108 ± 0.031 84.4 20.8 10 700 122.2 ± 2.4 0.609 ± 0.010 0.051 ± 0.004 2.063 ± 0.073 0.132 ± 0.030 333.5 10.0 11 750 166.9 ± 2.5 0.371 ± 0.008 0.073 ± 0.006 2.204 ± 0.067 0.143 ± 0.032 686.9 17.8 12 800 229.3 ± 2.5 0.249 ± 0.007 0.079 ± 0.006 2.703 ± 0.067 0.224 ± 0.033 1198.9 29.2 13 1000 637.2 ± 3.0 0.432 ± 0.009 0.328 ± 0.008 17.320 ± 0.215 0.845 ± 0.038 1665.9 27.3 14 1200 620.6 ± 2.3 0.299 ± 0.008 1.173 ± 0.020 51.385 ± 0.390 4.057 ± 0.040 1816.8 105.7 15 1400 238.0 ± 10.4 0.089 ± 0.007 0.513 ± 0.037 12.299 ± 0.174 2.224 ± 0.119 1804.2 239.2 16 1500 52.0 ± 11.3 0.024 ± 0.007 0.012 ± 0.040 0.190 ± 0.034 0.091 ± 0.122 2059.4 495.1 Total 247.1 6.4 *Null Value. All isotope values are in 1e-8 ccSTP/g and ages are in Ma.

The plateau plots for each set of splits from a single sample look very similar to one another, but vary significantly from one sample to the other. The four splits of the clast rich (“light lithology”)

MB020f,2 (SB1, SB2, SB3, and CH1) have low apparent ages for the first ~85% of the total 39Ar released, before increasing, typically to ~2000-3000 Ma, in the highest-temperature extractions.

Three of the splits give a series of concordant ages from about 35-50% of the 39Ar released with an age of 26.1 ± 11.0 Ma, which represents our best age for MB020f,2.

Much of the gas was released at low temperature steps, and consequently, the number of temperature steps was modified from 16 to 32 to better determine the age. The following results are presented in order by split name from MB020f,2.

SB1: The apparent age spectra of SB1 has a partial plateau of 28.9 ± 6.2 Ma (three temperature steps 450-550°C), containing ~40% of the released 39Ar (Fig. 4.6). This partial plateau also corresponds to its absolute minimum age. The summed age (equivalent to the K-Ar age) is 247.1

± 6.4 Ma. More details are given in Table 4.2, see supplementary data for similar tables of other samples.

SB2: It was not known beforehand which temperatures would correspond to the majority of the released argon. For the first analysis of a sample we use a ‘typical’ heating assignment that anticipates that the majority of the argon is released at intermediate temperatures (Fig. 4.7). The first Chelyabinsk sample measured was SB2, which was analyzed with the typical heating schedule that proved to not be ideal; most of the gas was released at lower temperatures than expected. In the plateau plot there are two steps that contain 52% of the released 39Ar and have a weighted age of 29.7 ± 9.7 Ma. 82

SB3: This split has slightly higher ages than the other three splits, with a minimum age of 47 ±

38.7 Ma (Fig. 4.8). Just over 40% of the gas was released in 3 steps from 400-500 °C with an age of ~68 ± 7 Ma. However, we do not quote this as a best age since the steps disagree with the other three splits of the sample. It is possible that a portion of this split was not degassed as thoroughly as the others in the resetting event, or the trapped component was not completely accounted for.

The summed age is 257.9 ± 22.5 Ma.

CH1: The split with the most heating steps, CH1, has a partial plateau of 19.8 ± 2.3 Ma, representing about 37% of the 39Ar (375-475 °C). The summed age is 259.5 ± 12.6 Ma (Fig. 4.9).

When determining our best age of the overall sample (Fig. 4.10), the average of the three splits that agreed (SB1, SB2, and CH1) was taken, with the standard deviation for the uncertainty, which results in 26.1± 11.0 Ma. This represents the upper limit to the most recent impact event.

Figure 4.6. The apparent age spectra of SB1 has a partial plateau of 28.9 ± 6.2 Ma (three temperature steps 450-550°C), containing ~40% of the released 39Ar. The summed age (equivalent to the K-Ar age) is 247.1 ± 6.4 Ma.

Figure 4.7. The apparent age plot for SB2 has two steps that have 52% of the released 39Ar and has a weighted age of 29.7 ± 9.7 Ma (the standard deviation of the two steps is ~16 Ma). The summed age is 260 ± 12 Ma.

Figure 4.8. SB3, has slightly higher ages than the other three splits. Just over 40% of the gas was released in 3 steps from 400-500 °C with an age of 68 ± 7 Ma. However, we do not quote this as a best age since the steps disagree with the other three splits of the sample. It is possible that a portion of this split was not degassed as thoroughly as the others in the resetting event, or the trapped component was not completely accounted for. The summed age is 257.9 ± 22.5 Ma.

Figure 4.9. CH1 has a partial plateau of 19.8 ± 2.3 Ma, representing about 37% of the 39Ar (375-475 °C). The summed age is 259.5 ± 12.6 Ma. The trapped component is different than the other samples, but least-squares fitting suggests ~65 ± 25 is a solution. More conservative estimates result in ages ~ 10 Ma younger, which would bring the bottom of the u-shaped plateau to ~10 Ma rather than ~20± 2 Ma.

Figure 4.10. The average of the three splits that agreed (SB1, SB2, and CH1) with the standard deviation for the uncertainty results in 26.1 ± 11.0 Ma as the upper age estimate for the most recent resetting impact event.

The ages of the three splits from the melt rich MB020f, 5 are characterized by an overall classic diffusion pattern with higher ages at higher temperatures up to a partial plateau (Fig. 4.11). None of them have any apparent ages as low as the partial plateaus in the clast-rich splits. The high temperature steps have apparent ages as high as 3134 ± 110 Ma, which represents a lower limit to the age of a previous event. An argument could be made for a plateau in the apparent age spectra of CH2, with an age of 2740 ± 17 Ma (~57% of released argon, 10 steps, 975-1400°C). It does not meet the definition of a plateau because three steps are not within two-sigma agreement. However, when considering that all of the splits from this sample ~agree with this age, I believe that it is significant. The average of the partial plateaus results with an age of 2706.2 ± 28.9 Ma. The ages of each split are summarized in Table 4.3. The following discusses each split of melt rich

MB020f,5 separately.

SB4: The plateau plot of this split has a minimum age of 422 ± 30 Ma, followed by a stair-step pattern that ends with a series of similar ages ~ 2692 ± 24 Ma (5 steps, 975-1050°C, 20% of released 39Ar). The summed age is 1795 ± 16 Ma.

SB5: The same pattern is seen in this sample, which has a minimum age of 474 ± 175 Ma.

There is a series of similar ages ~ 2688 ± 40 Ma (7 steps, 1000-1200°C, 41% of the released 39Ar).

The summed age is 2203 ± 13 Ma.

CH2: The plateau plot of CH2 is similar to the other two splits. It has a minimum age of 531 ±

40 Ma, and a much broader partial plateau of high ages at 2740 ± 17 Ma (10 steps, 975-1400°C,

57% of the released 39Ar). The summed age is 2224 ± 22 Ma.

Figure 4.11. Apparent age spectra for all splits of melt-rich MB020f,5 (SB4, SB5, and CH2) show a diffusion pattern at low temperatures that lead to a partial plateau. The average age is 2706.7 ± 28.9 Ma and is the best age estimate for a resetting impact event.

Table 4.4. Chelyabinsk Argon-Argon Age Summary Sample Split ID Min Age Total Age Plateau Age # Steps % 39Ar Best age MB020F,2 26.1 ± 11.0 Clast Rich SB1 23.3 ± 11.7 247.1 ± 6.4 (28.9 ± 6.2) 3 40 SB2 12.3 ± 14.8 260.7 ± 11.8 (29.7 ± 9.7) 2 52 SB3 47.0 ± 38.7 257.9 ± 22.5 (68.0 ± 7.2) 4 40 CH1 16.5 ± 5.0 259.5 ± 12.6 (19.8 ± 2.3) 5 37 MB020F,5 2706.7 ± 28.9 Melt Rich SB4 422.4 ± 30.1 1794.7 ± 15.8 (2692 ± 24.1) 4 20 SB5 474.6 ± 174.7 2203.1 ± 12.8 (2688.7 ± 39.6) 7 41 CH2 531.4 ± 39.9 2223.9 ± 22.8 (2740.1 ± 17.3) 10 57 Partial plateaus are reported in parenthesis (####). "Min" stands for minimum. All ages are in Ma. Best age for MB2 is using only SB1, SB2, and CH1 splits; with the standard deviation representing the uncertainty.

4.4 Impact Age Discussion

This work shows evidence of a recent impact event in Chelyabinsk MB020f,2 at 26.1 ± 11 Ma, which is significantly younger than other known LL chondrites. Most LL chondrites have impact ages of 4200-4350 Ma (Swindle et al., 2014) with very few examples of resetting in the last 1500

Ma. The youngest 40Ar/39Ar ages are for the LL impact melts (LLIM) Yamato-790964 (1260 ± 24,

Takagami and Kaneoka 1987), NWA 1701 (970 ± 80, Swindle et al., 2006), and LAR 06298/9 with an 40Ar/39Ar age of 200 ± 200 (Weirich et al., 2009; Swindle et al., 2011) and the LL6

Morokweng fossil meteorite (625 ± 163, Jourdan et al., 2010).

The clast rich sample MB020f,2 yields younger apparent ages than the melt rich sample, indicating that it was more thoroughly degassed in the most recent event. Although this seems counterintuitive, this is not unusual behavior (McConville et al. 1988; Bogard et al. 1995).

Minerals in Chelyabinsk experienced shock altering effects that can increase the retentivity of argon relative to the clast rich sample, i.e. the melt-rich sample is more resistant against thermal resetting. Work by Trieloff et al. (2017) confirms lower ages in the clast-rich lithology

(indistinguishable from the results in this work, ~30 Ma) and explains the high retention of the melt material is due to formation of high pressure phases of plagioclase as a result of shock metamorphism. One might consider that the clast rich sample’s young apparent age is a result of heating during Chelyabinsk’s entry in the atmosphere and subsequent explosion. However, to drop the apparent age to 30 Ma from 4000 Ma would require that 99.8% of the argon to be degassed, and there is no evidence seen in the petrography for such extreme heating, which would be from the outside in. Therefore, I believe 26 ± 11 Ma from the clast rich sample does in fact represent an upper limit to the most recent Ar-Ar resetting impact. 88

The melt rich MB020f,5 sample shows minimum ages of ~450 Ma and has a partial plateau age of 2706 ± 29 Ma. This partial plateau age is consistent with Sm-Nd ages (Righter et al., 2015) of

Chelyabinsk and is interpreted to represent a unique impact event not seen in other LL chondrites.

Trieloff et al. (2017) studied dark lithologies and melt from Chelyabinsk, with different results

(~1.7 Ga compared to ~2.7 Ga in this work). The shape of their age spectra is very reminiscent of the age spectra of MB020f,5. The difference might be attributed to different air corrections at high temperature steps; Trieloff et al. (2018) determined a trapped 40Ar/36Ar = 1870 ± 200 correction from a reverse isochron while this work finds a trapped 40Ar/36Ar = 100 ± 50 determined from two separate standard isochrons, and a higher trapped component of 550 ± 180 is evident in both the standard and reverse isochron of an additional aliquot. The only evidence of such high trapped corrections reported in Trieloff et al. (2018) is in the intermediate diffusion steps, which do not hold impact age data. Using a trapped correction of 1870 on my work results in maximum ages similar to Trieloff et al. (2018), but the shape of the plateau is much more disturbed. Some ages from high temperature steps have an age of ~0 when using the correction from Trieloff et al., implying an over correction. Applying the trapped correction from this work to their data results in a plateau age of ~2690 Ma, which is very similar to the quoted age of MB020f,5 (2706 ± 29

Ma). This confirms that the different ages are the result of different trapped corrections.

Furthermore, despite the different trapped corrections found in my work, the isochron ages agree

(determined from the slope of the isochrons). In summary, using their trapped correction results in an over correction in MB020f,5, while the measured trapped correction from my work applied to the sample from Trieloff et al., 2017 results in ages that agree with MB0202f,5. This should be further investigated with additional analysis; perhaps by using knowledge from the prior analysis to better define temperature steps that can more clearly determine a trapped correction. 89

The other dating methods applied to Chelyabinsk show many conflicting behaviors. Bouvier

(2013) studied a dark melt fragment from Chelyabinsk and obtained a Pb-Pb age of 4538.3 ± 2.1, while U-Pb ages of phosphates are around 4456 ± 18 Ma (Lapen et al., 2014). These ages are comparable to one another yet are much older than both the Ar-Ar and some Sm-Nd ages, which at first seems puzzling – since Pb-Pb generally resets more easily than Sm-Nd, and not as easily as Ar-Ar, we would expect the age of Sm-Nd > Pb-Pb > Ar-Ar. An impact with enough energy to reset the Sm-Nd chronometers would likely reset many other chronometers. However,

Chelyabinsk is shocked (S4, Badyokov et al., 2015) which affects different isotopic systems in distinct minerals in unknown ways. The Sm-Nd, Rb-Sr, and U-Pb are all disturbed from shock affects and reported ages are difficult to interpret. It is not well understood how energetic impacts

(that could completely degas argon) might affect more refractory dating systems. Experimental studies of shock and heat effects on Rb-Sr, Sm-Nd, Pb-Pb, and U-Pb isochrons (Gaffney et al.,

2011) have been performed on a lunar basalt (described as ~60% clinopyroxene and ~30% plagioclase). A control, shocked, and heated isochron were measured for each isotopic system to monitor the effects of a ~50 GPa shock and heating of 1000 °C. The Sm-Nd system was disturbed the least, followed by Rb-Sr, with Pb-Pb and U-Pb being the most susceptible to change. The authors note that shock or heating can degrade or destroy an isochron fit, but if a linear isochron remains, the age it determines likely reflects the crystallization age of the sample (based on experiments on their control sample). This provides a guide when considering Sm-Nd, Rb-Sr, and

U-Pb ages, but likely does not reflect these metamorphic effects on Chelyabinsk in detail. What are the isochron effects of different material that appears to have been heated to different temperatures and impacted multiple times? Chelyabinsk is thought to have experienced a shock of 90

~30-40 GPa and temperatures of up to ~1600 °C (Righter et al., 2015), which are different parameters than covered in the experiment.

Figure 4.12 includes a compilation of measurements using different chronometers (similar to

Righter et al., 2015), and implies that Chelyabinsk has experienced multiple impacts and has a complicated history. If each of these ages represent an impact event (at least eight implied by

Righter et al., 2015), then Chelyabinsk is indeed a heavily impacted and complex sample!

However, not all of the age signatures may represent a meaningful event, or at least individual events. For example, the K-Ar ages included in Figure 4.12 are essentially an average of different events recorded in different minerals in the sample. Two Ar-Ar ages in the figure are from mixed lithologies (clast rich and melt rich) which would result in various mixtures of argon gas representing two different events that are evident in this work. Therefore, they should not be considered as impact events.

Three studies of Sm-Nd yield four very different ages; ~300, ~2900, ~3700, ~4452 Ma

(Galimov et al., 2013; Righter et al., 2015; Bogomolov et al., 2013; and Righter et al., 2015 respectively). The Sm-Nd age ~300 Ma comes from unidentified rock separates, whole rock measurements, glass, and is highly disturbed. If real, this would imply a high energy impact that would likely reset at least the Ar-Ar ages if not others. A Sm-Nd age ~2900 Ma is indistinguishable from the Ar-Ar age of MB0202f,5 in this work. This Sm-Nd age is determined from removal

(whole rock, magnetic separate) and addition (glass, whole rock, and non-magnetic separate) of measurements by Galimov et al., 2013 with one dark impact melt breccia measurement from

Righter et al., 2015. Though the justification for addition and removal of measurements is questionable, it is possible that this represents an energetic impact. The ~3700 Ma age is determined by three mineral separates that have had the bulk rock measurement removed 91

(Bogomolov et al., 2013). The authors of this work note that if the bulk rock was included, the isochron represents an age ~700 Ma younger, which would be indistinguishable from the age reported in Righter et al., (2015) and may represent the same event. The oldest Sm-Nd age of

~4452 Ma comes from only two whole rock (light lithology and light lithology that has been darkened by shock) measurements. The ~300 Ma and ~4452 Ma ages are not considered to represent impacts due to disturbed isochrons and/or not enough measurements.

The Rb-Sr data for Chelyabinsk are also disturbed and difficult to interpret. Nakamura et al.

(2015) determined two Rb-Sr ages, a younger ‘re-processed’ age of 153 ± 58 Ma from glass and an ‘ancient’ age of 4567 Ma from areas of light lithology darkened by shock. The younger age

(153 ± 58 Ma) could represent the same impact event recorded in the Ar-Ar of MB020f,2 at ~30

Ma. Both ages are based on well-developed isochrons, and therefore the shock effects, though unknown, are assumed to be minimal (Gaffney et al., 2011). Combined Rb-Sr literature data in

Righter et al. (2015) show an isochron age of ~300 Ma that is not distinguishable from an isochron age of ~30 Ma (Ar-Ar age of this work). In total they find three ages at ~300, 880, 1400 Ma

(Righter et al., 2015) that are not reliable because this data is highly disturbed; any fit seems as likely as the other resulting in very different ages.

It is difficult to explain how Chelyabinsk could record four distinct Sm-Nd and four distinct

Rb-Sr ages that are real. The disturbed isochron plots for Sm-Nd and Rb-Sr are likely the result of shock effects that make it difficult to determine a reliable age. However, with additional comparison to other chronometers, some of the Sm-Nd and Rb-Sr ages may have value.

The Pb-Pb age of 4457 ± 35 Ma (Lapen et al., 2014) agrees well with the U-Pb ages of apatite grains that have upper concordia ages of ~4450 Ma (4454 ± 67 Ma, 4452 ± 21 Ma, 4433 ± 110 Ma from Lapen et al., 2014; Popova et al., 2013; and Kamioka et al., 2014 respectively). Apatite has 92 a less constrained lower concordia age of 585 ± 390 (Lapen et al., 2014), which might indicate the most recent time lead was lost from the system. Apatite isotopic composition ends re-equilibration with its surroundings at ~ 500°C, i.e. that is its closure temperature (Cherniak et al., 1991). Given the ages listed above, this would suggest an early post-accretion impact reset the apatite grains in

Chelyabinsk. Contrary to this, U-Pb ages from zircons are intermediate to the upper and lower concordia of apatite grains (~2850 ± 15 Ma and ~ 840 ± 10 Ma, Skublov et al., 2015). Zircons have a much higher closure temperature (> 1000°C; Cherniak et al., 1991) than apatite, making it unclear why the apatite grains were not also reset. Shock and heat experiments (Gaffney et al.,

2011) show that Pb-Pb and U-Pb are generally affected most strongly, however it is not clear how the behavior of these systems change based on the target mineral. It seems reasonable that the upper concordia of zircon should be older than the upper concordia of apatite based on closure temperatures, but apatite is ~2 Ga older! Some possible explanations include a local difference recorded by the respective samples, shock effects, and unknown experimental error. However, recrystallization of zircon would require a non-local event and subtle changes to the isochron as suggested in Gaffney et al. (2011) do not explain the respective age difference. Perhaps there was an experimental error which would most likely come from the zircon U-Pb ages; data come from one study while the apatite ages have been independently verified.

Considering the different isotopic dating methods and their context described above, a more reliable distribution of impact ages on Chelyabinsk is given in Figure 4.13 and Table 4.4 and is interpreted as follows: formation age ~ 4560 Ma (Re-Os, Rb-Sr), an early energetic impact ~ 4450

Ma (Pb-Pb, U-Pb), an energetic impact not seen in other LL chondrites at ~ 2800 Ma (Sm-Nd, Ar-

Ar), followed by the most recent impact event ~30 Ma (Rb-Sr, U-He, U-Pb, Ar-Ar). A less conservative approach (including the U-Pb of zircon and some of the poorly constrained ages from 93

Rb-Sr and Sm-Nd systems) results in an age distribution very similar to Figure 4.13, with the addition of weak evidence of an impact event ~800 Ma. A graphical summary for the impact history is given in Figure 4.14.

0.10 Chelyabinsk Impact Age Distribution 0.09

0.08 y

t 0.07

i l

i 0.06

b a

b 0.05

r p

0.04

e v

i 0.03

t a

l 0.02

e R 0.01

0.00 0 1000 2000 3000 4000 Age (Ma)

LL Chondrite Age Distribution

l i

b o

e v

a l

e R

0 1000 2000 3000 4000 Ar-Ar Age (Ma)

Figure 4.12. A) (Top) Combined chronology studies of Chelyabinsk show evidence for up 8 possible impact/thermal events from 22 analysis using multiple dating methods including Ar-Ar (Beard et al., 2014, Lindsay et al., 2015, Righter et al., 2015), U-He (Nyquist via Righter et al., 2015), Sm-Nd (Galimov et al., 2013, Righter et al., 2015), Rb-Sr (Galimov et al., 2013), U-Pb (Popova et al., 2013; Lapen et al., 2014), and Pb-Pb (Bouvier 2013). Ages circled in red were contributed from this work. Compare this figure to that in Righter et al. 2015. B) (Bottom) LL ordinary chondrite age distribution of 16 samples for comparison to event signatures of Chelyabinsk in A above, data from Swindle et al., 2014.

Figure 4.13. Revised impact age distribution suggested by this work. Formation age ~ 4560 Ma (Re-Os, Rb-Sr), an early energetic impact ~ 4450 Ma (Pb-Pb, U-Pb), an energetic impact not seen in other LL chondrites at ~ 2800 Ma (Sm-Nd, Ar-Ar-this work), followed by the most recent impact event ~30 Ma (Rb-Sr, U-He, U-Pb, Ar-Ar-this work). See Table 4.5 for references.

~4560 Ma ~4450 Ma ~2700 Ma ~ 30 Ma

• Formation age • Energetic Early impact • Energetic impact • Moderate impact • Re-Os • U-Pb: Apatite • Sm-Nd: whole rock • Rb-Sr: Melt rich • 4558 +/-12 Ma • 4454 +/- 67 Ma • 3700+/-110, or glass and • Rb-Sr: clast rich • 4453 +/- 21 Ma 3000 +/- 200 plagioclase • 4567 Ma • 4433 +/- 110 Ma • Sm-Nd: light and dark • 150 +/- 60 Ma • Pb-Pb: Apatite mix • U-Pb: 580 +/- 350 Ma • 4457 +/- 6 Ma • 2900 +/- 500 • Most recent lead • Recrystallized Apatite • Ar-Ar: 2706 +/- 30 Ma • Reset argon system loss • U-He: ~30 Ma IMPACT HISTORY OF CHELYABINSK • Ar-Ar: ~30 Ma

Figure 4.14. History of Chelyabisnk. See Table 4.5 for references.

These results provide additional data that is important for understanding the history of

Chelyabinsk and the evolution of the LL chondrites, which may include a shared past with the

Hayabusa target asteroid 25143 Itokawa, identified as an LL chondrite (Nakamura et al. 2011).

Itokawa and Chelyabinsk both have 21Ne CRE ages of ~1.5 Ma (Nakamura et al., 2011; Meier et al., 2014). The only other known LL chondrite with a matching exposure age is Appley Bridge

(Haymann et al., 1967), so all may derive from the same event (Meier et al., 2014).

Figure 4.15. Ar-Ar apparent age of Itokawa particle #0013 measured by Jourdan et al., 2017. The summed age is 2291 ± 139, which is similar to the summed age of MB,5 (though the max ages of MB,5 are ~500 Ma higher).

Furthermore, Jourdan et al., (2017) recently completed two Ar-Ar measurements of Itokawa; one from an un-shocked grain that appeared to have an ‘old’ age (larger errors), and the other from

Table 4.5. Age Summary of Chelyabinsk. Isotope Age (Ma) Reason for exclusion from impact distribution or suspicion of inclusion U-He ~271 Ar-Ar 26 ± 11 ~ 30 ± 303 312 ± 61,2 Measured from mixed phases (light and dark) 716 ± 301 Measured from mixed phases (light and dark) 1014 ± 241,2 Partial resetting of ~30 Ma event? 1184 ± 401,2 Measured from mixed phases (light and dark?) 1700 ± 1003 Represents the impact measured in this work with different correction? 2706 ± 29 K-Ar 865 ± 971 K-Ar age not appropriate for measuring multiple events 1000 4 K-Ar age not appropriate for measuring multiple events 1945 ± 1681 K-Ar age not appropriate for measuring multiple events 1952 ± 1691 K-Ar age not appropriate for measuring multiple events 2736 ±1991 K-Ar age not appropriate for measuring multiple events* Rb-Sr 153 ± 585 Shock effects?* 880 ± 1201 Two data points. Shock effects? 1400 ± 3001 No other indication of an event, difficult to explain. Age from two data points. Shock? 4567 5 Higher than Sm-Nd, U-Pb, Re-Os. Shock effects?* Sm-Nd ~2906 Highly disturbed. Shock effects? 2900 ± 5001 Disturbed. Shock effects? * 3733 ± 1107 No other indication of an event, Inclusion of whole rock measurement reduces age to ~3000 Ma.* 4452 1 Only two data points. Why not reset by 2900 Ma event?* U-Pb 585 ± 3908 834 ± 79 Zircon closure age after apatite? 2744 ± 139 Zircon closure age after apatite?* 2861 ± 159 Zircon closure age after apatite?* 4433 ± 11010 Why not affected by event seen in zircon? 4452 ± 2111 Why not affected by event seen in zircon? 4454 ± 678 Why not affected by event seen in zircon? Pb-Pb 4538 12 Measured on a mixture of phases, could increase shock effects? 4457 ± 68 Re-Os 4558 13 *Indicates inclusion/exclusion in age distribution does not change results. Blue text are results from this work, bold descriptions are for unreliable measurements not included in Fig. 4.13, and italicized descriptions are semi-reliable in Fig. 4.13. 1. Righter et al. (2015), 2. Lindsay et al. (2015), 3. Trieloff et al. (2015), 4. Haba et al. (2014), 5. Nakamura et al. (2015), 6. Galimov et al. (2013), 7. Bogomolov et al. (2015), 8. Lapen et al. (2014), 9. Skublov et al. (2015), 10. Kamioka et al. (2014), 11. Popova et al. (2013), 12. Day et al. (2014), 13. Bouvier (2013).

97 a highly shocked plagioclase grain that yielded an age of 2291 ± 139 Ma (Fig. 4.15). Interestingly, this age is similar (though not within error) to the Ar-Ar age in Chelyabinsk MB020f,5 (~2745 ±

34). Although the Itokawa age is lower, the spectrum is disturbed, and >50% of the released argon is from a single step, these ages are comparable. Itokawa has a similar low-inclination orbit as

Chelyabinsk (Popova et al., 2013) although the aphelion of Itokawa is between Mars and Earth while the aphelion for Chelyabinsk is in the asteroid belt. CRE ages are not typically an absolute chronometer, but we know the time that Chelyabinsk was closed to galactic cosmic-rays, so in this case it is an absolute chronometer; i.e. it was ejected from its parent body ~1.5 million years ago.

The youngest Ar-Ar age from this study places an upper limit of the most recent impact at < 30

Ma, which could be the result of an impact that is at ~1.5 Ma, though I believe these are separate events. For comparison, L4 and H5 chondritic material recovered from the fall of Almahata Sitta have Ar-Ar ages of > 4.2 Ga (Beard et al., 2013) compared to CRE ages of ~20 Ma (Meier et al.,

2012), i.e. the impact that began cosmic-ray exposure of Almahata Sitta is not recorded in its Ar-

Ar data.

The effect of partial resetting on the plateau of MB,5 has been modeled (Swindle and Weirich,

2017). The result, shown in Fig. 4.16, assumes complete degassing occurred at 2715 Ma and subsequently lost ~5% of its argon. Although not an exact match, the model resembles the data and shows that the plateau age is representative of the true age of crystallization (2688 Ma plateau in model is ~1% lower than 2715 Ma event).

4.5 Conclusions

The Ar-Ar analysis of melt rich and clast rich lithologies indicate Chelyabinsk experienced two impact events. The older event represents an impact that was energetic enough to degas and completely reset argon occurred ~2700 Ma ago. This provides a lower limit for the time of shock 98 alteration of minerals, including formation of high pressure phases of argon-retentive-plagioclase

(jadeite), metal and sulphide melt veins along grain boundaries, melt-filling of shock-altered silicates pore space (shock-darkening), and melt pockets. The Ar-Ar age of Itokawa has a similar age to the melt rich material and may indicate a shared impact on the LL parent body that was experienced by Chelyabinsk and Itokawa that is unique among LL chondrite ages. A younger impact event at ~30 Ma significantly reset the argon in the clast rich material, while only partially resetting the melt rich material and leaving the high retentive phases undisturbed.

The results of this work are the youngest and oldest Ar-Ar ages of Chelyabinsk and have provided additional constraints on possible impact ages determined by other isotopic systems that otherwise would remain ambiguous. Chelyabinsk does not have a history as complicated as others have suggested, though the isotopic data are complex. Chelyabinsk shows evidence of ~2-3 impacts before ejection from its parent body.

Figure 4.16. Comparison of the apparent age of Chelyabinsk sample CH2 with a model (Swindle et al., 2017), that demonstrates partial resetting of a sample that was completely degassed/reset at 2715 Ma. The model assumes that argon was released from two mineral phases (first 35% of gas released under E = 20 kcal/mol conditions, and remaining 65% released under E = 70 kcal/mol).

Chapter 5: Ar-Ar Difficulties with Brachinites

5.1 Introduction

The intent of this work was to combine Ar-Ar and CRE data of the brachinites (including brachinite-like and similar ungrouped achondrites) to explore the history and relationships of these meteorites (by finding common chronologic signatures, or the lack thereof). However, the Ar-Ar portion of this work was fraught with difficulties and ultimately unsuccessful.

An overview of the general Ar-Ar methods for this work can be found in Chapter 2, with specific and additional details for the experimental design and brachinite sample analysis provided here. The main difference is that a diode laser was used for heating instead of the furnace system used for Chelyabinsk. The laser system provides better signal to noise by reducing the volume of the extraction system and allows for direct viewing of sample heating. A drawback of the laser is that the actual temperature of the sample is not well constrained. Because it is possible to view the sample and observe melting, and therefore degassing, temperature uncertainty did not seem to be a problem beforehand. These results are not meaningful, but they are included to discuss some of the difficulties experienced with this experiment so that it may serve to help others know what to anticipate and plan for in their future work.

5.2 Analysis

Brachinite, brachinite-like, and ungrouped achondrite samples were divided into three aliquots

(splits) for each sample. The splits were irradiated at the Cadmium-Lined In-Core Irradiation Tube

(CLICIT) facility at Oregon State University along with proper irradiation monitors to determine the J-factor; roughly 2.8×10-2 among the samples (propagated from multiple standards at known positions, with uncertainty ~1×10-4 ). The achondrites were irradiated for 100 hours in the linear- with-height portion of the reactor to provide a more evenly distributed neutron flux. Upon return 100 from the reactor, samples were then allowed to cool for about 2 months to allow for short-lived isotopes to decay before being placed in cylindrically drilled holes in a copper disk (Fig. 5.1).

Copper was chosen because it does not couple with the laser.

Figure 5.1 (Below): The sample holder for the laser is a copper disk with molybdenum foil placed in the holes.

Figure 5.2 (Right): The sample holder was placed in a viewport that has a window that is transparent to the laser, allowing for direct heating and viewing of the samples.

Different hole configurations were produced to allow for multiple sample sizes. Sample splits and standards were placed in separate holes and placed into a viewport window piece that is attached to the mass spectrometer extraction line (Fig. 5.2). After the copper disk with the samples is placed in the viewport, it is clamped together with a copper gasket, providing a vacuum seal.

The viewport window is made from quartz, which is transparent to the laser that is focused to a ~1 mm beam width at the sample.

Argon was extracted from each sample using a computer-controlled step-heating procedure.

Each sample holder was uniquely mapped with the locations of each sample and standard. The laser was placed on a motorized stage that was programed with the map data. This allowed for an automated sample analysis procedure to be executed for multiple site analysis. To minimize uneven heating of samples, an additional motion program function called ‘jog’ was used to move the laser position over a sample during heating. Typical temperature steps were intended to be 101 from ~ 350-1400 °C in ~50-100 °C intervals depending on the amount of argon released and the accuracy desired. However, instrumental malfunctions prevented extraction steps from reaching their target temperature, sometimes only reaching ~1100 °C. The actual temperatures reached have a high uncertainty, and it is assumed that not all of the samples analyzed were degassed completely.

An addition to these complications, the samples analyzed do not have identifiable feldspar and have low potassium abundances ( e.g. 6.2 and 239 ppm from two measurements of NWA 1500,

Goodrich et al., 2006), which makes it even more difficult to distinguish if the sample is not being heated enough (the more likely scenario) or if it has already been degassed. As detailed in chapters

2 and 3, uncertainties and corrections were applied to account for blanks, machine discrimination, spallation-produced isotopes, and interfering isotopes produced in the reactor from Ca and K.

Isotopic measurements were regressed to a time of zero using linear regression techniques, after subtracting baseline values in order to get the most accurate measurements. The total number of heating steps varied between 15-25 per split. The power supplied to the laser was roughly calibrated to an optical pyrometer for an estimate of the target temperature, which is highly suspect.

More details can be found in the discussion section.

5.3 Brachinite Ar-Ar Results

The apparent ages (Figs. 5.3-5.6)) for the brachinites and achondrites are complicated, confusing, and may not represent a time when a meaningful event happened. Take these ‘results’ with a bag of salt. The quoted temperatures in this section may not actually represent the temperature that was achieved, but nonetheless serve as a description of which measurement step is being discussed. 102

Two measurements were taken for brachinite-like achondrite NWA 6077 (Fig. 5.3). The first split (BS1) has a minimum age of ~200 Ma in the first two steps (~50% of the released argon,

“350-375 °C”) which appears again at higher temperatures (~8% of the released argon, “650-

700°C”). This second minimum could be the result of recoil. There is a series of steps that appear to be better constrained at ~650 Ma (20% of the total argon, “400-550 °C”). The high temperatures steps indicate an age of ~1000 Ma with larger uncertainties. Note that the majority of the gas was released by “375 °C”, which is lower than expected (see discussion). The summed age of this disturbed sample is 395 ± 26 Ma. The best estimated trapped composition from the isochrons indicates a correction of ~20 ± 10, though they are difficult to interpret.

The second split of NWA 6077 (BS2) appears to have a U-shaped release pattern and has a minimum age at ~125 Ma, with a series of low ages at 151 ± 10 Ma (28% of the total argon, from

“725-900 °C”). The summed age over all steps is 140 ± 150 Ma (note the high uncertainty).

Excluding the steps that have large error, the summed age is 312 ± 60 Ma (80% of the total released argon, “425-1100 °C”) compared to ~395 for the other split of the same sample.

NWA 595 is also a brachinite-like achondrite but the age spectra (Fig. 5.4) are not in agreement with each other. The initial 70% of the 39Ar released from BS4 has an age of ~2800 ± 140 Ma

(from the “500-800 °C” steps), and a maximum age of ~3600 Ma (from “1400-1450 °C”). A series of high temperature steps give a minimum age of ~1200 Ma (10% of released argon, “1100-1250

°C”), but may be the result of recoil. The summed age over all steps is 2781 ± 128 Ma.

Contrary to BS4, the split BS5 shows much younger ages with 87% of the argon released from

425-900 indicating an age of 1562 ± 44 Ma, and an absolute minimum age around ~1000 Ma. The summed age over all steps is 1986 ± 60 Ma. 103

Brachinite NWA 7297 also shows conflicting release patterns (Fig. 5.5). Split BS7 has very high ages greater than 5000 Ma. BS8 has five steps that agree to ~ 2894 ± 229 Ma (22% of the released argon, “365-475 °C”), and ten steps that agree at higher temperatures with an age of 814

± 109 Ma (53% of the argon, from “650-1150 °C”). Note that this sample’s maximum heating was only ~ “1187°C” so it may not be degassed completely. The summed age is 750 ± 1000 Ma.

Brachinite-like NWA 1500 splits show two different patterns. Split BS10 has unrealistically high ages, while BS11 has a U-shaped release pattern (Fig. 5.6). The minimum age is ~816 ± 99

Ma (50% of the released 39Ar, from “800-900 °C”) and the summed age is ~3221 ± 87 Ma.

Figure 5.3. Apparent ages of NWA 6077. BS1 (blue) has two sets of minimum ages at ~200 Ma, a series of intermediate steps at ~650 Ma, and the maximum temperatures are < 2000 Ma. BS2 (orange) has a u- shaped pattern with minimum ages ~150 Ma. The summed age of BS1 is 395 ± 26 Ma, and the summed age of BS2 is 312 ± 60 Ma (from 80% of the argon released, after taking out high error steps).

104

Figure 5.4. Apparent ages of NWA 595. BS4 (gold) has an age of 2800 ± 140 from the first 70% of the released argon. BS5 (purple) has lower ages, 1562 ± 44 Ma from 87% of the released argon. The summed age of BS4 is 2781 ± 128 Ma, and the summed age of BS5 is 1986 ± 60 Ma.

Figure 5.5. Apparent ages of NWA 7297. BS8 (maroon) shows two sets of ages, the first has an age of 2894 ± 229 Ma, and the second has an age of 814 ± 109 Ma. The summed age of BS4 is 750 ± 1000 Ma.

105

Figure 5.6. Apparent ages of NWA 1500. BS10 (purple) has unrealistically high ages, with a minimum around 4200 Ma. BS11 is much different and has a u-shaped spectrum. The minimum age is 816 ± 99 Ma. The summed age of BS4 is 3221 ± 87 Ma.

5.4 Ar-Ar Discussion

NWA 6077’s splits both show signs of being relatively young <~1000 Ma. The Ca/K ratios of

NWA 595 were examined to help understand the difference between the splits, but even when the

Ca/K between the aliquots agree, the ages do not. Sample splits often do not agree and are generally quite puzzling. These poor results are attributed with low potassium content of the samples and difficulty in properly heating the samples. Below is a review of the efforts taken to increase heating efficiency.

As described in the methods, samples were placed in holes drilled into a copper disk. To create a surface that the laser could couple with, molybdenum disks were formed from foil and placed at the base of the copper holes (this heats the copper conductively for blank measurements).

Molybdenum disks were ultrasonically cleaned in distilled water followed by acetone before being 106 placed at the base of each hole. Early experiments also had difficulty with ‘pulsing’ of the laser, where the laser would go from 0 to 100% power in quick succession. This was eventually solved with multiple ‘auto-tuning’ of the WATLOW controller PID parameters.

Figure 5.7. A) (Left) Sample holder set up for laser experiments, with a sapphire disk on top of a glass ring. B) (Above) Cracked sapphire disk from laser heating.

During ‘high’ temperature extractions (beginning as low as 800 °C and as high as 1150 °C), sample debris began to deposit on the viewport window. When this happened, the extraction of the sample at that location had to be stopped until the viewport could be cleaned. Although the viewport is transparent to the laser, once debris coats the window it can couple with the laser. This heats the viewport and can cause it to crack, which presumably would happen while the laser is heating a sample and therefore sample gas would be lost. To avoid this, the laser was moved to another sample to begin lower temperature extractions, or the viewport was removed. The process of cleaning the viewport requires isolating the extraction line of the mass spec and breaking the vacuum to remove the viewport holder. At this point it can be carefully disassembled as to not contaminate or spill the samples, cleaned with acetone, and reassembled. During reassembly, a sapphire disk was placed on top of the copper sample holder to act as a buffer between sputtered sample material and the viewport. The sapphire is ~95% transparent to the laser but debris from 107 standards and sample built up on and cracked the sapphire before building up on the viewport once again.

In an effort to prevent this from happening again, various modifications continued to be made without much success. Molybdenum disks were included above (in addition to the disk that was already included below) the sample with the idea to create an ‘oven’ around the sample to increase heating efficiency while also adding additional protection to the sapphire from sputtering. A glass ring was placed on top of the copper holder to elevate the sapphire disk ~1.5 cm above the sample to allow debris to sputter diffusely rather than concentrated directly onto the sapphire (Fig. 5.7).

Finally, two sapphire disks were placed on top of the glass ring so if one failed, there would still be one layer of protection before material could build up on the viewport.

It would have been tremendously useful to verify how the temperature corresponds to the laser power as it would allow better comparisons of different splits and to better understand results.

This was attempted by melting different metals but the results cannot be trusted because they were in contact with other metals which changes the thermal properties. In the future, it might be useful to use a ceramic sample holder, or at least a ceramic piece to melt metals on to avoid this

Figure 5.8. A) (Left) Modified sample wrap, designed to allow for better heating with less sputtering debris. Two ‘pie tins’ of different sizes are formed from a tantalum-iridium alloy. The smaller pie tin is placed on a thin strip of the alloy, the large pie tin is set face down on top of the smaller pie tin, and the ends of the strip are pulled up over the top and twisted tight (Right).

108 problem. Another possibility is to attach a thermocouple to a piece of molybdenum and measure the temperature/power levels.

Discussions with other labs led us to redesign our procedure. Much of the redesign was led by the lab manager, Dr. Clark Isachsen. Many labs use platinum tubes to contain their samples by crimping the ends of the tube. This couples well with the laser and essentially acts as an oven baking a sample. This was not possible for this work because our samples are much larger rock chips that would make the cost of enough platinum too high. Instead we found that a tantalum- iridium alloy has similar properties to platinum, is much cheaper, and could be purchased in foil, which would make sample preparation much easier. The geometry of the hole also seemed to be problematic. Because the holes are relatively deep, sample particles released during heating were collimated upward directly onto the sapphire disk. Some labs use a shallower and wider pit or a trough that runs the length of the sample holder to allow particles to travel more diffusely. For preliminary testing, the existing copper holders were modified (Fig. 5.9) to affectively elevate the laser target (by elegantly filling the holes with stainless steel nuts). A tantalum-iridium foil was purchased and a tin punch was used to create ‘pie tins’ of two sizes. The smaller pie tin was placed on a narrow strip of tantalum-iridium and filled with a sample for testing. The larger pie tin was placed face down on top of the smaller tin, and the two ends of the strip of tantalum-iridium were folded over the top and tightly twisted. The twisting of the strip affectively clamps the two pie tins together and is surprisingly secure once achieved (suitable for whole rock pieces, approximately larger than sand). Due to laboratory malfunctions, these new design ideas were not able to be tested further. 109

Figure 5.9. The modified sample wrap placed on the copper sample holder. The pie tin is set in a shallower hole than used for the experiments. This design avoids the effects of the deep pit and covers the sample more thoroughly to prevent debris building up on the sapphire window.

5.5 Brachinite Argon Conclusion

The Ar-Ar results are poor and difficult to interpret stemming from high sensitivity to trapped corrections. This sensitivity is a consequence of having non-detectable amounts of potassium, which combined with laser heating problems makes it difficult to report results with any confidence. Doing standard and logical corrections yields many uninterpretable results. The remaining brachinite and brachinite-like achondrites not yet analyzed (NWA 3151, NWA 4874,

NWA 4876, NWA 4882, NWA 4969, NWA 6474, NWA 6962, NWA 7297, NWA 8077, and

NWA 10637) contain more potassium, have been irradiated, and are being prepared for analysis by Dr. Barbara Cohen at NASA Goddard.

110

Chapter 6: Cosmic-Ray Exposure Ages of Ureilites

Part of the planned analysis of the brachinite/brachinite-like achondrites was to compare their

CRE ages with Ar-Ar ages and oxygen isotopes. To ensure that I had the capability of independently measuring their ages, a test case was done using published noble gas data (e.g.

Schultz and Franke, 2004). The calculated ages from this work agree with literature values, with the addition of comparisons of the forsterite content and oxygen isotopes. The ureilites and brachinte/brachinite-like achondrites both have distinct sample groups making this an ideal proof of concept (ureilites can be thought of as two sample types- monomict and polymict, which is broadly analogous to two sample types- brachinite and brachinite-like achondrites). Shielding parameters for the brachinite/brachinite-like achondrites were determined using the methods described in Chapter 2. The shielding parameters for the ureilites are based on different methods that are described in the methods (6.2).

6.1 Introduction

(Goodrich, 1992; Mittlefehldt et al., 1998). Igneous properties include bulk element chemistry, mineral chemistry, and petrography suggesting that they are products of a high degree of igneous fractionation on a differentiated body (Goodrich, 1992). Primitive characteristics of ureilites include their variable oxygen isotopes (see Fig. 1.12; Clayton and Mayeda, 1988; Greenwood et al., 2017), high trace siderophile element abundances, and trapped noble gases of chondritic abundance with a fractionated pattern (Mittlefehldt et al., 1998; Mittlefehldt 2004) that are unlikely 111 to survive significant geologic processing. Therefore, it is hard to make a clear distinction as to whether ureilites represent primitive or differentiated materials, and if they can be linked to a common heterogeneous source (inferred from variable oxygen isotopes).

The collection of ureilites is dominated by the main group (or monomict) ureilites that many believe to be mantle residues. Petrologic models predict that complementary melt rocks, which are not present in the meteorite collection (referred to as the missing basalts), exist (Goodrich et al.

1987). A subclass of the ureilites consists of polymict, fragmental breccias (e.g., Berkley et al.,

1980; Goodrich et al., 2004; Downes et al., 2008) that are inherently heterogeneous on the scale of thin sections (Rai et al., 2003). These polymict ureilites contain materials spanning the same range of magnesium to iron ratios in olivine (given as Fo, defined as 100*molar Mg/(Mg+Fe)) as monomict ureilites, suggesting both classes derive from the same source (Goodrich et al., 2004;

Downes et al., 2008). Polymict breccias are thought to be the regolith material formed on or near the surface of the original UPB and contain a few percent feldspathic clasts that are considered to be representative of the missing basalt (Warren and Kellemeyn, 1989; Cohen et al., 2004; Goodrich et al., 2004; Downes et al., 2004).

Goodrich et al., (2004) suggest that delivery of ureilites to Earth is not from impacts on the

UPB, but instead from offspring or daughter bodies (new bodies that formed after breakup of the original parent body). They argue that a common pressure and temperature history evident in monomict ureilites (Fo peak around ~79) suggests that they derived from a common body, which later experienced very rapid cooling with a sudden drop in pressure. This could be a result of deep excavation from a major impact. The reassembled material could form several offspring bodies that are described as reassembled debris of ~ meter sized monomict materials with a range of Fo’s that have a prevalence of Fo ~79 (Fig. 1.11). One or more of these bodies could be the source of 112 ureilite meteorites. More recent work by Goodrich et al. (2015) suggests a common history that links all ureilites to a single Ureilite Daughter Body (UDB) that formed after the initial UPB breakup (rather than multiple daughter sources). They suggest either a single breakup of the UDB at > 46 Ma with subsequent collisions producing smaller fragments, or a cratering event at 46 Ma followed by UDB break up at ~20 Ma based on cosmic-ray exposure ages of Almahat Sitta (Fig.

1.13).

Alternatively, the ejected material could have reaccreted back onto the UPB as a rubble pile

(Warren and Kallemeyn, 1989). In these models, either for a rubble pile on the UPB or the creation of offspring bodies, polymict ureilites represent the regolith (Fig. 6.1) that overlies the main group ureilite material and is developed after reassembly (Goodrich et al., 2004).

The high degree of heterogeneity and foreign clast content of the polymict ureilite Almahata

Sitta (Horstmann and Bischoff 2014) is explained to be the outer regolith material of the source

UDB (Goodrich et al., 2015). This regolith material is from the same body as the polymict ureilites

(deep regolith of UDB) and the main group ureilites (interior of UDB), but has foreign clasts

(enstatite, ordinary, and Rumuruti-type chondrites) implanted at low velocities.

If it is assumed that the ureilites come from a single daughter body, an impact could break apart material from different heterogeneous or homogeneous regions (Warren and Kallemeyn, 1989;

Goodrich et al., 2002), which might be evident in their CRE data. A goal of this work is to examine

CRE ages for evidence of clusters that may improve our understanding of the origin of ureilites.

Possible relationships of CRE ages with other parameters will be examined and implication on how they might relate to the structure (Fig. 6.1) and heterogeneity of the proximate ureilite source body will be discussed. This will be done by comparing oxygen isotopic ratios (Δ17O) and the ratios of magnesium to iron in olivine (Fo) with their corresponding CRE ages to look for evidence 113 of possible common impact events. Although Δ17O and Fo (Fig. 1.14) tend to correlate (Clayton and Mayeda 1988; 1996; Mittlefehldt et al., 1998; Rumble et al., 2010), an impact into a homogeneous area (Fig. 6.1) would be expected to produce a very tight cluster in both parameters.

Since ureilites with similar O isotope compositions and Fo would imply a genetic relationship, one might expect clusters of CRE ages to have similar properties if they are in fact from a common source of homogenous material.

Figure 6.1 (Left) Pre-Impact Disruption: Possible schematic of the UPB’s petrologic structure (after Goodrich et al., 2004), with numbers indicating the Fo number of olivine. After catastrophic disruption, material either reaccreted onto the UPB or formed offspring bodies (UDB). In either case, the post-disruption material may have ‘monomict’ main group ureilite material underlying polymict material on top. (Right) Post Impact Disruption: Excavation of material from the UDB as a result of an impact could result in a number of different mixtures. A) Part of the material could be ejected as a single, whole, monomict piece (e.g. Fo 83) with the above polymict material still attached. B) Some of the ejected material may not contain any polymict/regolith, and could be multiple monomict domains of different, or similar, Fo (e.g. 79,80). C) Alternatively, some of the released material might be a mixture of multiple monomict domains with the regolith still attached. The analysis in this work would assign scenario A to be a heterogeneous grouping, with monomict and polymict samples if Fo or oxygen isotopes do not agree. Scenario B would be classified as homogeneous (if the Fo are the same in addition to matching oxygen isotopes) or heterogeneous (if the Fo of the multiple monomict regions or oxygen isotopes are different) and would not contain polymict samples. This scenario would also be suitable to describe a larger impact where deeper material that has no contact to the polymict region would be released. Scenario C would be considered heterogeneous with polymict and monomict samples present if Fo and oxygen isotopes are not in agreement.

114

6.2 Methods

CRE (21Ne) ages are calculated for 31 ureilites by using Ne isotopic measurements from

Okazaki et al. (2003), Rai et al. (2003), and Schultz and Franke (2004); along with production rates and age formulations presented by Hohenberg et al. (1978) and Wieler (2002), respectively.

Production rates from Hohenberg et al. (1978) were adjusted to 4π exposure by multiplying the reported rate by a factor of two. Shielding effects are important to consider when determining an appropriate production rate of cosmogenic isotopes (see Chapter 2). The cosmogenic 21Ne/22Ne ratio of bulk samples vary by only ~10%. This corresponds to a shielding range of 20-100 g/cm2, which changes the possible production rates by up to ~15%. Shielding of 40 g/cm2, near the peak in production rate, was assumed for all samples. While these corrections are imperfect, they are self-consistent.

21 21 22 21 22 Cosmogenic Ne is calculated assuming ( Ne/ Ne)trapped = 0.03 and ( Ne/ Ne)cosmogenic = 0.9 as described in Rai et al. (2003). When calculating the 21Ne production rate, literature values are used for each sample’s chemical composition (Na, Mg, Al, Si, Ca, and Fe) when available. For samples whose chemical composition could not be found, an average of the remaining ureilites is used as a representative composition (Table A2, Rai et al., 2003 and references therein). To determine the effects of the chemistry, test calculations were performed using the full range of chemical compositions. This had little effect on the age, and therefore it is reasonable to use the average chemical composition when necessary.

Multiple CRE ages are calculated for samples that have multiple isotopic measurements available, yielding an average exposure age with the uncertainty expressed as the standard deviation. The standard deviations varied from 5% to 21% (average of 13%), with the exception of one sample with only two measurements that differed by 44%. For samples for which only one 115 measurement was available, we assumed a 20% uncertainty, a conservative value based on the standard deviation among other samples that have data from multiple measurements.

The ureilite Almahata Sitta was also added to this study and has many different reported ages.

This work used the average of 6 measurements of Almahata Sitta ureilite samples measured by

Welten et al. 2010, resulting in a 21Ne age of 15.8 Ma that agrees with the 21Ne age reported by

Welten et al. (2010) to within 10% (although they prefer a Ne-Al age of 19 Ma). Eight measurements in Table 6.1 marked with an asterisk come directly from an abstract by Park et al.

(2014). The CRE ages reported in their abstract are used for these samples, since isotopic data have not been published. Since these ages are not determined with the same method as done in this work, they do not contain the same systematic uncertainties. However, it is reasonable to include them in our analysis since indistinguishable results were obtained independently for common samples included in both studies. In all, 39 ureilite CRE ages are tabulated (Table 6.1).

CRE ages are plotted against literature values of oxygen isotopes (Δ17O) and forsterite content of olivine (Fo), i.e. 100*molar Mg/(Mg+Fe) (Clayton and Mayeda, 1988, 1996; Goodrich, 1999;

Kita et al., 2004; Table 23 of Mittlefehldt et al., 1998; Weber and Bischoff, 1998) to find if there are any groupings in all three variables. Uncertainties for Fo and Δ17O either reflect the uncertainty or deviation (specifically important for polymict ureilites where the entire range was used) in the literature or were assigned a conservative value of ± 2 when uncertainties could not be found for

116

Table 6.1. Key data for the 39 ureilites used in this work. Ureilite CRE age (Ma) CRE Error (Ma) Δ17O Fo # Polymict Acfer 277 0.53 0.1 -1.06 83.2 Allan Hills (ALHA) 77257 9.10 1.8^ -0.9 84.9 Allan Hills (ALH) 78019 0.09 0.01 -0.83 76.2 Allan Hills (ALHA) 81101 10.86 2.6 -0.23 79 Allan Hills (ALH) 82130 1.51 0.3 -2.53 95 Almahata Sitta 15.80 0.9 -1.22 88.5 Asuka (A-) 881931 3.85 1.9 - 76 Dar al Gani (DaG) 084 19.10 3.8 - 79.1 Dar al Gani (DaG) 319 21.67 2.2 - 79 X Dar al Gani (DaG) 340 8.74 0.1 -1.2 79.9 Dingo Pup Donga 4.24 0.8^ -1.29 84 Dyalpur 9.09 1.9 -1.57 84 Elephant Moraine (EET) 83225* 2.77 0.6 - 87 Elephant Moraine (EET) 83309 38.01 7.5 -1.2 82 X Elephant Moraine (EET) 87511* 3.30 0.7 - 85 Elephant Moraine (EET) 87517* 7.77 1.6 - 92 Elephant Moraine (EET) 87720 7.34 0.3 -0.53 83 X Goalpara 20.50 0.6 -0.4 78.6 Graves Nunataks (GRA) 95205* 10.09 2.1 - 79 Graves Nunataks (GRA) 98032* 15.32 0.8 - 75 Grosvenor Mountains (GRO) 95575* 5.22 1.0 - 78.7 Hajmah (a) 1.10 0.2^ - 85 Hammadah al Hamra (HaH) 064 0.76 0.2^ -0.74 78 Haverö 19.29 1.1 -0.52 79 Kenna 24.74 5.5 -1.02 79 Lahrauli 15.42 0.6 - 79 Lewis Cliff (LEW) 85328 11.52 2.3^ -0.57 80 Meteorite Hills (MET) 01083* 7.17 1.4 - 91.9 Meteorite Hills (MET) 01085* 25.13 5.0 - 90 Nilpena 9.02 1.2 -1.39 80 X North Haig 0.13 0.03^ -0.97 84 X Novo-Urei 5.90 0.8 -0.99 79 Pecora Escarpment (PCA) 82506 2.81 0.6^ -0.88 78.3 Reckling Peak (RKP) A80239 21.89 4.3 -1.12 84 Roosevelt County (RC) 027 2.11 0.4^ -1.08 79.4 Sahara 98505 13.20 2.5 - 81.2 Yamato (Y-) 74123 4.01 1.0 -0.81 79.2 Yamato (Y-) 790981 11.66 2.3^ -0.52 77.5 Yamato (Y-) 791538 2.64 0.5 -1.9 91.8 CRE age calculations are based on data from Okazaki et al., (2003), Rai et al., (2003) and Schultz and Frank, (2004). Oxygen isotopes and forsterite numbers were compiled from Clayton and Mayeda, (1988,1996); Goodrich, (1999); Kita et al., (2004); Mittlefehldt et al., (1998): see Table 23; Weber and Bischoff, (1988). CRE ages from meteorites denoted by an asterisk (*) are from Park et al., (2014). Cases where ages were determined from one measurement were assigned a 20% error and are denoted by a (^). 117

Fo. Oxygen isotopic values could not be found for every sample in this study – more Δ17O data would increase the effectiveness of this type of analysis if included in the future.

If all three sets of data (CRE, Fo, Δ17O) agree for a group of ureilites, it is considered to be a homogenous cluster, potentially a homogeneous event grouping. If one or two variables disagree, the CRE cluster is classified as a heterogeneous cluster. In some cases, there are homogenous subsets within a heterogeneous cluster. Possible explanations of this include coincident impacts into homogenous material on multiple independently homogenous bodies/regions, or a single impact into heterogeneous material. Reported cluster ages are determined by taking the average exposure age of the cluster and taking the standard deviation within the group as the uncertainty.

Alternatively, the uncertainty could be calculated by taking the root mean square of the CRE uncertainty, which does not change results significantly.

For the five polymict ureilites, I would expect the CRE ages of the various clasts (within the same sample) to agree, unless pre-compaction exposure occurred. I am not aware of any evidence of pre-compaction exposure within ureilites, although it cannot be ruled out. The measured O isotopes are expected to represent an average over the polymict materials measured, while the entire range of literature Fo values are used for polymict ureilites in comparisons with CRE ages.

Discussion of cluster classification is confusing, and the terms are listed here for a reference:

• Homogenous- Age, Fo, and Δ17O agree, likely representing collisional event on a parent body that is at least locally homogenous.

• Heterogenous- Age and Fo, or age and Δ17O agree, possibly representing a collisional event on a heterogenous parent body or multiple impacts on multiple heterogenous or homogenous bodies.

• Homogenous subgroup- subset of samples in a heterogenous group whose age, Fo, and Δ17O agree; could represent a separate impact at the same time or could be part of the heterogenous group.

118

6.3 Ureilite Exposure Age Results

CRE ages are reported in Table 6.1 and agree well with prior work where comparisons can be made (Rai et al., 2003). These ages are also shown in Figure 6.2 as a relative probability plot (each age is represented as a Gaussian distribution, all with the same area under the curve, so the width is proportional to the uncertainty and more precisely defined ages have narrower and taller

Gaussians) and Figure 6.3 in the form of a histogram for easy comparison to prior publications.

In general, there is a wide range of mostly randomly distributed CRE ages, with no definitive cluster identified. However, there is evidence of possible clusters, which we will discuss in order to illustrate the method of searching for and assigning clusters. Ages for potential clusters are around <1 Ma, ~1 Ma, ~2-3 Ma, ~4-5 Ma, ~9 Ma, and ~20 Ma based on ages alone.

When comparing the Fo to CRE age (Fig. 6.4), another parameter is added for determining the likelihood of source event clusters. Now possible clusters and subclusters include <1 Ma, ~1 Ma,

~3 Ma, ~4-5 Ma, ~9 Ma, ~12 Ma, ~15 Ma, and ~20 Ma.

Possible clusters from CRE and Δ17O comparisons (Fig. 6.5) include groups at <1 Ma, ~3-4

Ma, ~9-11 Ma, and ~20 Ma. Final determination of possible clusters can be seen in Table 6.2. The data were scrutinized carefully to look for signs of age clusters but nothing stands out. The clusters briefly discussed below are possible.

119

Figure 6.2. Relative probability plot (ideogram) of the CRE ages of ureilites in this study. Possible groupings appear to be at ~0.1, ~1, ~3, ~9, ~15 and ~20 Ma. Values of y-axis are arbitrary units.

Figure 6.3. Histogram of 39 ureilite CRE ages. Note that the data are much more scattered than that for the acapulcoites and lodranites in Fig. 1.8. Data is represented in ‘bins’ of 2 Ma.

120

Possible Homogenous Clusters:

Combining all three sets of data (CRE, Fo, and Δ17O), possible evidence of three homogenous clusters is present at 0.1 ± 0.02 Ma, 5.6 ± 0.5 Ma, and 20.1 ± 1.2 Ma. 1.) One main group and one polymict ureilite form the youngest cluster, 0.1 ± 0.02 Ma. 2.) The next cluster at 5.6 ± 0.5 Ma, from two main group ureilites, is missing one of the Δ17O values but is homogeneous in Fo and age, so although there is missing data, it is categorized as homogenous. 3.) The last homogenous cluster is at 20.1± 1.2 Ma and is formed from three main group and one polymict ureilite. Note that all homogenous parings have Fo ~79, but the two more recent groups have Δ17O of ~-1.0‰ while the older event has Δ17O of ~-0.5‰.

Possible Heterogeneous Clusters: Several clusters agree in CRE age and Fo, but not Δ17O. There is no case of agreement between

CRE age and Δ17O only. Seven heterogeneous clusters are at 0.7 ± 0.1 Ma, 1.3 ± 0.3 Ma, 3.4 ± 0.7

Ma, 7.4 ± 0.3 Ma, 9.3 ± 0.4 Ma, 12.4 ± 0.7, and 15.5 ± 0.3 Ma, four of which show the possibility of homogenous subclusters (see methods and discussion).

1.) There are two homogeneous subclusters that form the group at 3.4 ± 0.7 Ma. 2.) The heterogeneous cluster at 7.4 ± 0.3 Ma is a combination of a homogeneous subcluster and a single sample. 3.) Another distinct cluster with two possible homogeneous subclusters is at 9.3 ± 0.4 Ma.

4.) The last heterogeneous cluster of this type is at 12.4 ± 0.7 Ma. All the meteorites in these clusters are main group except for Nilpena and Elephant Moraine 87720, which are polymict and are each in different groupings. Half of these groups contain samples that have an Fo ~79.

There are three possible clusters that are heterogeneous with no evidence of homogeneous subclusters. 1.) The youngest such cluster is at 0.7 ± 0.1 Ma, followed by 2.) a heterogeneous 121 cluster at 1.3 ± 0.3 Ma. 3.) The last cluster of this type is at 15.5 ± 0.3 Ma and includes Almahata

Sitta. At least one sample in each of these clusters has Fo of 79.

Figure 6.4. Ureilite Fo vs CRE Age. This shows an additional constraint on Fig. 6.2 since it shows samples that have the same CRE age and Fo (within error), which might suggest these samples are material ejected from the same impact (shaded groupings). Triangles denote samples that are polymict.

Table 6.2. Possible Clusters. Age (Ma) Type Meteorites 0.1 ± 0.02 Homogeneous ALH 78019, North Haig (P) 0.7 ± 0.1 Heterogeneous Acfer 277, Hah 064 1.3 ± 0.3 Heterogeneous ALH 82130, Hajmah (a) 3.4 ± 0.7 Heterogeneous* (A-881931, PCA 82506, Y-74123), (Dingo Pup Donga, EET 83225, EET 87511), Y-791538 5.6 ± 0.5 Homogeneous (?) GRO 95575, Novo-Urei 7.4 ± 0.3 Heterogeneous* (EET 87515, MET 01083), EET 87720 (P) 9.3 ± 0.4 Heterogeneous* (ALHA77257, Dylpur, (Nilpena (P), DaG 340) 12.4 ± 0.7 Heterogeneous* (LEW 85328, Sahara 98505, Y-790981), (ALHA81101, GRA 95205) 15.5 ± 0.3 Heterogeneous Almahata Sitta, GRA 98032, Lahrauli 20.1 ± 1.2 Homogeneous DaG 084, DaG 319 (P), Goalpara, Haverö Heterogeneous clusters that have an asterisk signify the presence of homogeneous subclusters, which are listed in parenthesis. The homogeneous cluster at 5.6 Ma has a question mark (?) since it is homogenous in Fo and age, but oxygen data is lacking. Polymict ureilites are denoted by a “P” in parentheses. Details about individual meteorites are in Table 6.1.

122

Figure 6.5. Ureilite Δ17O vs CRE ages showing additional constraints on Figs. 6.2 and 6.4. Now samples can be 17 evaluated to check for agreement in CRE age, Fo, and Δ O values. Triangles denote polymict samples.

6.4 Discussion

If the parent body of the collection of ureilites is in fact as described in Figure 6.1 (note different excavation possibilities; A-C), what would the resulting debris from an impact look like? It is reasonable to postulate that an impact could break apart main group ureilite material (with a single

Fo) along predefined fracture/weak points, which would then be launched simultaneously (i.e., homogeneous cluster). Alternatively, an impact could excavate ureilite materials composed of various Fo (i.e., a heterogeneous cluster), as would be expected if the UPB was completely disrupted and reaccreted (Bischoff et al., 2010; Goodrich et al., 2015).

However, the ureilites do not show significant evidence for groupings, certainly not at the level of, for example, the acapulcoites/lodranites (Fig. 1.10). The CRE data for the ureilites are rather 123 scattered, but possible groupings that, as a whole, are unconvincing will be discussed. Defining a cutoff for the range of Fo, age, or Δ17O is arbitrary at this point so at best these hint at weakly developed groups. It is clear that most ureilites did not originate from one or two impact events, but rather multiple impacts.

This work shows possible evidence of three homogenous clusters, which would suggest a homogeneous source body, or a source body that is at least locally homogenous, if considered alone. The cluster at ~0.1 Ma is based on very weak evidence since it has only two samples and the polymict Fo covers a wide range of Fo and Δ17O values. The cluster at 5.6 ± 0.5 Ma only has one known oxygen value, so there is a possibility that this is a heterogeneous cluster. The cluster at 20.1 ± 1.2 Ma includes one polymict ureilite and three main group ureilites. It was decided to include all four samples into this source cluster due to the agreement in age and Fo, despite the fact that no Δ17O data could be found for two of the meteorites in this cluster. The Fo for the homogeneous clusters are all ~79 (not surprising given the prevalence of Fo ~79 among ureilites), which, combined with the fact that there are only two distinguishable oxygen isotope values among these clusters, might suggest that the source body is dominated by Fo ~79 material.

However, there are many more possible heterogeneous clusters identified in this work, suggestive of a heterogeneous source. Cases where there is a cluster in ages with heterogeneous

Fo and/or Δ17O could represent a single impact into a heterogeneous source, or alternatively two or more separate impact events on different bodies around the same time.

Classification of homogeneous subclusters within a cluster was determined based on the natural segregation of the Fo (i.e., the cluster at 3.4 Ma, the most likely example of a cluster, has two sets of three samples that have the same Fo, making it unclear if this set represents an impact into heterogeneous material or different impacts at roughly the same time). 124

Several samples in this study lack oxygen isotopic data (Table 6.1), which if known would help to better determine the classification scheme used in this work. Two samples in three heterogeneous clusters (7.4 ± 0.3 Ma and 9.3 ± 0.4 Ma and 12.4 ± 0.7) have a sample that has an

Fo that overlaps with two clusters. A decision was made to include it in one of the clusters since inclusion in one or the other does not significantly change results.

To be clear, a heterogeneous cluster does not exclude that these samples were liberated from the same impact event but could reflect the degree of homogeneity of the parent (or daughter) body. For example, if a cluster agrees in CRE age, and one sample has Fo of 79 and another sample has Fo of 84, they could be two separate and neighboring ‘monomict’ regions ejected from the same impact (Fig. 4.1). Depending on the true scale of heterogeneity of the ureilite source body, the case of a small disagreement in oxygen isotopes does not exclude that these samples are from the same event. If the two sets of meteorites are from the same event, the heterogeneity in isotopes and chemistry may be due to the fact that they were located far apart during a large collision.

Further, if the UPB was completely disrupted and reaccreted, CRE age clusters might be independent of petrogenetic properties, and we would expect heterogeneous clusters or locally homogeneous clusters, which is supported by this work.

The models of offspring bodies (Goodrich et al., 2002, 2004) suggest the possibility that rubble of polymict material overlays heterogeneous patches of main group ureilite material. Of the possible source event parings suggested in this work, three events have a mixture of main group and polymict ureilites, while the seven other tentative events contain only main group ureilites. If these proposed events are truly from a small number of impacts on the ureilite source(s), then the cases in which there are both polymict and main group ureilites means these events potentially have material from both the surface and underlying layers that survived the impact and journey to 125

Earth. Alternatively, these clusters could be coincidences and are not real. Furthermore, this study includes just 39 out of over 350 individual ureilites (Meteoritical Bulletin Database), so much more insight into possible source paired events could be achieved from further study using this approach.

Goodrich et al. (2015) suggest the possibility of the breakup of the UDB at ~20 Ma, consistent with our oldest proposed cluster.

6.5 Conclusion

Ureilites are a heterogeneous group of achondrites with both primitive and differentiated characteristics. Goodrich et al. (2004) proposed a model that has polymict material overlaying main group (a.k.a. monomict) material of various Fo. We do not see evidence of a large ~50 Ma break up, unless smaller pieces were larger than multi-meter sized (providing shielding from cosmic rays), but we do see evidence of a possible break up event ~20 Ma.

CRE ages of ureilites generally show a wide distribution and show that samples do not come from just one or two major impact events. Several possible clusters have been further constrained by considering Fo and oxygen isotopic data. Although no ureilite clusters or events can be confirmed, the technique for classification of clusters and its implications are discussed. This correlation of data has produced possible homogenous clusters at ~0.1 ± 0.02 Ma, 5.6 ± 0.5Ma, and 20.1 ± 1.2 Ma, which suggest some degree of homogeneity in the ureilite source(s), at least regionally. Two of these clusters contain both polymict and main group ureilites, which may hold interesting consideration for heterogeneity, impact parameters, and the structure of the ureilite source.

Signs of seven heterogeneous clusters have been identified, that if real, could represent coincidental source events into different bodies or different areas of a heterogeneous body. Rai et al. (2003) suggest possible clusters in ureilites (~1 and ~10 Ma), which we also see in our 126 heterogeneous clusters, with the addition of more data and further chemical and isotopic constraints that can be used as a future analysis technique. This work demonstrates that the ureilites as a whole show a wide variety among CRE ages, unlike the H-chondrites and acapulcoites/lodranites (Fig. 1.10) that show more distinct clusters (Eugster et al., 2006).

127

Chapter 7: Cosmic-Ray Exposure Ages of Brachinites and other Achondrites 7.1 Introduction

This work includes brachinites, brachinite-like achondrites, and ungrouped achondrites that may have links to the brachinites based on mineralogy, oxygen isotopes, textures, or other properties. When referring to all of these types of samples together, the term ‘UMA’ will be used

(brachinite/like/ungrouped Ultra-Mafic Achondrites). Regardless how they are categorized, these meteorites are ultramafic, mostly unshocked (with exception of EET 99402- Mittlefehldt et al.,

2003), primarily unbrecciated, and are mostly olivine (~70-96 vol.%, ~Fo64-73) with various amounts of clinopyroxene (up to 15%, ~En40-63Wo36-48), orthopyroxene (up to 20%, ~En69-73Wo2-

4), and plagioclase (up to 10%, ~An15-33) with accessory chromite, sulfide, and phosphates (Day et al., 2012; Keil 2014). Mineral compositions are homogenous within a meteorite (Keil 2014). Their olivines have medium/course-grained textures (~0.1-1.5 mm, e.g. Mittlefehldt et al., 2003) that have been described as equigranular, xenomorphic-granular, and protogranular (Nehru et al., 1983;

Mittlefehldt et al., 2003; and Irving et al., 2009 respectively) and tend to have triple junctions (e.g.

Mittlefehldt et al., 2003; Krot et al., 2014). Their O-isotopes span from Δ17O = –1.3 to +0.66‰

(Fig. 7.1), but most are in the range of –0.08 to –0.39‰ (Day et al., 2012; Greenwood et al., 2012;

2017).

Brachinite-like achondrites are not clearly defined other than that they share a similar composition to brachinites. Some consider them to have a more magnesian olivine and a higher volume % of orthopyroxene (Day et al., 2012). It is not clear if they are part of the brachinite group and some authors have classified them as being ‘brachinite-like’ or as an ungrouped achondrite

(e.g. Day et al., 2012; Singerling et al., 2013). Two feldspar-rich achondrites, Graves Nunataks

(GRA) 06128 and 06129, are also suggested to be partial melts from an oxidized parent body and may be related to brachinites (Day et al., 2009; 2012). 128

Brachinite petrogenesis is often considered to be from either a cumulate origin or partial melt residue. A cumulate origin is based on weak evidence of a ‘possible’ grain orientation in olivine grains of ALH 84025 (Warren and Kellemeyn, 1989) and the observation of a common birefringence in olivine grains that were used to speculate about possible similar crystallographic orientation (Mittlefehldt et al., 2003). However, their results are not conclusive, rather only suggest a possibility, and brachinites contain Fe-Ni metal, which is not expected to be present in cumulate rocks.

The range in oxygen isotopes (Fig. 7.1) could indicate formation on several genetically unrelated parent bodies that sample similar, yet different, oxygen reservoirs (Greenwood, 2007), or could be the result of formation from a single undifferentiated parent body that did not experience thorough homogenization of oxygen isotopes from sufficient heating (Rumble, 2008).

A residue formation of brachinites would preserve the observed oxygen heterogeneities, and a cumulate origin would likely require multiple different parent bodies (though Mittlefehldt et al.,

2003 suggested different oxidation states of different magma regions on a brachinite parent body could result in heterogenous oxygen isotopes).

129

Figure 7.1. Oxygen isotopes of achondrites (brachinites, winonaites, acapulcoites, lodranites, brachinite-like achondrites, and ungrouped achondrites (Greenwood et al., 2017). Each meteorite group occupies relative distinct areas that can be used as part of classification criteria. Samples included in this work are circled in purple.

The Fe-Mg-Mn compositions of olivine are useful for comparing meteorite groups and offer insights into petrogenesis (Goodrich and Delaney, 2000). The relationship of the UMA Fe-Mg-

Mn system demonstrates that these samples experienced similar Mn-Mg and metal/silicate fractionation under highly oxidizing conditions (Fig. 7.2) of primarily chondritic material

(Goodrich et al., 2010). High CaO content of brachinite olivine (~0.1-0.24% compared to <0.04% in ordinary chondrites) indicates processing under igneous temperatures (Smith et al., 1980) consistent with olivine-chromite temperatures ranging from ~950-1250°C, higher than the Fe,Ni-

FeS eutectic (950°C) and in most cases higher than the eutectic of ordinary chondrites (1050°C) 130 indicating silicate melt was produced (Gardner-Vandy et al., 2013, Goodrich et al., 2010; McCoy et al., 2006). High oxidization state has also been confirmed by oxygen fugacity experiments

(Table 7.1) that show conditions for forming brachinite mineralogy occur at ~ iron-wustite buffer

(IW) ~0 to ~IW -1.2, and are more reducing for other UMA samples (Gardner-Vandy et al., 2013).

TableTable 7.1. 1. Δ ΔIWIW of of br abrachiniteschinitic and and oth eotherr ach oacndhondrites.rites. Meteorite Type Olivine–chromite (°C) ΔIW Reference Brachina Brachinite 1250 ± 19 -0.3 1 ALH 84025 Brachinite 970 ± 46 -1.0 1 Hughes 026 Brachinite 1038 ± 21 -1.1 1 EET 99402 Brachinite 980 ± 16 - 1 NWA 3151 Brachinite 1054 ± 27 -1.0 1 NWA 4969 Brachinite 1032 ± 15 -1.0 1 NWA 5400 Brachinite-Like 1028 ± 22 −1.1 1 ALH 84027 Ungrouped -1.4 1 LEW 88763 Ungrouped -0.7 1 RBT 04239 Ungrouped -1.6 1 Divnoe Ungrouped -1.8 2 Winonaites -2.5 to -3.5 3,4 IAB, IIICD, IIE irons -2.5 to -4 3,4 Acaplcoites -2 4 Lodranites -2 4,5 H -2.5 4 L/LL -2.5 to -2.5 4 CK Chondrites 7.5 6 R Chondrites 0 to +3 6

ΔIWΔIW is is thethe deviationdeviation ooff tthehe ccalculatedalculated f fOO22 f rfromom t hthee ir ironon-w-uwustitestite bu bufferffer in linog log uni tunits.s. 1.) Gardner-Vandy et al., 2013. 2.) Pataev et al., 1994. 3.) Benedix et al., 2005. 4.) Righter 1.) Gardner-Vandy et al., 2013. 2.) Pataev et al., 1994. 3). Benedix et al., 2005. and Drake, 1996. 5.) Benedix et al., 2009. 6.) Righter and Neff, 2007. 4.) Righter and Drake (1996(. 5.) Benedix et al., 2009. 6.) Righter and Neff, 2007.

131

Figure 7.2. (A.-Top) Fe-Mn-Mg plot for brachinites, lodranites, winonaites, augite-bearing ureilites. Brachinites are the most oxidized of these groups and appear to have a redox relationship. Linear trends that pass through the origin signify addition or loss of Fe at a constant Mg/Mn ratio. (B.-Bottom). Horizontal trends indicate large changes in Fe/Mg with little change in the Fe/Mn, as expected from formation of fractional melts and cumulates. The relative locations of FeO/MgO and FeO/MnO are shown for chondritic material that undergoes equilibrium melting (2% melt) which is then fractionated or forms a cumulate. At higher degrees of melting (55% in the figure), the melt returns closer to the bulk composition. The bulk material develops into a residue with a lower Mn/Mg. Figures are from Goodrich et al., 2010 and Goodrich and Delaney, 2000.

132

Abundance patterns for highly siderophile elements in brachinite and brachinite-like samples

(Fig. 7.3) show low iridium, platinum, and palladium relative to ruthenium and osmium, which could be used as a potential distinguishing characteristic of brachinites. This might be an effect of different partition coefficients of HSE during melting (Chabot and Jones, 2003) as a result of different sulfur concentrations in the Fe-Ni-S melt (Day et al., 2012; S. Crossley-personal communication).

Fig. 7.3. High Siderophile Elemental abundance patterns for brachinites (solid lines) and brachinite-like and select ungrouped achondrites (dashed lines). Perhaps a distinguishing characteristic is that brachinites (excluding Brachina) have low Ir, Pt, and Pd relative to Ru and Os. Day et al., 2012.

7.2 Brachinite, Brachinite-like, and Ungrouped Classification An updated version of the Fe-Mn-Mg plot of olivine (e.g. Goodrich and Righter, 2000;

Goodrich et al., 2010) for brachinites can be found in Fig. 7.4, with additional UMA samples that were discussed above. Two approximately continuous sets of data that are separated around Fe/Mg

= 0.45 are observed. The further samples plot to the upper right of Fig. 7.4, the more oxidized they are (enrichment in Fe, see Goodrich and Delaney, 2000). Most of the samples in the reduced group 133 are considered brachinite-like, while most of the samples in the oxidized group are often considered brachinites (with exception of LEW 88763). NWA 6962 and NWA 4518 were included in this work based on brachinite characteristics (e.g. basic minerology, Δ17O) but plot outside the Mn/Mg space defined by the brachinite and brachinite-like samples (upper boundary set at Mn/Mg = 158, lower boundary set to Mn/Mg = 129). Igneous processing (e.g. fractional crystallization from a low degree (~few %) of equilibrated melt), can increase the Fe/Mg ratio relative to the bulk composition and is a possible explanation for the Fe-Mn-Mg location of NWA

6962. On the other hand, it may be from another parent body. NWA 4518 likely requires a different

Figure 7.4 Fe-Mn-Mg characteristics for brachinite and brachinite-like samples. Data appears to be separated into two groups. The relatively oxidized group could be a way to distinguish ‘brachinites’ from ‘brachinite-like’ achondrites. Data used for this figure can be found in Keil, 2014; Gardner- Vandy, 2013; Lorenz, 2011; Goodrich and Delaney, 2000; Goodrich et al., 2017; Goodrich et al., 2012; Warren and Rubin, 2012; and Greenwood et al., 2017.

134 explanation, such as crystallization from a high degree (~50%) of partial melt or formation from a cumulate (Goodrich and Delaney, 2000) if it is in fact related to the brachinites. Alternatively, it could be igneous processing from another source (HED or IIAB, Lorenz et al., 2011).

What defines a brachinite, brachinite-like, or an ungrouped achondrite with brachinite affinity?

The separation of oxidized and reduced UMA samples offers a natural place to distinguish between the two. This is not the only possibility, but it at least provides a framework for discussion and provides consistency. Plotting the Δ17O (Fig. 7.5) data for these groups further emphasizes their distinction and shows that the oxygen isotope values cover the same range in the brachinites

(oxidized) as the brachinite-like (reduced) group, with a general trend of increasing Δ17O with increasing Fe/Mg. This positive correlation is also seen in ureilites, lodranites, and H, L, LL chondrites and is attributed to highly reducing conditions early in the solar nebular which is coupled with metal/silicate fractionation (Goordich and Delaney, 2000).

We define a sample to be a ‘brachinite’ if the following apply: 1) Δ17O values are between

0.04‰ and -0.40‰, 2) Fe/Mg in olivine is 0.449-0.563 (Fo = 64-69), and 3) the Mn/Mg in olivine is 129-159. The first criterion is based on the reasonable range of variation in Δ17O for the majority of brachinitic samples including Brachina (and removes apparent outliers such as LEW 88763).

Criteria 2 and 3 are based on the characteristics of the UMA samples in the Fe-Mn-Mg system

(Fig. 7.4), with the more oxidized samples considered brachinites, and the more reduced samples considered brachinite-like. ‘Brachinite-like’ samples satisfy criteria 1 and 3, but have Fe/Mg

0.350-0.450 in their olivine. If a sample does not satisfy the brachinite or brachinite-like sets of parameters, it is considered an olivine-rich ungrouped achondrite (Table 7.2).

Other criteria that could be considered are the CaO and Cr2O3 content of olivine (suggested limits: 0.09-0.25 wt. % and 0.025-0.1% wt. % respectively), oxygen fugacity (IW: 0 to -1.5), and 135 the presence of platinum and palladium depletion relative to iridium, and/or iridium depletion relative to ruthenium and osmium (Goodrich et al., 2011, Garder-Vandy et al., 2013, Day et al.,

2012, S. Crossley- personal communication). These criteria are not included in this work because data are not available for many samples (Table 7.3), however they do not change the classification in Table 7.2 in cases where the data are available.

Figure 7.5. Δ17O vs Fe/Mg of oxidized and reduced UMA (as separated in Figure 7.4), Δ17O values cover the same range in the brachinites (oxidized) as the brachinite-like (reduced) group, with a general trend of increasing Δ17O with increasing Fe/Mg. This positive correlation is also seen in ureilites, lodranites, and H, L, LL chondrites and is attributed to highly reducing conditions early in the solar nebular that is coupled with metal/silicate fractionation. Oxygen uncertainites are all less than 0.08 per mil.

136

Table 7.2. Conditions for brachinite, brachinite-like, and ungrouped achondrites. Sample Δ17O (‰) Fe/Mg Mn/Mg Classification NWA 3151 -0.173 0.552 146.7 Brachinite NWA 4874 - 0.517 154.0 Brachinite NWA 4876 - 0.502 135.0 Brachinite NWA 4882 -0.241 0.541 131.1 Brachinite NWA 4969 - 0.524 139.0 Brachinite NWA 6474* -0.062 0.558 131.5 ± 4.5 Brachinite RaS 309 -0.190 0.532 - Brachinite ALH 84025 -0.300 0.492 134.9 Brachinite Brachina -0.318 0.466 152.2 Brachinite Eagles Nest -0.193 0.462 131.2 Brachinite EET 99402 -0.127 0.554 136.5 Brachinite Hughes 026 -0.213 0.517 132.9 Brachinite NWA 1500* -0.235 0.451 ± 0.027 134.3 Brachinite-like NWA 595 -0.154 0.395 139.1 Brachinite-like NWA 6077 -0.050 0.429 - Brachinite-like NWA 7297* - 0.440 142.3 Brachinite-like NWA7605 -0.176 0.351 130.9 Brachinite-like NWA 10637 -0.216 0.367 141.9 Brachinite-like NWA 4518 -0.209 0.471 75.65 Ungrouped NWA 6962 -1.037 0.923 76.0 Ungrouped NWA 8777 -0.291 0.342 137.0 Ungrouped LEW 88763 -1.260 0.561 139.0 Ungrouped Criteria: Δ17O (‰) Fe/Mg Mn/Mg Brachinite: -0.04 to -0.40 0.449-0.563 129-159 Brachinite-like: -0.04 to -0.40 0.350-0.450 129-159 Bold values indicate conditions for a brachinite are met. * indicates difficult classification due to uncertainties. Oxygen uncertainty is always below 0.07 per mil. Fe/Mg, Mn/Mg is in olivine. See references in Keil, 2014. 137

Table 7.3. Additional conditions for brachinite classification.

Sample Fo CaO wt.% Cr2O3 wt. % Pt & Pd Depletion ΔW NWA 3151 64.4 0.24 b.d. Yes -1.0 NWA 4874 65.9 - - NWA 4876 NWA 4882 64.9 - - Yes NWA 4969 65.5 0.12 -1.0 NWA 6474* 64.1 - - RaS 309 66.9 - - yes-pc ALH 84025 66.9 0.0997 0.046 -1.0 Brachina 68.4 0.2485 0.075 Yes -0.3 Eagles Nest 68.0 0.095 0.04 EET 99402 64.2 0.105 0.028 Yes Hughes 026 65.5 0.1225 0.026 -1.1 NWA 1500* 65-73 0.09 0.04 Yes NWA 595 71.7 0.082 0.028 - - NWA 6077 69.6 - - - - NWA 7297* 69.5 0.076 0.025 NWA7605 74.0 - - No-pc NWA 10637 73.2 0.072 0.035 - - NWA 4518 68.0 0.06 0.04 - NWA 6962 52.5 0.45 0.04 - - NWA 8777 69.1 - - - - LEW 88763 63.5 0.13 0.04 -0.7 Criteria 63-69 0.09-0.25 0.025-0.1 yes 0 to -1.5 Bold values indicate conditions for a brachinite are met. * indicates difficult

classification due to uncertainties. Fo, CaO wt% and Cr2O3 wt% are referring to olivine. Day et al., 2012; Gardner-Vandy, 2013; Keil, 2014.

7.3 UMA Noble Gases

Trapped heavy noble gases for brachinites indicate trapped components are a mixture of the quintessence phase “Q” with air ( 132Xe =1-1.75e-10 cm3/g; 36Ar/132Xe ~100-550 for Hughes 026,

Reid 013, and EET 99402, 132Xe =0.4-2.4e10 cm3/g; 36Ar/132Xe ~200-550 for Eagles Nest, and

132Xe =0.4-0.6e-10 cm3/g; 36Ar/132Xe ~200-900 for Brachina and ALH 84025, Patzer et al., 2003;

Swindle et al., 1998; Ott et al., 1985;1993). Phase Q refers to the primordial noble gas representative of the early solar nebula, and has been detected in achondrites (Busemann et al., 138

129 129 2000; Busemann et al., 2003). Many samples show production of Xe from decay of I (t1/2 =

16 Ma), indicating brachinites contained 129I during accretion relative early in the solar system which has not been degassed from the sample.

Brachinites formed in the first ~5 Ma of the solar system, based on 53Mn-53Cr age of 4563.7 ±

0.9 Ma for Brachina (Wadhwa et al. 1998), which is considered the best estimate of the crystallization age of the brachinite group. The 53Mn-53Cr age of brachinite NWA 4882 (Dunlap et al., 2016) is 4550.2 ± 0.8 Ma, which is considerably younger and suggests either a different parent body than Brachina or a shared parent body large enough to maintain significant heat over

~14 Ma. Other dating methods used on this group include U/Th-4He, K-Ar, and 40Ar-39Ar (Patzer et al., 2003; Ott et al., 1985; Swindle et al., 1998; Mittlefehldt et al., 2003). Argon retention ages range from 1200-3000 Ma with U/Th-4He ranging from 150-250 Ma (EET 99402, Hughes 026, and Reid from Patzer et al., 2003). The result of these prior studies show various ages younger than the crystallization age, interpreted as degassing from impact events. 40Ar-39Ar studies have had mixed success, often suffering from experimental difficulty due to weathering and other effects (Swindle et al., 1998, Chapter 5), although Mittlefehldt et al., (2003) obtained a reliable resetting age of 4130 ± 60 Ma for brachinite EET 99402.

Cosmic-ray exposure (CRE) ages measure the transfer time in space to Earth since a meteorite was ejected from its parent body resulting from an impact or break-up event. Analysis of CRE clusters was used in addition to oxygen isotopic measurements to support the idea that the howardites, eucrites, and diogenites (HED) meteorites originated from the same parent body

(Eugster and Michel, 1995) as well as e.g. demonstrating a common break up event among acapulcoites (Eugster and Lorenzetti, 2005). On the other hand, in Chapter 6, I showed that ureilite

CRE ages do not cluster and are not dominated by a single event. Prior studies of CRE ages of 139 brachinites are limited in sample size (Patzer et al. 2003, Ott et al., 1985, Swindle et al., 1998,

Bogard et al., 1983). The combined literature data lists 21Ne measurements of 7 brachinites, which spread between ~3-50 Ma, with a group of three 21Ne CRE ages at ~50 ± 10 Ma (Eagles Nest, EET

99402, Hughes 026; Patzer et al., 2003.).

The goal of this work is to use noble gas mass spectrometry to expand the noble gas data set and to better understand the exposure and impact/collisional history from a suite of 16 brachinite and brachinite-like ungrouped achondrites. The combination of these new measurements with literature data may help address the classification of brachinites and brachinite-like achondrites with respect to their parent bodies from a different perspective. Furthermore, for the first time we hoped to provide a systematic study of 4He and 40Ar gas retention ages and of the potentially present minor trapped heavy noble gases in these samples.

7.4 Experiment and Methods

Details of the experimental procedure can be found in Chapter 2 (section 4). Additional compositional measurements of samples measured in this work are summarized in Table 7.4.

140

Table 7.4. Compositional measurements from this work, in wt.%. NWA 595 Olivine Orthopyroxene Clinopyroxene Chromite

Na2O 0.410 0.344 0.007

K2O

SiO2 37.426 54.165 53.066 0.142 MgO 36.234 26.753 16.094 4.957

Al2O3 0.410 0.836 11.477 CaO 0.074 1.125 21.987

TiO2 0.079 0.172 1.189 FeO 25.912 16.414 6.184 26.260 MnO 0.449 0.393 0.174

Cr2O3 0.027 0.207 0.571 54.262 Total 100.122 99.957 99.429 98.295 n 8 3 5 2

RaS 309 Olivine Clinopyroxene Chromite

Na2O 0.416 0.011

K2O

SiO2 36.195 52.530 0.139 MgO 32.529 15.102 4.628

Al2O3 0.003 1.019 13.704 CaO 0.074 23.060

TiO2 0.128 0.928 FeO 30.512 6.106 27.569 MnO 0.420 0.099

Cr2O3 0.024 0.605 51.196 Total 99.758 99.064 98.176 n 7 7 3

NWA 6077 Olivine Orthopyroxene

Na2O 0.030

K2O

SiO2 37.502 54.679 MgO 34.650 26.611

Al2O3 0.285 CaO 0.084 1.109

TiO2 0.008 0.076 FeO 27.113 16.061 MnO 0.454 0.432

Cr2O3 0.174 Total 99.811 99.458 n 14 6

Analysis was performed on CAMECA SX50 and SX100 microprobes under guidance of K. Domanik. Olivine, chromite, and pyroxene were measured at 15kV, 20 nA, plagioclase was measured with a defocused 5 m beam, 15kV, 10 nA. Compositions are very homogenous for each sample, for example, the standard deviation among 14 measurements of olivine is ~0.05%.

141

7.5 Cosmogenic Results

All noble gas data for the 16 meteorites measured in 30 aliquots are given in Table 7.5-7.8. The second aliquots for 6 samples have been analyzed for all Kr and Xe isotopes (see Experimental) as the first measurements (He-Ne, Ar-84Kr, 129,132Xe only) indicated the presence of well- measurable amounts of Kr and Xe (Tables S7.3 and S7.4).

There is no evidence of trapped neon among these samples. The measured 20Ne/22Ne ~1 (actual data ranges from 0.831-0.99) (Fig. 7.6) is suggestive of a purely cosmogenic origin of the detected

Ne. The 21Ne/22Ne ratios that indicate shielding conditions/radius of meteoroid and that are used to determine the production rates (3He, 21Ne, 38Ar) can be found in Table 7.5. The model often allows multiple possible meteoroid radii and sample depth combinations. Because no unique solution of shielding parameters could be found, the average production rate from all possible radii and depths was used to determine sample ages (Table 7.6).

142

Figure 7.6. Neon three-isotope plot for the samples measured in this work suggests a cosmogenic origin. Inset figure: Two measurements of brachinite-like NWA 6962 (left most green diamonds) and ungrouped achondrite LEW 88763 (left most purple triangles) lie outside of the range of model shielding possibilities and likely experienced effects of solar cosmic-rays (SCRs). Blue squares are brachinites, pink circles are brachinite-like, and green diamonds are ungrouped ultramafic achondrites all measured in this work. Purple triangles are recalculated literature values. Dashed line is between solar energetic particles, galactic cosmic rays, and air values. The other star symbols are for solar wind, Q, and HL (Weiler 2002; Ott 2002).

The measured 3He is assumed to be cosmogenic. The argon isotopes are a mix of a cosmogenic and trapped component, which can be solved for (as described in the methods section).

The 3He, 21Ne, and 38Ar exposure ages are listed in Table 7.6. In many cases the 38Ar age does not agree with 3He or 21Ne, which is attributed to chemical heterogeneity. The production rate of

38Ar, and therefore the age, is very sensitive to the assigned chemistry, most notably from Ca- bearing minerals. For example, an abundance increase in Ca of 1 wt.% is sufficient to lower the

38Ar age by up to 60% (and does not affect helium or neon results). Hence, I suspect that the variable 38Ar ages are symptomatic of chemical heterogeneity. Helium can experience heat- induced loss (through close approach orbits with the sun and/or atmospheric entry) but because the 3He and 21Ne ages generally agree, I do not think significant heat-induced loss occurred. 143

The sixteen meteorites’ 21Ne CRE ages range from ~10 to 65 Ma (Table 7.6, Fig. 7.7). In general, there is good agreement between one aliquot and another. NWA 6962 has low 21Ne/22Ne

(0.809 ± 0.005, 0.820 ± 0.003) which may be due to effects of SCR (Leya et al., 2000) in the uppermost centimeters of the meteoroid. It is not possible to determine the effects of SCR on the production of neon, and therefore the ages reported here are an upper limit for this sample.

Table 7.5. Cosmogenic abundances of samples in this work. Values are in (10-8 cm3 STP/g).

3 21 38 21 22 129 132 84 132 Sample Hecosm Necosm ArCosm Ne/ Ne Xe/ Xe Kr/ Xe NWA 10637 91.6 ± 2.1 22.2 ± 0.8 1.24 ± 0.47 0.939 ± 0.005 102.9 ± 1.0 3.6 ± 0.1 NWA 1500_1 14.4 ± 0.6 3.5 ± 1.0 0.26 ± 0.90 0.928 ± 0.007 104.0 ± 0.6 1.8 ± 0.0 NWA 1500_2 14.4 ± 0.5 3.5 ± 0.9 0.23 ± 1.05 0.933 ± 0.006 104.1 ± 1.8 1.5 ± 0.1 NWA 3151_1 36.9 ± 2.1 7.9 ± 0.7 1.08 ± 0.34 0.927 ± 0.003 104.5 ± 0.9 3.1 ± 0.1 NWA 3151_2 35.3 ± 2.3 7.6 ± 0.8 1.27 ± 0.27 0.927 ± 0.004 101.1 ± 0.9 3.1 ± 0.1 NWA 4518_1 13.5 ± 0.5 3.4 ± 0.8 0.23 ± 1.26 0.879 ± 0.004 133.0 ± 1.0 4.3 ± 0.1 NWA 4518_2 12.2 ± 0.5 3.2 ± 0.8 0.19 ± 2.20 0.875 ± 0.004 126.6 ± 1.4 3.1 ± 0.1 NWA 4874_1 43.7 ± 0.5 8.3 ± 0.7 0.67 ± 0.09 0.891 ± 0.004 104.2 ± 2.4 1.1 ± 0.1 NWA 4874_2 39.3 ± 2.0 7.9 ± 0.8 0.69 ± 0.07 0.898 ± 0.005 102.7 ± 1.0 1.5 ± 0.1 NWA 4876_1 108.4 ± 2.1 22.2 ± 1.5 1.94 ± 0.11 0.923 ± 0.017 119.0 ± 1.2 1.6 ± 0.1 NWA 4876_2 115.3 ± 2.5 23.2 ± 1.3 0.00 ± 0.00 0.919 ± 0.011 120.0 ± 0.6 1.4 ± 0.0 NWA 4882_1 114.9 ± 2.4 22.6 ± 2.4 1.69 ± 0.12 0.903 ± 0.022 119.1 ± 0.8 1.0 ± 0.0 NWA 4882_2 115.3 ± 2.5 23.3 ± 1.0 1.56 ± 0.14 0.927 ± 0.009 121.4 ± 1.0 1.2 ± 0.0 NWA 4969_1 102.6 ± 2.1 21.1 ± 2.5 1.71 ± 0.13 0.941 ± 0.019 110.8 ± 0.9 1.9 ± 0.1 NWA 4969_2 117.3 ± 2.5 - 3.75 ± 0.04 0.964 ± 0.026 107.2 ± 0.6 2.5 ± 0.0 NWA 595_1 40.4 ± 2.6 9.2 ± 0.8 3.94 ± 0.05 0.928 ± 0.005 102.3 ± 1.5 6.1 ± 0.1 NWA 595_2 51.1 ± 2.0 11.7 ± 0.9 0.61 ± 0.58 0.922 ± 0.005 103.7 ± 0.6 6.2 ± 0.1 NWA 6077_2 45.0 ± 2.0 10.4 ± 0.9 0.56 ± 0.77 0.941 ± 0.006 131.5 ± 0.7 4.3 ± 0.1 NWA 6474_1 63.6 ± 2.0 11.4 ± 0.7 0.51 ± 0.11 0.895 ± 0.004 104.9 ± 0.7 1.3 ± 0.0 NWA 6474_2 62.1 ± 2.0 11.3 ± 0.8 1.13 ± 0.20 0.910 ± 0.004 101.2 ± 1.6 1.3 ± 0.1 NWA 6962_1 24.4 ± 0.5 2.7 ± 0.9 1.11 ± 0.14 0.809 ± 0.005 92.0 ± 16.0 1.3 ± 0.7 NWA 6962_2 25.0 ± 2.2 2.9 ± 0.7 0.98 ± 0.09 0.820 ± 0.003 95.6 ± 12.4 0.7 ± 0.2 NWA 7297_1 43.4 ± 2.0 10.3 ± 0.9 0.67 ± 0.01 0.952 ± 0.006 173.9 ± 2.0 3.9 ± 0.1 NWA 7297_2 50.2 ± 0.6 12.0 ± 0.9 0.00 ± 0.00 0.943 ± 0.007 155.6 ± 0.7 3.8 ± 0.1 NWA 7605_1 78.5 ± 2.3 18.1 ± 3.4 0.61 ± 0.01 0.934 ± 0.034 148.1 ± 0.8 1.4 ± 0.0 NWA 7605_2 77.5 ± 2.1 17.9 ± 1.3 0.95 ± 0.35 0.924 ± 0.012 140.0 ± 1.0 1.8 ± 0.0 NWA 8777_2 32.7 ± 0.5 6.6 ± 1.1 0.39 ± 1.07 0.881 ± 0.007 97.0 ± 1.1 1.0 ± 0.0 RaS 309_2 77.0 ± 2.0 17.9 ± 2.2 1.34 ± 0.21 0.959 ± 0.023 98.9 ± 0.8 7.6 ± 0.1 Abundances are in (10-8 cm3 STP/g). The ‘1’ and ‘2’ signify different aliquots of the same sample. Exposure ages are in Table 7.6.129Xe/132Xe values are multiplied by 100.

144

Table 7.6. Shielding parameters and 3He, 21Ne, and 38Ar CRE ages. 3 21 38 Sample He Age (Ma) Ne Age (Ma) Ar Age (Ma) Radius (cm) Depth (cm) NWA 3151 20.5 ± 0.8 22.1 ± 0.6 37.4 ± 4.1 20-85 5-8 NWA 4874 24.5 ± 1.8 27.4 ± 1.1 23.3 ± 0.7 20-25 2-3 NWA 4876 62.6 ± 2.8 63.1 ± 1.8 63.3 ± 2.1 20-85 3-10 NWA 4882 65.2 ± 2.4 66.5 ± 2.5 63.1 ± 2.3 20-85 2-9 NWA 4969 57.5 ± 5.5 50.9 ± 2.5 66.9 ± 2.7 20-85 9-55 NWA 6474 37.5 ± 1.4 36.1 ± 2.9 31.6 ± 2.9 20-85 2-4 RaS 309^ 40.2 ± 2.4 41.1 ± 3.2 50.0 ± 3.0 35-85 17-56 NWA 1500* 8.1 ± 0.3 9.5 ± 0.3 9.1 ± 0.8 20-85 5-10 NWA 595* 25.4 ± 4.2 27.4 ± 4.7 24.6 ± 2.6 20-85 4-8 NWA 6077*^ 24.2 ± 1.2 26.6 ± 1.1 30.3 ± 1.2 20-85 7-17 NWA 7297* 24.6 ± 2.5 27.9 ± 2.9 6.1 ± 3.5 30-85 12-26 NWA 7605* 42.8 ± 1.7 48.0 ± 2.4 43.7 ± 1.5 20-85 3-10 NWA 10637* 48.7 ± 2.6 53.1 ± 2.2 52.5 ± 2.7 20-85 6-14 NWA 4518' 8.2 ± 0.6 13.7 ± 1.0 5.4 ± 0.7 ~10 3-4 NWA 6962' 17.1 ± 0.4 16.3 ± 1.1 18.0 ± 1.1 10.00 0-2 NWA 8777' 19.4 ± 0.8 23.1 ± 1.0 29.7 ± 1.0 20-85 1-2 3He, 21Ne, and 38Ar CRE ages for 15 samples measured in this work, in addition to sample ages recalculated from literature isotopic measurements. Ages and uncertainty are in millions of years, with the uncertainty reported as the more conservative value of standard propogation of error or the standard deviation among repeated measurements. The size of the pre-atmospheric body as determined from the model in Leya and Masarik (2009) is givin in the ‘Radius’ column, with the depth from the surface given in the ‘Depth’ column, both of which are in centimeters. An * denotes samples that are are brachinite-like, ’ are for ungrouped achondrites, and ^ results are from one aliquot only. NWA 6962’ is an upper estimate because of possible SCR effects (see text).

145

Figure 7.7. 21Ne CRE ages for brachinites (blue squares), brachinite-like (pink circles), and ungrouped olivine- rich achondrites (green diamonds), with recalculations of literature data (purple triangles and diamond). The histogram has a 2 Ma bin size, which does not reflect the age uncertainty. The histogram axis is on the right.

7.6 Discussion

A simple histogram does not address the age uncertainty when assigning data to a bin and can be misleading, but for comparison to prior work (Patzer et al., 2003), a histogram of 21Ne CRE ages is provided in Fig. 7.7. Data points are plotted with the uncertainties in addition to the histogram, which has a 2-million-year bin size. Each sample’s CRE age in each isotopic system

(3He, 21Ne (Fig. 7.8), and 38Ar) are comparable, although argon is particularly variable and does not agree as well as helium and neon . Solar heating can cause diffusive loss of helium, significant amounts of trapped argon can inflate uncertainties, and argon is susceptible to being lost by terrestrial weathering (Gibson and Bogard 1978; Scherer et al., 1994; Patzer and Schultz 2001).

For these reasons, neon is typically considered to provide the most robust CRE ages. The 3He and

21Ne ages typically agree within ~10%, with no correlation with find size or amount of radiogenic 146 helium, which implies that these samples have experienced little helium loss since separation from their parent body. An exception to this is ungrouped achondrite NWA 4518, which has a 3He age of 8.2 Ma and a 21Ne age of 13.7 Ma (40% difference), indicating some helium loss. Though the majority of the samples in this work have experienced little 3He loss, the discussion will be focused on the neon results.

Figure 7.8. 3He and 21Ne CRE ages show close agreement. Error bars are 15% of the measured value, blue squares are brachinites, pink circles are brachinite-like, and green squares are ungrouped ultra-mafic achondrites. Purple triangles are ages recalculated from literature data (Ott et al., 1988; 1993; Swindle et al., 1998; Patzer et al., 2003).

Literature data of six brachinites is included in addition to the sixteen UMA measured in this work, for a more thorough representation. Instead of using a histogram, we prefer a relative probability plot of the 21Ne CRE ages (Fig. 7.9), which includes the uncertainty. This is a useful way to look for a signal of a common break up event or grouping (see Beard et. al., 2017, for an example of using this approach on ureilite CRE groupings). Note that in a relative probability plot, each age is represented as a Gaussian distribution, all with the same area under the curve, so the 147 width is proportional to the uncertainty and more precisely defined ages have narrower and taller

Gaussians. To help further scrutinize the possible groupings, Figure 7.10 includes the Δ17O measurements from the literature.

1.2

1.0 21Ne CRE Age Relative Probability

0.8

i b

a 0.6

v 0.4

l e R 0.2

0.0 0 10 20 30 40 50 60 70 Exposure Age (Ma) Figure 7.9. Relative probability plot of 21Ne exposure ages. Each age is represented as a Gaussian distribution, all with the same area under the curve. The width is proportional to the uncertainty and more precisely defined ages have narrower and taller Gaussians. Note that the peak at ~4 is from only one sample and do not suggest a common event.

Suggested groupings are summarized in Table 7.7. From the distribution of ages in Fig. 7.9, there is an indication of up to five possible groupings based on age alone. However, the Δ17O values (Table 7.8, Fig. 7.10) suggest that not all of these samples actually come from the same parent body. The youngest possible grouping, Group 1 in Figure 7.9, is at ~9.4 ± 1.4 Ma (group ages are the average of the meteorites’ ages in the group, with the standard deviation used as the uncertainty unless otherwise stated) and includes two brachinite samples from literature (ALH

84025 and Reid 013; Ott et al., 1985 and Patzer et al., 2003 respectively) and one brachinite-like 148 sample from this work, NWA 1500. Of the three samples, oxygen values are known for only two, which agree in Δ17O: NWA 1500 = –0.235 ± 0.049 ‰, and ALH 84025 = –0.30 ± 0.02 ‰ (NWA

1500: Bartoschewitz et al., 2003; Goodrich et al. 2011; Kita et al., 2009. ALH 84025: Clayton and

Mayeda 1996; Greenwood et al. 2012). Using this additional oxygen data, it is reasonable to say that NWA 1500 and ALH 84025 could represent a common break up event at 10.1 ± 0.8 Ma based on CRE ages and Δ17O. Oxygen isotopic measurements of Reid 013 are needed to better define this possible group. If this group shared an impact event at ~10 Ma, the fact that NWA 1500 is brachinite-like suggests that perhaps some brachinite and brachinite-like samples come from the same parent body. Alternatively, this group could be the result of 2 or more independent impacts.

Figure 7.10. 21Ne CRE ages in this work plotted against Δ17O values from literature. Blue squares represent brachinites measured from this work, pink circles represent brachinite-like samples from this work, purple triangles represent brachinite ages based on literature measurements, and green diamonds represent ungrouped achondrites from this work. Outliers at the bottom are LEW 88763 and NWA 6962. Bartoschewitz et al., 2003; Goodrich et al. 2011; Kita et al., 2009. ALH 84025: Clayton and Mayeda 1996; Greenwood et al., 2012.

149

Table 7.7. Impact groupings for brachinite and brachinite-like samples. 21 Group Sample Δ17O Ne Age (Ma) Type Group 1 Reid 013* - 7.9 ± 1.0 Brachinite NWA 1500^ -0.235 ± 0.049 9.5 ± 0.3 Brachinite-like ALH 84025* -0.300 ± 0.020 10.7 ± 1.0 Brachinite Total -0.268 ± 0.046 10.1 ± 0.8 Mix Group 2 NWA 3151 -0.173 ± 0.032 22.1 ± 0.7 Brachinite NWA 4874 - 27.4 ± 1.1 Brachinite NWA 595 -0.154 ± 0.033 27.4 ± 4.7 Brachinite-like NWA 6077 -0.057 ± 0.054 26.6 ± 1.1 Brachinite-like NWA 7297^ - 27.9 ± 2.9 Brachinite-like Total -0.128 ± 0.062 26.3 ± 2.4 Mix Group 3 RaS 309 -0.190 ± 0.019 41.1 ± 3.2 Brachinite Hughes 026* -0.213 ± 0.039 39.3 ± 2.7 Brachinite Total -0.202 ± 0.017 40.2 ± 2.1 Brachinite Group 4 NWA 7605 -0.176 ± 0.018 48.0 ± 2.4 Brachinite-like Eagles Nest* -0.193 ± 0.015 48.3 ± 7.0 Brachinite EET 99402* -0.127 ± 0.018 48.7 ± 1.9 Brachinite NWA 4969 - 50.9 ± 2.5 Brachinite NWA 10637 -0.216 ± 0.078 53.1 ± 2.2 Brachinite-like Total -0.216 ± 0.079 50.4 ± 2.0 Mix Samples with ‘*’ have ages recalculated from literature values. A ‘^’ signifies the sample classification is not certain. Oxygen values are reported in ‰ and ages are in Ma. ‘Mix’ means the sample types in the group are both brachinite and brachinite -like. Note 3 of the 4 groups are mixed and no ungrouped achondrites are in these groupings. See references in Table 7.8.

At first glance it would appear there are eight samples (LEW 88763, NWA 3151, NWA 595,

NWA 8777, NWA 6077, NWA 4874, NWA 7297) that form one or two groups around 25 Ma,

“Group 2”. The CRE ages are bimodal, but it is difficult to define what sample belongs to which mode. The simplest approach is to consider all these samples to be from the same group, which would result in an age of ~25.9 ± 4.3 Ma. However, the age of LEW 88763 (Swindle et al., 1998) is highly suspect. The measured 21Ne/22Ne ratios of 0.710 ± 0.005 and 0.728 ± 0.004, respectively, 150 in two aliquots are unusual (possible effects of SCR, note small find size of ~4.1 g) and do not produce a reliable age using the shielding model (Leya and Masarik, 2009). The ungrouped achondrite NWA 8777 is unlikely to be part of this group despite having a similar age, due to a distinct mineralogy (Irving et al., 2015). Furthermore, the oxygen isotopes of both LEW 88763 and NWA 8777 (–1.26 ± 0.13 ‰ and –0.29 ± 0.04 ‰ Bartoschewitz et al., 2003 and Ziegler,

Meteoritical Bulletin Database 2018, respectively) do not agree with the other samples in this age group and likely come from a different parent body.

This leaves five samples that agree in age in Group 2; three brachinite-like samples, NWA

6077, NWA 595, and NWA 7297 Ma, and two brachinites NWA 3151 and NWA 4874. These may represent a shared impact at 26.3 ± 2.4 Ma. The Δ17O is known for the brachinite-like samples,

NWA 595 = –0.15 ± 0.03 ‰ (Irving et al., 2005; Irving and Rumble 2006, Gardner-Vandy et al.,

2013, Greenwood et al., 2012, Day et al., 2012), NWA 6077 = –0.057 ± 0.002 ‰ (Rumble,

Meteoritical Bulletin Database 2018) and for brachinite NWA 3151 = –0.17 ± 0.03 ‰ (Gardner-

Vandy et al., 2013; Greenwood et al., 2012; Rumble, Meteoritical Bulletin Database 2018). This demonstrates another case where brachinite-like and brachinite samples appear in the same age grouping, which could represent distinct impacts at the same time on different bodies, or brachinite and brachinite-like samples affected by the same impact on the same parent body. The Δ17O measurements of NWA 4874 and NWA 7297 are still needed.

Group 3 includes brachinites RaS 309 and NWA 6474, and Hughes from literature (Patzer et al., 2003) and suggest an event at ~38.8 ± 2.6 Ma. However, the oxygen isotopes may suggest a different parent body. The Δ17O of RaS 309, -0.19 ± 0.02, (Franchi and Greenwood, Meteoritical

Bulletin Database 2018) agrees with Hughes, Δ17O = -0.21 ± 0.04 (Clayton and Mayeda, 1996;

Gardner-Vandy et al., 2013; Greenwood et al., 2012), while Δ17O = -0.062 ± 0.013 for NWA 6474. 151

This difference of oxygen isotopes suggests a common impact event among all three samples is unlikely. The average CRE for the brachinites is 40.2 ± 2.1 Ma.

Group 4 is at ~50.4 ± 2.0 Ma and includes five samples, three from this work: brachinite NWA

4969, and brachinite-like NWA 7605 and NWA 10637, with two others from literature, Eagles

Nest and EET 99402 (Swindle et al., 1998 and Patzer et al., 2003 respectively). Four of these samples have oxygen data in the literature: Δ17O NWA 7605= -0.18 ± 0.02 ‰, NWA 10637 = -

0.21 ± 0.08 ‰, EET 99402 = -0.13 ± 0.03 ‰, and Eagles Nest = -0.19 ± 0.2 ‰ (Day et al., 2012;

Greenwood et al., 2012; Clayton and Mayeda, 1996; Gardner-Vandy et al., 2013; Zeigler,

Meteoritical Bulletin Database 2018). The similar CRE ages and Δ17O data are a strong indication that these five brachinites represent samples from a common parent body. Including or excluding

NWA 4969 does not significantly affect the average age, though its oxygen isotopes still need to be measured.

Figure 7.11. Fe/Mg of brachinite (blue) and brachinite-like achondrites (pink) in relation to 21Ne CRE ages. Brachinites are mostly in a narrow range of Fe/Mg (~0.5 to 0.55) compared to brachinite-like samples. Brachinite-like samples show tighter clusters of ages than brachinites although there is a small sample size (n =6)

152

There are four samples from this work that do not fall into any apparent group (ungrouped olivine-rich achondrites NWA 4518 = 13.7 ± 1.0 Ma, NWA 6962 = 16.25 ± 1.1, and brachinites

NWA 4876 = 63.1 ± 1.8 Ma, and NWA 4882 = 66.5 ± 2.5 Ma). NWA 4874 has been suspected to be paired with NWA 4876 (Tony Irving, personal communication), but their CRE ages demonstrate that this is not the case. On the other hand, NWA 4876 is very petrologically similar to NWA 4882 (Goodrich et al., 2006; Wittke and Bunch, Meteoritical Bulletin Database 2018;

Rumble et al., 2008), and they have the same age, so they might be paired.

The Δ17O values of the meteorites in the two older (~26 Ma, ~50 Ma) clusters are similar, while the members of the younger cluster (~10 Ma) have slightly more depleted Δ 17O values, and might suggest a different parent body or reflect oxygen heterogeneity within the same source. Dunlap et al., (2018) show a ~15 Ma range in the Mn-Cr ages of two brachinites (Brachina and NWA 4882) which suggests an extended thermal history. Such a prolonged thermal history could be questionable as it would imply a rather large parent body comparable to 4 Vesta.

153

Table 7.8. 21Ne (Ma), Δ17O (‰), and classification. 17 21 Sample Δ O Ne Age (Ma) Classification Reference NWA 3151 -0.173 ± 0.032 22.1 ± 0.7 Brachinite 1,2,3 NWA 4874 - 27.4 ± 1.1 Brachinite - NWA 4876 - 63.1 ± 1.8 Brachinite - NWA 4882 -0.241 ± 0.010 66.5 ± 2.5 Brachinite 6

NWA 4969 - 50.9 ± 2.5 Brachinite - NWA 6474^ -0.062 ± 0.013 36.1 ± 2.9 Brachinite 1 RaS 309 -0.190 ± 0.019 41.1 ± 3.2 Brachinite 4 NWA 1500^ -0.235 ± 0.049 9.5 ± 0.3 Brachinite-like 10, 11, 12 NWA 595 -0.154 ± 0.033 27.4 ± 4.7 Brachinite-like 3, 7, 8, 9 NWA 6077 -0.057 ± 0.054 26.6 ± 1.1 Brachinite-like 1 NWA 7297^ - 27.9 ± 2.9 Brachinite-like -

NWA 7605 -0.176 ± 0.018 48.0 ± 2.4 Brachinite-like 5 NWA 10637 -0.216 ± 0.078 53.1 ± 2.2 Brachinite-like 5 NWA 4518 -0.209 ± -0.029 13.7 ± 1.0 Ungrouped 4 NWA 6962 -1.037 ± 0.006 16.3 ± 1.1 Ungrouped 13 NWA 8777 -0.291 ± 0.040 23.1 ± 1.0 Ungrouped 5

Lew 88763* -1.260 ± 0.126 22.4 ± 3.4 Ungrouped 10, 17 ALH 84025* -0.300 ± 0.020 10.7 ± 1.0 Brachinite 3, 14, 15, 18 Brachina* -0.318 ± 0.078 4.2 ± 0.5 Brachinite 3, 8, 15, 19 Eagles Nest* -0.193 ± 0.015 48.3 ± 7.0 Brachinite 3, 15, 17 EET 99402* -0.127 ± 0.018 48.7 ± 1.9 Brachinite 3, 9, 16, 20 Reid 013* - 7.9 ± 1.0 Brachinite 20 Hughes 026* -0.213 ± 0.039 39.3 ± 2.7 Brachinite 2, 3, 15, 20

‘^’ indicates classification is suspect, NWA 1500 has reverse-zoned rims and the Fe/Mg span a wide range. ‘*’ indicates sample ages were recalculated using literature values. References are for the oxygen isotopes. 1. Rumble, Meteoritical Bulletin Database 2018. 2. Gardner-vandy et al., 2013. 3. Greenwood et al., 2012. 4. Franchi and Greenwood, Meteoritical Bulletin Database 2018. 5. Ziegler, Meteoritical Bulletin Database 2018. 6.) Rumble et al., 2008. 7.) Irving et al., 2005. 8.) Irving and Rumble III 2006. 9.) Day et al., 2012. 10.) Bartoschewitz et al., 2003. 11.) Goodrich et al., 2011. 12.) Kita et al., 2009. 13.) Tanaka, Meteoritical Bulletin Database 2018. 14.) Clayton, in Warren and Kalleymeyn (1989). 15.) Clayton and Mayeda 1996. 16.) Mayeda and Clayton, Meteroit. Bull. Database 2018. 17.) Swindle et al., 1998. 18.) Ott et al., 1993. 19.) Ott et al., 1985. 20.) Patzer et al., 2003.

154

7.7 Conclusions

The measured 3He and 21Ne are purely cosmogenic, while the 38Ar has a significant trapped component. The 38Ar exposure age is very sensitive to the amount of calcium in the sample, and

Ca is contained in minor phases. Brachinites’ 21Ne CRE ages range from ~4 to 65 Ma, while the brachinite-like samples range from ~10 to 50 Ma. NWA 1500 (9.5 Ma) and NWA 7605 (50.1 Ma), are the only brachinite/brachinite-like samples that have the presence of reduced rims, are clearly from different source events (CRE age), but may be from the same parent body (similar petrologic characteristics and Δ17O values). Combining literature data of seven brachinites with sixteen additional samples measured in this work, the brachinites and and brachinite-like achondrites show evidence of three possible groupings at ~10 Ma, ~26 Ma, and ~50 Ma. Literature data on brachinites shows evidence of three impacts on the parent body, which is expanded to at least six impacts by this work, four of which were previously unknown.

There is at least one brachinite-like sample in each group, and no brachinite-like samples that indicate an age outside of these groups. NWA 1500 and ALH 84025, one a brachinite and one a brachinite-like achondrite, have the same 21Ne age and Δ17O. Another possible mixing of the two classifications is with brachinites NWA 3151 and NWA 4874 with brachinite-like NWA 6077,

NWA 7297, and NWA 595 at ~26 Ma. No Δ17O measurements have been made on the brachinites in this group and therefore no comparison between them can be made. Finally, brachinite-like

NWA 7605 and NWA 10637 share approximately the same 21Ne age and Δ17O values as the brachinites Eagles Nest, EET 99402, and NWA 4969. The interesting point is that there are two cases where brachinite and brachinite-like achondrites agree in 21Ne age and Δ17O values, and an additional case where the ages agree but no oxygen measurements can be compared. This does not definitively confirm that they form a grouping event, i.e. that they were liberated from the same 155 impact on the same parent body at the same time, but it is worth consideration that the brachinite and brachinite-like achondrites originate from the same parent body. Future models of the origin of the brachinite parent body should try to explain the formation of both of these relatively oxidized/reduced groups.

156

Chapter 8: Conclusions

Chelyabinsk has been the focus of many geochronology studies that have resulted in several different ages. One of the goals of this work was to use Ar-Ar dating to better understand the impact history of Chelyabinsk. Seven measurements on the two main lithologies of Chelyabinsk show well defined impact ages ~30 Ma and ~2700 Ma that I then used to help put other results into better context. Chelyabinsk experienced 3-4 impacts after formation at ~4450 Ma, ~2700 (this work) Ma, ~30 Ma (this work), before becoming exposed to cosmic rays ~1.5 Ma. An impact at

~2700 Ma has not been observed in other LL chondrites and is similar to Ar-Ar ages of Itokawa, which has been hypothesized to have a closely shared relationship with Chelyabinsk. The light lithology of Chelyabinsk records an impact event at ~30 Ma, which is significantly younger than ages measured in other LL chondrites. Most of the discrepancies in age measurements have been accounted for, however, the U-Pb ages of zircon and apatite behaving opposite to what would be expected based on closure temperatures still needs a resolution.

Another goal of this work was to evaluate the extreme heterogeneity in oxygen isotopes among the ureilites by comparing them to samples that have similar CRE ages. A shared CRE age does not necessarily mean that the samples shared the same impact, but if a cluster of ages was discovered it would provide a starting point to discuss the extent of heterogeneity between samples that might have shared the same history However, no strong indication of common CRE ages was found among the ureilites.

The last goal of this work was to use Ar-Ar and CRE ages to evaluate the likelihood of a genetic relationship among the brachinites, brachinite-like, and other ungrouped achondrites. I also developed a better-defined classification for determining what is or is not a brachinite. My criteria of a brachinite are broadly similar to what others have done, but use defined parameters, which 157 can be expanded upon, provides for an opportunity to minimize the ambiguity. The Ar-Ar portion of this work was met with difficulties and has been delayed. However, the CRE ages have shown interesting results. Brachinite-like samples, as I have defined them, appear in more tightly packed age clusters than the brachinites. There are four clusters of ages among all the UMA samples in this work, three of the four show a mix of brachinite and brachinite-like samples, which suggest a possible shared history. The extent of their relationship is still unknown, but further chronology data combined with well-defined classifications of samples might provide a clearer picture in the future.

Studying the brachinite and brachinite-like achondrites in detail has shown me that these samples need to be more studied. For example, I was not able to evaluate helium and argon retention ages because there have been too few, if any bulk elemental measurements on these samples. The wide variation in the data from the few abundance measurements makes them difficult to be useful. Oxygen isotopic measurements are missing for key samples that are important for determining age groupings. There have only been two Mn-Cr ages of brachinites, which, if from the same parent body, show a significantly long cooling period that might imply a large parent body. Further Mn-Cr ages would be useful for comparing the early history of these samples. Combining Ar-Ar with other chronometers can be a powerful tool for understanding the history of a sample and its parent body. The CRE ages of this work suggest three shared events between brachinite and brachinite-like achondrites, at 10, 26, and 50 Ma, and increases the total number of samples studied from seven to twenty-three. It will be very interesting to see how other chronometers compare among these groups.

158

Supplementary Materials

Figure S.5.1. Left) Ar-Ar plateau plot for SB4. Below) Ar-Ar plateau plot for CH2. Each split from this sample (SB4, SB5, and CH2) have a partial plateau age of ~2700 Ma.

159

Figure S.5.2. Isochrons for Chelyabinsk not shown in the main text- either show different temperature ranges or different isochron type (standard or reverse).

400-600 °C

160

Figure S.5.3. Combined MB,2 isochron. SB1-SB3 are found in a similar range of 3-isotope space while CH1 (light blue) is clearly different. 161

Sample Mass (mg) J Factor Split Table S.5.1. Measured argon data and ages for each temperature. MB020f,2 10.3 0.001055785 CH1 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 250 797.0 ± 11.5 0.930 ± 0.025 0.615 ± 0.028 3.336 ± 0.365 2.536 ± 0.116 8.581 109.307 2 300 154.8 ± 5.3 1.417 ± 0.018 0.191 ± 0.022 2.732 ± 0.347 0.489 ± 0.062 21.307 33.008 3 350 109.1 ± 5.1 2.661 ± 0.013 0.142 ± 0.022 3.195 ± 0.332 0.515 ± 0.062 * * 4 375 37.6 ± 4.6 2.312 ± 0.022 0.087 ± 0.022 2.917 ± 0.325 0.221 ± 0.059 19.816 6.728 5 400 35.1 ± 4.5 2.593 ± 0.016 0.050 ± 0.022 2.681 ± 0.324 0.142 ± 0.058 18.917 5.174 6 425 31.5 ± 4.5 2.651 ± 0.015 0.094 ± 0.020 2.914 ± 0.323 0.155 ± 0.057 16.519 5.001 7 450 34.2 ± 4.4 2.670 ± 0.025 0.096 ± 0.021 2.819 ± 0.318 0.132 ± 0.056 19.499 4.696 8 475 37.1 ± 4.4 2.599 ± 0.026 0.091 ± 0.020 2.673 ± 0.319 0.102 ± 0.056 23.601 4.599 9 500 146.3 ± 4.4 2.410 ± 0.020 0.128 ± 0.022 2.786 ± 0.329 0.234 ± 0.055 102.750 5.868 10 525 54.6 ± 4.3 2.215 ± 0.025 0.091 ± 0.021 2.646 ± 0.323 0.134 ± 0.055 40.627 5.550 11 550 47.8 ± 4.3 2.077 ± 0.022 0.068 ± 0.020 2.252 ± 0.312 0.058 ± 0.055 41.303 5.456 12 575 50.0 ± 4.3 1.598 ± 0.013 0.059 ± 0.020 2.073 ± 0.312 0.028 ± 0.054 58.481 6.888 13 600 55.3 ± 4.3 1.421 ± 0.010 0.028 ± 0.020 1.895 ± 0.309 0.022 ± 0.054 71.219 7.690 14 625 65.6 ± 4.3 1.169 ± 0.013 0.046 ± 0.020 1.698 ± 0.307 0.081 ± 0.055 97.042 9.658 15 650 79.1 ± 4.3 0.997 ± 0.007 0.062 ± 0.019 1.551 ± 0.308 0.046 ± 0.054 143.291 10.582 16 700 156.1 ± 4.3 1.096 ± 0.008 0.093 ± 0.020 2.257 ± 0.311 0.138 ± 0.056 243.067 10.156 17 750 255.1 ± 4.4 0.839 ± 0.013 0.097 ± 0.020 3.238 ± 0.316 0.304 ± 0.056 471.402 17.678 18 800 392.9 ± 4.6 0.607 ± 0.012 0.137 ± 0.021 3.848 ± 0.331 0.245 ± 0.056 919.442 21.004 19 850 462.1 ± 4.7 0.460 ± 0.012 0.167 ± 0.021 5.016 ± 0.342 0.375 ± 0.063 1269.379 31.765 20 900 421.0 ± 4.6 0.333 ± 0.009 0.194 ± 0.021 6.347 ± 0.340 0.371 ± 0.057 1497.591 38.390 21 950 378.4 ± 4.6 0.290 ± 0.008 0.160 ± 0.021 8.425 ± 0.369 0.321 ± 0.059 1536.455 40.365 22 975 302.4 ± 4.6 0.187 ± 0.008 0.128 ± 0.021 8.814 ± 0.349 0.315 ± 0.060 1766.216 64.347 23 1000 355.0 ± 4.7 0.167 ± 0.007 0.171 ± 0.022 12.980 ± 0.359 0.393 ± 0.061 2109.976 64.404 24 1025 330.6 ± 4.4 0.124 ± 0.007 0.210 ± 0.025 18.615 ± 0.506 0.429 ± 0.059 2460.403 92.498 25 1050 196.6 ± 4.3 0.091 ± 0.006 0.195 ± 0.025 19.533 ± 0.215 0.410 ± 0.058 2198.869 125.185 26 1075 150.5 ± 4.2 0.061 ± 0.006 0.215 ± 0.025 20.046 ± 0.470 0.448 ± 0.056 2422.393 196.606 27 1100 142.5 ± 4.1 0.050 ± 0.004 0.242 ± 0.023 15.175 ± 0.314 0.640 ± 0.066 2427.195 238.991 28 1150 947.4 ± 4.2 0.094 ± 0.006 0.979 ± 0.025 21.098 ± 0.329 4.515 ± 0.088 4099.226 295.639 29 1200 797.2 ± 4.3 0.095 ± 0.008 1.223 ± 0.031 27.707 ± 0.438 5.890 ± 0.094 3459.858 556.671 30 1300 460.9 ± 16.7 0.130 ± 0.011 1.477 ± 0.020 28.452 ± 0.282 6.629 ± 0.095 606.470 1995.721 31 1400 89.2 ± 16.2 0.056 ± 0.009 0.378 ± 0.019 3.785 ± 0.129 1.610 ± 0.087 * * 32 1500 59.2 ± 17.5 0.017 ± 0.012 0.108 ± 0.020 -0.579 ± 0.145 0.482 ± 0.092 1793.910 1148.633 Total 259.5 12.6 *Null Value. All values are in 1e-8 ccSTP/g

162

Table S.5.2. Measured argon data and ages for each temperature. Sample Mass (mg) J Factor Split MB020f,2 10.3 1.062E-03 SB2 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 300 1881.5 ± 10.2 1.720 ± 0.018 0.993 ± 0.014 1.773 ± 0.068 4.730 ± 0.077 * * 2 400 637.5 ± 9.7 3.629 ± 0.024 0.367 ± 0.007 3.615 ± 0.071 1.281 ± 0.064 75.0 20.0 3 500 347.1 ± 9.3 5.798 ± 0.032 0.336 ± 0.010 6.929 ± 0.171 0.930 ± 0.066 12.3 14.7 4 600 137.3 ± 9.2 3.815 ± 0.021 0.164 ± 0.006 5.291 ± 0.148 0.202 ± 0.059 43.0 12.9 5 700 188.9 ± 9.1 1.799 ± 0.021 0.099 ± 0.005 3.086 ± 0.099 0.144 ± 0.058 154.0 25.5 6 800 400.2 ± 9.3 0.651 ± 0.009 0.091 ± 0.005 2.413 ± 0.107 0.146 ± 0.060 894.3 21.8 7 900 427.4 ± 9.6 0.326 ± 0.009 0.142 ± 0.005 5.605 ± 0.086 0.275 ± 0.061 1547.3 40.2 8 925 163.9 ± 9.7 0.102 ± 0.008 0.097 ± 0.007 3.823 ± 0.111 0.138 ± 0.062 1785.3 117.6 9 950 167.7 ± 9.7 0.095 ± 0.008 0.096 ± 0.006 4.787 ± 0.094 0.121 ± 0.062 1910.0 128.3 10 975 214.8 ± 9.9 0.097 ± 0.008 0.125 ± 0.007 5.758 ± 0.199 0.218 ± 0.063 2166.0 133.4 11 1000 223.3 ± 10.0 0.083 ± 0.008 0.165 ± 0.007 8.290 ± 0.233 0.284 ± 0.065 2437.6 153.6 12 1025 159.3 ± 5.8 0.051 ± 0.009 0.177 ± 0.006 10.339 ± 0.226 0.336 ± 0.044 2665.1 285.1 13 1050 84.1 ± 5.6 0.032 ± 0.009 0.153 ± 0.006 9.671 ± 0.209 0.269 ± 0.042 2498.9 478.9 14 1075 70.5 ± 5.5 0.029 ± 0.009 0.145 ± 0.005 13.649 ± 0.313 0.292 ± 0.043 2491.8 647.2 15 1100 51.0 ± 5.4 0.025 ± 0.008 0.098 ± 0.003 7.452 ± 0.175 0.258 ± 0.042 1941.3 549.5 16 1200 178.9 ± 5.6 0.067 ± 0.009 0.448 ± 0.008 19.023 ± 0.522 1.156 ± 0.063 2083.2 215.8 17 1300 328.4 ± 13.8 0.072 ± 0.006 0.464 ± 0.010 9.914 ± 0.244 1.623 ± 0.097 2770.7 170.4 18 1400 73.8 ± 13.4 0.016 ± 0.007 0.042 ± 0.005 0.684 ± 0.038 0.213 ± 0.086 2931.6 727.9 19 1500 39.4 ± 14.5 0.014 ± 0.007 0.027 ± 0.007 0.347 ± 0.023 0.073 ± 0.093 2375.9 883.2 17 260.7 11.8 *Null Value. All values are in 1e-8 ccSTP/g

Table S.5.3. Measured argon data and ages for each temperature. Sample Mass (mg) J Factor Split MB020f,2 13.47 1.062E-03 SB3 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 250 2011.1 ± 2.7 1.308 ± 0.012 1.103 ± 0.029 2.136 ± 0.024 5.475 ± 0.033 * * 2 300 587.3 ± 2.0 0.983 ± 0.011 0.303 ± 0.026 1.192 ± 0.024 1.334 ± 0.019 57.1 77.5 3 350 418.6 ± 1.5 1.535 ± 0.013 0.233 ± 0.024 1.683 ± 0.058 0.920 ± 0.032 47.0 38.7 4 400 231.8 ± 0.8 1.947 ± 0.015 0.147 ± 0.024 2.138 ± 0.044 0.393 ± 0.015 76.8 14.1 5 450 164.4 ± 2.2 2.355 ± 0.015 0.135 ± 0.023 2.220 ± 0.070 0.255 ± 0.023 59.2 11.9 6 500 230.8 ± 0.9 2.338 ± 0.020 0.195 ± 0.023 2.300 ± 0.050 0.397 ± 0.017 72.6 11.6 7 550 104.5 ± 0.6 1.797 ± 0.015 0.094 ± 0.022 2.315 ± 0.050 0.207 ± 0.015 97.1 5.8 8 600 93.0 ± 0.6 1.365 ± 0.012 0.071 ± 0.021 1.819 ± 0.028 0.207 ± 0.008 110.6 8.0 9 650 103.8 ± 0.6 1.072 ± 0.012 0.075 ± 0.021 1.789 ± 0.034 0.090 ± 0.005 171.1 3.9 10 700 116.9 ± 0.2 0.666 ± 0.009 0.076 ± 0.021 1.652 ± 0.032 0.155 ± 0.012 290.1 10.6 11 750 153.5 ± 0.4 0.384 ± 0.009 0.067 ± 0.021 1.587 ± 0.031 0.160 ± 0.010 609.8 19.5 12 800 188.3 ± 0.5 0.251 ± 0.006 0.061 ± 0.021 2.127 ± 0.066 0.183 ± 0.008 1018.8 30.3 13 1000 411.3 ± 0.6 0.302 ± 0.008 0.242 ± 0.023 11.953 ± 0.147 0.511 ± 0.014 1577.1 44.1 14 1200 526.5 ± 7.4 0.326 ± 0.007 1.045 ± 0.019 61.695 ± 0.367 3.484 ± 0.066 1465.9 287.0 15 1400 296.7 ± 7.8 0.160 ± 0.004 0.905 ± 0.025 22.825 ± 0.191 3.783 ± 0.058 822.6 915.7 16 1500 37.7 ± 8.5 0.016 ± 0.005 0.007 ± 0.025 0.117 ± 0.030 0.044 ± 0.059 2136.3 487.3 Total 257.9 12.6 *Null Value. All values are in 1e-8 ccSTP/g

163

Table S.5.4. Measured argon data and ages for each temperature. Sample Split Mass (mg) J Factor MB020f,5 SB4 9.32 1.06E-03 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 400 820.3 ± 6.2 1.210 ± 0.018 0.338 ± 0.018 1.335 ± 0.046 1.762 ± 0.061 422.4 30.1 2 500 596.2 ± 5.2 1.597 ± 0.016 0.212 ± 0.015 1.849 ± 0.055 0.588 ± 0.053 463.5 18.2 3 600 626.3 ± 5.0 1.484 ± 0.015 0.134 ± 0.015 2.582 ± 0.057 0.405 ± 0.041 566.4 15.5 4 700 958.4 ± 5.0 1.137 ± 0.020 0.220 ± 0.015 2.332 ± 0.057 0.885 ± 0.048 913.7 22.4 5 800 887.7 ± 5.1 0.770 ± 0.015 0.205 ± 0.016 2.491 ± 0.058 0.716 ± 0.042 1207.2 26.7 6 900 1242.9 ± 5.3 0.714 ± 0.013 0.224 ± 0.015 4.159 ± 0.101 0.769 ± 0.077 1822.8 37.3 7 925 1790.0 ± 5.4 0.695 ± 0.013 0.280 ± 0.016 4.581 ± 0.086 1.018 ± 0.048 2308.3 39.2 8 950 2384.2 ± 6.2 0.741 ± 0.013 0.300 ± 0.016 5.475 ± 0.067 1.051 ± 0.041 2624.8 34.9 9 975 2228.2 ± 5.8 0.665 ± 0.015 0.293 ± 0.015 6.924 ± 0.143 1.181 ± 0.042 2670.7 44.5 10 1000 2293.9 ± 5.9 0.685 ± 0.014 0.396 ± 0.016 7.354 ± 0.125 1.358 ± 0.077 2663.6 44.3 11 1025 2090.8 ± 4.1 0.608 ± 0.015 0.328 ± 0.024 7.114 ± 0.105 1.154 ± 0.080 2705.9 46.6 12 1050 1625.1 ± 3.8 0.441 ± 0.013 0.286 ± 0.023 7.127 ± 0.177 1.275 ± 0.081 2771.9 63.9 13 1075 989.7 ± 3.6 0.232 ± 0.011 0.178 ± 0.022 5.489 ± 0.118 0.808 ± 0.084 2980.6 86.9 14 1100 465.4 ± 3.5 0.118 ± 0.008 0.148 ± 0.022 3.671 ± 0.070 0.698 ± 0.066 2767.6 241.6 15 1200 1227.4 ± 3.6 0.356 ± 0.012 0.336 ± 0.023 5.427 ± 0.093 1.238 ± 0.075 2650.5 139.4 16 1300 669.0 ± 22.6 0.177 ± 0.012 0.292 ± 0.022 4.576 ± 0.323 1.340 ± 0.084 2624.2 323.8 17 1400 159.4 ± 21.9 0.056 ± 0.010 0.074 ± 0.019 1.740 ± 0.313 0.327 ± 0.076 2253.0 443.7 18 1500 173.4 ± 23.8 0.070 ± 0.011 0.032 ± 0.020 1.491 ± 0.334 0.133 ± 0.080 2241.7 298.9 Total 1794.7 15.8 *Null Value. All isotope values are in 1e-8 ccSTP/g and ages are in Ma.

164

Table S.5.5. Measured argon data and ages for each temperature. Sample Split Mass (mg) J Factor MB020f,5 SB5 9.96 1.06E-03 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 300 4536.8 ± 6.4 0.773 ± 0.015 2.709 ± 0.032 0.238 ± 0.066 14.636 ± 0.159 329.8 190.3 2 400 744.4 ± 6.8 0.830 ± 0.017 0.330 ± 0.015 0.619 ± 0.064 1.373 ± 0.041 661.7 30.2 3 450 472.9 ± 4.1 0.665 ± 0.015 0.086 ± 0.012 0.721 ± 0.064 0.360 ± 0.034 831.6 27.8 4 500 654.9 ± 4.0 0.754 ± 0.013 0.189 ± 0.013 0.884 ± 0.063 0.815 ± 0.049 832.5 31.1 5 600 954.3 ± 3.9 1.105 ± 0.015 0.134 ± 0.012 1.764 ± 0.066 0.427 ± 0.034 1062.9 16.8 6 700 1314.6 ± 4.2 1.151 ± 0.014 0.182 ± 0.015 2.611 ± 0.067 0.438 ± 0.034 1346.1 16.6 7 750 948.4 ± 4.0 0.630 ± 0.012 0.168 ± 0.013 2.077 ± 0.072 0.669 ± 0.038 1482.4 28.2 8 800 979.6 ± 4.2 0.542 ± 0.012 0.106 ± 0.011 2.313 ± 0.068 0.456 ± 0.037 1763.8 31.4 9 850 906.9 ± 4.1 0.391 ± 0.010 0.100 ± 0.012 2.493 ± 0.075 0.236 ± 0.032 2156.1 38.1 10 900 1055.2 ± 4.2 0.474 ± 0.011 0.151 ± 0.012 3.266 ± 0.077 0.301 ± 0.032 2105.9 34.4 11 950 2695.3 ± 4.9 0.829 ± 0.013 0.197 ± 0.015 6.389 ± 0.120 0.618 ± 0.032 2528.6 63.5 12 1000 5293.0 ± 10.8 1.466 ± 0.017 0.416 ± 0.014 12.884 ± 0.230 1.179 ± 0.044 2683.4 59.3 13 1050 7241.2 ± 5.9 1.879 ± 0.026 0.645 ± 0.016 19.280 ± 0.194 2.233 ± 0.055 2696.8 88.7 14 1100 979.6 ± 4.2 0.542 ± 0.012 0.106 ± 0.011 2.313 ± 0.068 0.456 ± 0.037 1763.8 31.4 15 1125 906.9 ± 4.1 0.391 ± 0.010 0.100 ± 0.012 2.493 ± 0.075 0.236 ± 0.032 2156.1 38.1 16 1150 1055.2 ± 4.2 0.474 ± 0.011 0.151 ± 0.012 3.266 ± 0.077 0.301 ± 0.032 2105.9 34.4 17 1175 2695.3 ± 4.9 0.829 ± 0.013 0.197 ± 0.015 6.389 ± 0.120 0.618 ± 0.032 2528.6 63.5 18 1200 1075.1 ± 3.6 0.201 ± 0.014 0.204 ± 0.013 4.778 ± 0.072 0.735 ± 0.041 2802.8 278.1 19 1225 683.4 ± 3.6 0.132 ± 0.014 0.182 ± 0.012 2.039 ± 0.069 0.635 ± 0.035 2453.9 433.2 20 1250 450.5 ± 3.7 0.077 ± 0.013 0.189 ± 0.010 1.028 ± 0.060 0.678 ± 0.042 1571.7 135.0 21 1300 483.6 ± 17.4 0.085 ± 0.014 0.193 ± 0.017 0.831 ± 0.093 1.304 ± 0.082 2446.0 432.1 22 1350 117.5 ± 16.8 0.043 ± 0.013 0.041 ± 0.015 0.264 ± 0.086 0.195 ± 0.051 1637.4 545.3 23 1400 124.2 ± 16.8 0.047 ± 0.014 0.038 ± 0.015 0.318 ± 0.086 0.197 ± 0.045 1641.7 492.1 24 1500 76.8 ± 18.2 0.014 ± 0.014 0.039 ± 0.014 0.348 ± 0.093 0.150 ± 0.049 2315.3 1640.1 Total 2062.7 29.1 *Null Value. All isotope values are in 1e-8 ccSTP/g and ages are in Ma.

165

Table S.5.6. Measured argon data and ages for each temperature. Sample Split Mass (mg) J Factor MB020f,5 CH2 10.30 1.06E-03 Step # Temp (°C) 40Ar 39Ar 38Ar 37Ar 36Ar Apparent Age ± 2σ 1 300 1696.8 ± 6.9 0.275 ± 0.018 1.084 ± 0.032 0.933 ± 0.198 7.325 ± 0.305 * * 2 400 546.3 ± 6.3 0.722 ± 0.017 0.197 ± 0.026 0.271 ± 0.172 1.057 ± 0.050 531.4 39.9 3 450 406.6 ± 6.2 0.674 ± 0.018 0.037 ± 0.024 0.337 ± 0.169 0.356 ± 0.040 697.5 32.5 4 500 532.6 ± 6.1 0.850 ± 0.020 0.097 ± 0.024 0.679 ± 0.168 0.558 ± 0.048 678.4 30.9 5 550 485.9 ± 6.0 0.834 ± 0.022 0.035 ± 0.024 0.681 ± 0.167 0.123 ± 0.039 813.4 28.6 6 600 509.2 ± 6.0 0.804 ± 0.018 0.049 ± 0.023 0.822 ± 0.167 0.120 ± 0.037 908.3 21.7 7 650 644.5 ± 6.0 0.745 ± 0.021 0.056 ± 0.024 0.852 ± 0.165 0.198 ± 0.037 1144.0 30.4 8 700 764.5 ± 6.0 0.750 ± 0.020 0.079 ± 0.023 1.014 ± 0.166 0.159 ± 0.038 1301.7 29.0 9 750 863.9 ± 6.1 0.654 ± 0.017 0.057 ± 0.023 1.155 ± 0.166 0.231 ± 0.039 1546.6 33.7 10 800 765.8 ± 6.1 0.519 ± 0.017 0.092 ± 0.023 1.263 ± 0.170 0.277 ± 0.039 1658.6 43.2 11 850 895.5 ± 6.2 0.499 ± 0.020 0.094 ± 0.025 1.727 ± 0.176 0.394 ± 0.045 1866.5 56.7 12 900 849.2 ± 6.3 0.432 ± 0.015 0.059 ± 0.025 1.992 ± 0.177 0.254 ± 0.040 1991.6 49.4 13 925 825.0 ± 6.4 0.362 ± 0.015 0.029 ± 0.024 1.809 ± 0.179 0.152 ± 0.040 2188.0 56.9 14 950 1049.1 ± 6.4 0.357 ± 0.016 0.067 ± 0.025 2.001 ± 0.183 0.150 ± 0.041 2534.8 65.3 15 975 1369.2 ± 6.7 0.414 ± 0.018 0.084 ± 0.025 2.614 ± 0.186 0.234 ± 0.041 2692.5 63.1 16 1000 1610.0 ± 6.7 0.469 ± 0.018 0.065 ± 0.026 3.141 ± 0.187 0.281 ± 0.045 2742.7 57.4 17 1025 2574.4 ± 9.2 0.781 ± 0.024 0.158 ± 0.012 5.221 ± 0.186 0.569 ± 0.048 2680.4 48.4 18 1050 3085.8 ± 9.1 0.897 ± 0.021 0.187 ± 0.014 6.654 ± 0.199 0.662 ± 0.053 2741.9 39.1 19 1075 3085.8 ± 9.1 0.897 ± 0.021 0.187 ± 0.014 6.654 ± 0.199 0.662 ± 0.053 2741.9 39.1 20 1100 4302.8 ± 9.1 1.240 ± 0.026 0.314 ± 0.017 10.502 ± 0.234 1.168 ± 0.052 2746.7 38.3 21 1150 3058.0 ± 9.0 0.801 ± 0.020 0.269 ± 0.015 8.311 ± 0.189 1.008 ± 0.060 2875.4 46.8 22 1200 4315.8 ± 9.4 1.109 ± 0.023 0.812 ± 0.019 12.606 ± 0.194 4.296 ± 0.094 2801.5 99.1 23 1300 13342.9 ± 14.6 3.845 ± 0.039 1.992 ± 0.035 45.605 ± 0.552 10.465 ± 0.164 2672.6 72.9 24 1400 3358.2 ± 10.6 0.977 ± 0.023 0.676 ± 0.021 9.495 ± 0.212 2.832 ± 0.083 2653.0 81.4 25 1500 288.2 ± 12.1 0.133 ± 0.016 0.053 ± 0.017 0.446 ± 0.205 0.470 ± 0.048 1931.6 210.7 Total 2217.3 22.9 *Null Value. All isotope values are in 1e-8 ccSTP/g and ages are in Ma.

166

Supplementary data for Chapter 7.

Figure S.7.1. Xe isotopes show a mixture of phase Q and terrestrial air.

Blue squares are brachinites, pink circles are brachinite-like. Gold stars indicate standards, Ott, 2002. 167

129 129 Figure S.7.2. Some samples show excess Xe (derived from I), suggesting these samples were not heated sufficiently to lose Xe and the system was ‘closed’ to Xe early in the solar system. Blue squares are brachinites, pink circles are brachinite-like, and green diamonds are ungrouped. Gold stars indicate standards, Ott, 2002.

84Kr/132 vs 36Ar/132Xe 250 Air 200

150

3 1

/ 100

6 3 Q50

0 0 2 4 6 8 -50 84Kr/132 Xe

Figure S.7.3. 36Ar/132Xe of brachinites (blue), brachinite-like (orange), and ungrouped achondrites (grey) shows an Ar-rich trapped component ~90-150 which is in the range of literature values (Patzer et al., 2003). Both 36Ar/132Xe and 84Kr/132Xe cover the range of ordinary chondrites, ureilites, enstatite chondrites, Mars, and Earth (see e.g. Swindle et al., 1998); they are not diagnostic. 168

NWA 6962

NWA 8777

NWA 10637

4 Figure S.7.4. Measured Herad in brachinite and UMA samples is very low, making it difficult to determine radiogenic ages. Non-zero values are labeled with the appropriate sample name.

Figure S.7.5. Zoomed in version of S.7.4- showing several values ~0.

169

Figure S.7.6. Ne-3 Isotope plot showing measured values.

Figure S.7.7. Ar vs Ne CRE ages for brachinite, brachinite-like, and ungrouped achondrites. Clearly these do not agree as well as He and Ne (see Chapter 7).

170

Table S.7.1. Xe isotopes for UMA.

Sample 132Xe 129Xe/132Xe 84Kr/132Xe 129Xe 1/132Xe NWA 10637 10.8 ± 0.18 102.88 ± 0.99 3.6 ± 0.1 11.1 ± 0.21 0.093 ± 0.002 NWA 1500_1 4.1 ± 0.06 104.04 ± 0.58 1.8 ± 0.0 4.3 ± 0.07 0.241 ± 0.004 NWA 1500_2 4.7 ± 0.09 104.14 ± 1.83 1.5 ± 0.1 4.9 ± 0.13 0.213 ± 0.004 NWA 3151_1 5.8 ± 0.09 104.45 ± 0.87 3.1 ± 0.1 6.0 ± 0.11 0.174 ± 0.003 NWA 3151_2 5.6 ± 0.09 101.06 ± 0.91 3.1 ± 0.1 5.6 ± 0.11 0.180 ± 0.003 NWA 4518_1 3.7 ± 0.06 132.99 ± 1.01 4.3 ± 0.1 4.9 ± 0.09 0.272 ± 0.004 NWA 4518_2 6.9 ± 0.12 126.61 ± 1.42 3.1 ± 0.1 8.7 ± 0.18 0.145 ± 0.003 NWA 4874_1 1.2 ± 0.03 104.20 ± 2.43 1.1 ± 0.1 1.2 ± 0.04 0.853 ± 0.019 NWA 4874_2 1.2 ± 0.02 102.70 ± 0.98 1.5 ± 0.1 1.3 ± 0.02 0.811 ± 0.013 NWA 4876_1 4.2 ± 0.07 119.03 ± 1.20 1.6 ± 0.1 5.0 ± 0.10 0.240 ± 0.004 NWA 4876_2 4.6 ± 0.07 119.13 ± 0.79 1.4 ± 0.0 5.5 ± 0.09 0.218 ± 0.003 NWA 4882_1 4.0 ± 0.06 121.40 ± 0.98 1.0 ± 0.0 4.9 ± 0.09 0.249 ± 0.004 NWA 4882_2 3.2 ± 0.05 110.75 ± 0.88 1.2 ± 0.0 3.6 ± 0.07 0.309 ± 0.005 NWA 4969_1 5.0 ± 0.08 107.21 ± 0.63 1.9 ± 0.1 5.3 ± 0.09 0.201 ± 0.003 NWA 4969_2 9.9 ± 0.17 102.32 ± 1.53 2.5 ± 0.0 10.2 ± 0.23 0.101 ± 0.002 NWA 595_1 8.3 ± 0.13 103.65 ± 0.59 6.1 ± 0.1 8.6 ± 0.14 0.120 ± 0.002 NWA595_2 3.7 ± 0.06 98.50 ± 0.82 6.2 ± 0.1 3.6 ± 0.07 0.272 ± 0.004 NWA 6077_2 9.9 ± 0.15 131.54 ± 0.69 4.3 ± 0.1 13.1 ± 0.21 0.101 ± 0.002 NWA 6474_1 2.1 ± 0.03 104.94 ± 0.66 1.3 ± 0.0 2.2 ± 0.04 0.487 ± 0.008 NWA 6474_2 2.2 ± 0.04 101.19 ± 1.63 1.3 ± 0.1 2.3 ± 0.06 0.448 ± 0.008 NWA 6962_1 0.5 ± 0.06 91.96 ± 15.97 1.3 ± 0.7 0.5 ± 0.10 1.875 ± 0.216 NWA 6962_2 9.0 ± 0.17 173.94 ± 2.04 0.7 ± 0.2 15.6 ± 0.34 0.111 ± 0.002 NWA 7297_1 11.5 ± 0.18 155.56 ± 0.74 3.9 ± 0.1 17.9 ± 0.29 0.087 ± 0.001 NWA 7297_2 4.7 ± 0.07 148.14 ± 0.80 3.8 ± 0.1 6.9 ± 0.11 0.214 ± 0.003 NWA 7605_1 5.3 ± 0.08 139.96 ± 0.99 1.4 ± 0.0 7.4 ± 0.13 0.188 ± 0.003 NWA 7605_2 0.9 ± 0.01 97.05 ± 1.06 1.8 ± 0.0 0.8 ± 0.02 1.156 ± 0.019 NWA 8777_2 4.9 ± 0.08 122.16 ± 0.59 1.0 ± 0.0 6.0 ± 0.10 0.202 ± 0.003 RaS 309_2 8.9 ± 0.14 98.94 ± 0.76 7.6 ± 0.1 8.8 ± 0.16 0.113 ± 0.002

171

Table S.7.2. Radiogenic helium and argon. Table X. Radiogenic helium and argon amounts. 4 40 Sample Herad Ar NWA 10637 123.1 ± 43 1158 ± 34 NWA 1500_1 14.8 ± 7 269 ± 22 NWA 1500_2 20.5 ± 7 180 ± 87 NWA 3151_1 13.8 ± 17 965 ± 78 NWA 3151_2 13.6 ± 17 1129 ± 17 NWA 4518_1 28.4 ± 6 1399 ± 59 NWA 4518_2 22.0 ± 6 1675 ± 27 NWA 4874_1 -30.2 ± 20 46 ± 2 NWA 4874_2 -25.5 ± 18 76 ± 12 NWA 4876_1 -23.7 ± 51 175 ± 70 NWA 4876_2 -19.8 ± 55 87 ± 22 NWA 4882_1 -20.2 ± 54 - NWA 4882_2 -36.0 ± 55 67 ± 2 NWA 4969_1 -26.5 ± 48 113 ± 40 NWA 4969_2 -24.9 ± 56 316 ± 14 NWA 595_1 8.1 ± 19 2090 ± 56 NWA595_2 7.7 ± 24 2615 ± 61 NWA 6077_2 48.2 ± 21 733 ± 20 NWA 6474_1 -45.3 ± 30 16 ± 10 NWA 6474_2 -48.6 ± 29 64 ± 38 NWA 6962_1 1993.0 ± 32 3475 ± 3779 NWA 6962_2 1744.4 ± 17 3327 ± 533 NWA 7297_1 64.2 ± 20 923 ± 14 NWA 7297_2 94.1 ± 23 1079 ± 29 NWA 7605_1 66.4 ± 37 109 ± 6 NWA 7605_2 79.4 ± 36 317 ± 4 NWA 8777_2 582.2 ± 15 165 ± 14 RaS 309_2 2.8 ± 36 3849 ± 57 Values are in 10e-8 ccSTP/g

Table S7.3. Krypton isotopes for select brachinite and brachinite-like achondrites. 172

84Kr* 78Kr/84Kr 80Kr/84Kr 82Kr/84Kr 86Kr/84Kr 81Kr/84Kr 78Kr/86Kr 80Kr/86Kr 82Kr/86Kr 83Kr/86Kr 84Kr/86Kr NWA 10637 38.5 0.61 4.0 20.7 30.9 0.0007 1.98 12.9 66.9 66.4 323.4 NWA 3151 17.1 0.61 4.0 20.5 31.1 0.0011 1.97 12.9 65.8 66.2 321.7 NWA 4882 4.6 0.61 4.2 20.4 30.8 n.d. 1.99 13.7 66.3 64.9 324.5 NWA 595 60.8 0.60 3.9 20.0 30.0 n.d. 2.01 12.9 66.6 66.4 333.2 NWA 7297 34.9 0.60 4.2 20.4 30.5 n.d. 1.97 13.7 66.8 66.2 327.9 NWA 7605 9.4 0.61 4.6 20.5 30.4 n.d. 2.02 15.1 67.4 66.0 329.2 All ratios are multiplied by 100. * indicates that values are in [10E-10 cc/g STP]. Uncertainties are typically ~1%.

Table S7.4. Xenon isotopes for select brachinite and brachinite-like achondrites. 132Xe* 124Xe/132Xe 126Xe/132Xe 128Xe/132Xe 129Xe/132Xe 129Xe* 130Xe/132Xe 131Xe/132Xe 134Xe/132Xe 136Xe/132Xe 84Kr/132Xe 1/132Xe NWA 10637 10.78 0.38 0.33 7.50 102.88 11.09 15.29 79.57 38.52 32.47 3.57 0.09 NWA 3151 5.56 0.40 0.36 7.66 101.06 5.62 15.64 80.29 37.67 31.97 3.08 0.18 NWA 4882 4.02 0.43 0.38 8.12 121.40 4.89 15.87 81.01 38.19 31.74 1.15 0.25 NWA 595 9.93 0.36 0.33 7.17 102.32 10.16 15.22 78.29 38.11 32.23 6.13 0.10 NWA 7297 8.99 0.39 0.35 7.73 173.94 15.64 15.27 79.52 38.00 32.31 0.00 0.11 NWA 7605 5.31 0.42 0.39 8.21 139.96 7.44 15.96 81.63 38.46 31.43 0.00 0.19 All ratios are multiplied by 100. * indicates that values are in [10E-10 cc/g STP]. Uncertainties are typically ~1%.

173

REFERENCES

Ammon K., Leya I., Lavielle B., Gilabert E., David J.‐C., Herpers U., and Michel R. (2008). Cross sections for the production of helium, neon, and argon isotopes by proton‐induced reactions on iron and nickel. Nuclear Instruments and Methods in Physics Research B 266: 2–21.

Badyukov D. D., Raitala J., Kostama P., and Ignatiev A. V. (2015). Chelyabinsk meteorite: Shock metamorphism, black veins and impact melt dikes, and the Hugoniot. Petrology 23:103– 115.

Bartoschewitz, R., Wlotzka, F., Clayton, R.N., Mayeda, T.K., (2003). NWA 1500: the first basaltic ureilite? (Abstract). Meteoritics and Planetary Science. 38 (Suppl.), A64.

Bartoschewitz (2018), Meteoritical Bulletin Database

Beard S. P., Swindle T. D., Isachsen C., Jenniskens P., and Shaddad M. (2013). Ar‐Ar analysis of Almahata Sitta ordinary chondrites (abstract #2311). 44th Lunar and Planetary Science Conference.

Beard S. P., Kring D. A., Isachsen C. E., Lapen T. J., Zolensky M. E., and Swindle T. D. (2014). Ar‐Ar analysis of Chelyabinsk: Evidence for a recent impact (abstract #1807). 45th Lunar and Planetary Science Conference.

Beard, S. B.; Swindle, T. D.; Domanik, K. J. (2015). Searching for Potassium Host Phases in Preparation for Ar-Ar Analysis of Brachinite and Brachinite-Like Achondrites. No. 1832, p. 2568. 46th Lunar and Planetary Science Conference.

Beard S. P. and Swindle T. D. (2017). Search for evidence of source event grouping among ureilites. Meteoritics & Planetary Science 52:2343–2352.

Beckinsale, R.D., Gale, N.H., (1969). A reappraisal of the decay constants and branching ratio of 40K. Earth Planetary Science Letters 6, 289-294.

Becquerel, A.H.: On the rays emitted by phosphorescence. Comptes Rendus, 122, 420–421, 1896.

Benedix G. K., Lauretta D. S. and McCoy T. J. (2005) Thermo- dynamic constraints on the formation conditions of winonaites and silicate-bearing IAB irons. Geochimica Et Cosmochimica Acta 69, 5123–5131.

Benedix G. K., McCoy T. J. and Spratt J. (2009) Thermodynamic constraints on the formation history of lodranites. Lunar and Planetary Science Conference XL. #1494 (abstr.).

174

Binns W. R., Wiedenbeck M. E., Christian E. R., Cummings A. C., George J. S., Israel H. M., Leske R. A., Mewaldt R. A., Stone E. C., Rosenvinge T. T., and Yanasak N. E. (2001) GCR neon isotopic abundances: Comparison with Wolf-Rayet star models and meteorite abundances. In Solar and Galactic Composition (R. F. Wimmer-Schweingruber, ed.), pp. 257–262. American Institute of Physics, New York.

Bischoff A., Horstmann M., Pack A., Laubenstein M., and Haberer S. (2010). Asteroid 2008 TC3—Almahata Sitta: A spectacular breccia containing many different ureilitic and chondritic lithologies. Meteoritics & Planetary Science 45:1638–1656.

Bogard, D. D., Nyquist, L. E., Johnson, P., Wooden, J., & Bansal, B. (1983). Chronology of Brachina. Meteoritics, 18, 269-270.

Bogard, D. D. (1995). Impact ages of meteorites: A synthesis. Meteoritics, 30, 244-268.

Bogomolova E. S., Skublov S. G., Marin Y. B, Stepanov S. Y., Antonov A. V., and Galankina O. L. (2013). Sm–Nd age and isotope geochemistry of minerals of the Chelyabinsk meteorite. Doklady Earth Sciences 452:1034–1038.

Bouvier A. (2013). Pb‐Pb chronometry of the dark melt lithology of the Chelyabinsk LL chondrite. Large Meteorite Impacts and Planetary Evolution V. LPI Contributions 1737:3087.

Brown, P.G., Assink, J.D., Astiz, L., Blaauw, R., Boslough, M.B., Borovicˇka, J., Brachet, N., Brown, D., Campbell-Brown, M., Ceranna, L., Cooke, W., de Groot-Hedlin, C., Drob, D.P., Edwards, W., Evers, L.G., Garces, M., Gill, J., Hedlin, M., Kingery, A., Laske, G., Le Pichon, A., Mialle, P., Moser, D.E., Saffer, A., Silber, E., Smets, P., Spalding, R.E., Spurny ́, P., Tagliaferri, E., Uren, D., Weryk, R.J., Whitaker, R., Krzeminski, Z., (2013). A 500-kiloton airburst over Chelyabinsk and an enhanced hazard from small impactors. Nature 503, 238–241.

Busemann H., Baur H., and Wieler R. (2000). Primordial noble gases in “phase Q” in carbonaceous and ordinary chondrites studied by closed‐system stepped etching. Meteoritics & Planetary Science 35:949–973.

Cherniak D.J., Lanford W.A., Ryerson F.J. (1991). Lead diffusion in apatite and zircon using ion implantation and Rutherford bascattering techniques. Geochimica Et Cosmochimica Acta 55:1663-1673

Clayton, R. N. and T. K. Mayeda (1983). Oxygen isotopes in eucrites, shergottites, nakhlites, and chassignites. Earth Planetary Science Letters 62: 1-6.

Clayton, R. N. and T. K. Mayeda (1988). Formation of ureilites by nebular processes. Geochimica Et Cosmochimica Acta 52: 1313-1318.

175

Clayton, R. N. (1993). Oxygen isotopes in meteorites. Annual Reviews in Earth Planetary Science 21: 115-149. Clayton, R. N. and T. K. Mayeda (1996). Oxygen isotope studies of achondrites. Geochimica Et Cosmochimica Acta 60: 1999-2017.

Clayton, R.N. (2003). 1.06 - Oxygen Isotopes in Meteorites”. Treatise on Geochemistry 1, 129– 142.

Cohen, B.A., Goodrich, C.A., Keil, K., (2004). Feldspathic clast populations in polymict ureilites: Stalking the missing basalts from the ureilite parent body. Geochimica Et Cosmochimica Acta. 68, 4249-4266.

Day, J. M. D., et al. (2012). Origin of felsic achondrites Graves Nunataks 06128 and 062129, and ultramafic brachinites and brachinite-like achondrites by partial melting of volatile-rich primitive parent bodies. Geochimica Et Cosmochimica Acta 81: 94-128.

Day J. M. D., Corder C. A., Dhaliwal J. K., Liu Y., and Taylor L. A. (2014). The Chelyabinsk fall highly siderophile element abundance and 187Os/188Os composition and comparison with ordinary and carbonaceous chondrites (abstract #5078). Meteoritics & Planetary Science 49.

Dahymple, G.B., Lanphere, M.A., (1969). Potassium-Argon Dating. W.H. Freeman, San Francisco, 258 pp

Deino, A.L., (2001). User’s manual for Mass Spec v. 5.02. Berkeley Geochronology Center Special Publication 1a. 119 pp

Domanik K. and Stephan F., Meteoritical Bulletin Database 2018.

Downes H., Mittlefehldt D. W., Kita N. T., and Valley J. W. (2008). Evidence from polymict ureilites for a disrupted and re‐accreted single ureilite parent asteroid gardened by several distinct impactors. Geochimica et Cosmochimica Acta 72:4825–4844.

Dunlap, D. R.; Ku, Y. J.; Garvie, L. A. J.; Wadhwa, M. (2015) Petrology of Ungrouped and Anomalous Achondrites SaU 493, NWA 4470, NWA 6962, and NWA 5297, 46th Lunar and Planetary Science Conference. No. 1832, p.2570

Dunlap D. R., Romaniello S.J., Wadhwa M. (2016). 53Mn-53Cr Systematics of the Brachinite NWA 4882. 47th Lunar & Planetary Science Conference Abstract #3055.

Eugster O. (1988). Cosmic‐ray production rates for 3He, 21Ne, 38Ar, 83Kr, and 126Xe in chondrites based on 81Kr‐Kr exposure ages. Geochimica et Cosmochimica Acta 52:1649–1662.

Eugster, O. and T. Michel (1995). Common asteroid break-up events of eucrites, diogenites, and howardites and cosmic-ray production rates for noble gases in achondrites. Geochimica Et Cosmochimica Acta 59: 177-199. 176

Eugster O., Herzog G. F., Marti K., and Caffee M. W. (2006). Irradiation records, cosmic‐ray exposure ages, and transfer times of meteorites. In Meteorites and the early solar system, edited by Lauretta D. S.and McSween H. Y. Jr. Tucson: The University of Arizona Press. pp. 829–851.

Fireman, E. L., J. DeFelice, (1960). Journal of Geophysical. Research, 65, 3035.

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. Meteoritics & Planetary Science 46:35–52.

Galimov, E. M. (2013). Chelyabinsk meteorite - an LL5 chondrite. Solar System Research 47:255-259.

Gardner-Vandy, K. G., et al. (2013). A petrologic, thermodynamic and experimental study of brachinites: Partial melt residues of an R chondrite-like precursor. Geochimica Et Cosmochimica Acta: 36-57.

Gibson, K. E., D. D. Bogard, (1978). Chemical alterations of the Holbrook chondrite resulting from terrestrial weathering, Meteoritics, 13, 277–289.

Goodrich C. A., Jones J. H., and Berkley J. L. (1987). Origin and evolution of the ureilite parent magmas: Multi‐stage igneous activity on a large parent body. Geochimica et Cosmochimica Acta 51:2255–2274.

Goodrich, C. A. (1992). "Ureilites: A critical review." Meteoritics and Planetary Science. 27: 327-352.

Goodrich, C. A. (1999). Are ureilites residues from partial melting of chondritic material? The answer from MAGPOX. Meteoritics and Planetary Science. 34, 109-119.

Goodrich C. A. and Righter K. (2000). Petrology of unique achondrite Queen Alexandra Range 93148: A piece of the pallasite (howardite‐eucrite‐diogenite?) parent body? Meteoritics & Planetary Science 35:521–535.

Goodrich C.A. and Delaney J.S. (2000). Fe/Mg–Fe/Mn relations of meteorites and primary heterogeneity of primitive achondrite parent bodies. Geochimica Et Cosmochimica Acta, 64 (2000), pp. 2255-2273.

Goodrich C.A., Krot , A.N. , Scott E.R.D. , Taylor G.J. , Fioretti A.M. , Keil K. (2002). Formation and evolution of the ureilite parent body and its offspring. Lunar and Planetary Science 33. Abstract 1379

Goodrich C. A. (2004). Ureilitic breccias: Clues to the petrologic structure and impact disruption of the ureilite parent asteroid. Chemie der Erde 64:283–327

177

Goodrich C. A., Wlotzka F., Ross D. K., and Bartoschewitz R. 2006. Northwest Africa 1500: Plagioclase‐bearing monomict ureilite or ungrouped achondrite?Meteoritics & Planetary Science 41:925–952.

Goodrich, C. A., Kita, N. T., Spicuzza, M. J., Valley, J. W., Zipfel, J., Mikouchi, T., & Miyamoto, M. (2010). The Northwest Africa 1500 meteorite: Not a ureilite, maybe a brachinite. Meteoritics and Planetary Science., 45, 1906-1928.

Graf T, Baur H, Signer P (1990) A model for the production of cosmogenic nuclides in chondrites. Geochimica Et Cosmochimica Acta 54:2521-2534

Graf T. and Marti K. (1991) Exposure ages of LL‐ and L/LL‐ chondrites and implications for parent body history (abstract). Meteoritics 26, 340.

Graf, T. and K. Marti (1995). Collisional history of H chondrites. Journal of Geophysical. Research 100: 21247-21263.

Grasty R.L. and Mitchell J.G. (1966), Single sample potassium- argon ages using the omegatron, Earth Planetary Science Letters 1:121.

Greenwood R. C., Franchi I. A., Gibson J. M., and Benedix G. K. (2007). Oxygen isotope composition of the primitive achondrites (abstract #2163). 38th Lunar and Planetary Science Conference.

Greenwood R. C., Franchi I. A., Gibson J. M., and Benedix G. K. (2012). Oxygen isotope variation in primitive achondrites: The influence of primordial, asteroidal and terrestrial processes. Geochimica et Cosmochimica Acta 94:146–163.

Greenwood R.C., Burbine T.H. , Miller M.F. , Franchi I.A. (2017). Melting and differentiation of early-formed asteroids: the perspective from high precision oxygen isotope studies Chemie der Erde 77 pp. 1-43.

Haba M. K., Sumino H., Nagao K., Mikouchi T., Komatsu M., and Zolensky M. E.(2014). Noble gases in the Chelyabinsk meteorite (abstract #1732). 45th Lunar and Planetary Science Conference.

Heymann D, Anders E, Rowe MW. (1967), Meteorites with short cosmic-ray exposure ages. Geochimica Et Cosmochimica Acta 31:1791-1809

Hoffman, Zurfluh, and Gnos, Meteoritical Bulletin Database 2018

178

Hohenberg, C. M., Marti, K., Podosek, F. A., Reedy, R. C., & Shirck, J. R. (1978). Comparisons between observed and predicted cosmogenic noble gases in lunar samples. Proceedings of Lunar and Planetary Science, 9th, 2311-2344.

Horstmann M. and Bischoff A. (2014). The Almahata Sitta polymict breccia and the late accretion of asteroid 2008 TC3. Chemie der Erde 74:149–183.

Irving, A.J., Kuehner, S.M., Rumble III, D., (2005). Brachinite NWA 3151 and (?) brachinite NWA 595 (abstract). Meteoritics and Planetary Science 40, A73.

Irving A. J. and Rumble D. III. (2006). Oxygen isotopes in Brachina, Sahara 99555 and NWA 1054 (abstract #5288). Meteoritics & Planetary Science 41:A84.

Irving, A.J., Kuehner, S.M., Ziegler, K., (2013). Petrology and oxygen isotopic composition of brachinite-like achondrites Northwest Africa 7388 and Northwest Africa 7605, and evidence for late-stage methane-triggered reduction (abstract). Lunar and Planetary Science 44

Irving and Kuehner, Meteoritical Bulletin Database (2018).

Ivanova M. A., Badyukov D. D., Ryazantsev K. M., Lorenz C. A., Demidova S. I., Sadilenko D. A., Artemieva N., Korochantsev A. V., Skripnik A. Ya., Ivanov A. V., and Nazarov M. A. (2013). Fall, searching and first study of the Chelyabinsk meteorite. 76th Annual Meeting of the Meteoritical Society. Meteoritics & Planetary Science.

Johnson, J. E., J. M. Scrymgour, E. Jarosewich, B. Mason, Brachina meteorite — A chassignite from South Australia, Rec. S. Australia. Mus., 17, 309–319, 1977.

Jourdan, F., Andreoli, M.A.G., McDonald, I., Maier, W.D., (2010). 40Ar/39Ar thermochronology of the fossil LL6-chondrite from the Morokweng crater. South Africa, Geochimica Et Cosmochimica Acta 74, 1734–1747.

Jourdan F, Timms NE, Eroglu E, Mayers C, Frew A, Bland PA, Collins GS, Davison TM, Abe M. Yada T. (2017) Collisional history of asteroid Itokawa. Geology 45, 819–822.

Kallemeyn, G. W., A. E. Rubin, D. Wang and J. T. Wasson (1989). Ordinary chondrites: Bulk compositions, classification, lithophile-element fractionations, and composition-petrographic type relationships. Geochimica Et Cosmochimica Acta 53: 2747-2767.

Kamioka M., Terada K., Hidaka H., Kimura K., and Skublov S. (2014). Secondary ion mass spectrometry (SHRIMP) U‐Pb dating of Chelyabinsk meteorite. The 62nd Annual Conference on Mass Spectrometry, Osaka, Japan: P2‐05.

Keil, K. (2014). Brachinite meteorites: Partial melt residues from an FeO-rich asteroid. Chemie der Erde 74: 311-329.

179

Kita, N. T., Ikeda, Y., Togashi, S., Liu, Y., Morishita, Y., & Weisberg, M. K. (2004). Origin of ureilites inferred from a SIMS oxygen isotopic and trace element study of clasts in the Dar al Gani 319 polymict ureilite. Geochimica Et Cosmochimica Acta, 68, 4213-4235.

Koning A. J., Hilaire S., and Duijvestijn M. C. (2005). TALYS: comprehensive nuclear reaction modelling. Proceedings of the International Conference on Nuclear Data for Science and Technology. pp. 1154–1159.

Korochantseva, E. V., Trieloff, M., Lorenz, C. A., Buykin, A. I., Ivanova, M. A., Schwarz, W. H.,Jessberger, E. K. (2007). L-chondrite asteroid breakup tied to Ordovician meteorite shower by multiple isochron Ar-40-Ar-39 dating. Meteoritics and Planetary Science, 42, 113-130.

Lapen T. J., Kring D. A., Zolensky M. E., Andreasen R., Righter M., Swindle T. D., and Beard S. P. (2014). Uranium‐lead isotope evidence in the Chelyabinsk LL5 chondrite meteorite for ancient and recent thermal events (abstract #2561). 45th Lunar and Planetary Science Conference.

Leya I, Lange HJ, Neumann S, Wieler R, Michel R (2000) The production of cosmogenic nuclides in stony meteoroids by galactic cosmic ray particles. Meteoritics and Planetary Science 35:259-286

Leya I, Neumann S, Wieler R, Michel R (2001) The production of cosmogenic nuclides by GCR-particles for 2π exposure geometries. Meteoritics and Planetary Science 36:1547-1561

Leya I. and Masarik J. (2009). Cosmogenic nuclides in stony meteorites revisited. Meteoritics & Planetary Science 44:1061–1086.

Lindsay F. N., Herzog G. F., Park J., Turrin B. D., Delaney J. S., and Swisher C. C. (2015). Chelyabinsk Ar ages—A young heterogeneous LL5 chondrite (abstract #2226). 46th Lunar and Planetary Science Conference.

Lingenfelter R. E., Higdon J. C., and Ramaty R. (2000) Cosmic ray acceleration in superbubbles and the compositon of cosmic rays. In Acceleration and Transport of Energetic Particles Ob- served in the Heliosphere (R. Mewaldt et al., eds.), pp. 375. American Institute of Physics, New York.

Lorenz, C.A., Teplyakova, S.N., Humayun, M., Ivanova, M.A., Franchi, I.A.,Greenwood, R.C., 2011. Origin of the ungrouped achondrite NWA 4518: Mineralogy and geochemistry of FeNi- metal (abstract). Lunar and Planetary Science 42(#1291).

Marti K. and Graf T. (1992) Cosmic ray exposure history of ordinary chondrites. Annual Reviews in Earth Planetary Science 20, 221–243.

Masarik J, Reedy RC (1994) Effects of bulk composition on nuclide production processes in meteorites. Geochim Cosmochim Acta 58:5307-5317 180

Masarik J, Nishiizumi K, Reedy RC (2001) Production rates of cosmogenic 3He, 21Ne, and 22Ne in ordinary chondrites and the lunar surface. Meteoritics and Planetary Science 36:643- 650

Matsuhisa Y., Goldsmith J. R., and Clayton R. N. (1978) Mechanisms of hydrothermal crystallization of quartz at 250°C and 15 kbar. Geochim. Cosmochim. Acta 42, 173–182.

Mcconville P., Kelley S. and Turner G.(1988) Laser probe 40Ar‐39Ar studies of the Peace River shocked L6 chondrite. Geochim. Cosmochim. Acta 52, 2487–2499.

McCoy, T.J., Mittlefehldt, D.W., and Wilson, L. (2006) Asteroid differentiation. In D.S. Lauretta and H.Y. McSween, Jr., Eds., Meteorites and the Early Solar System II, p. 733–746. University of Arizona Press, Tucson.

Meier M. M. M., Alwmark C., Bajt S., Böttger U., Busemann H., Fujiya W., Gilmour J., Heitmann U., Hoppe P., Hübers H.‐W., Marone F., Ott U., Pavlov S., Schade U., Spring N., Stampanoni M., and Weber I. (2014). A precise cosmic‐ray exposure age for an olivine grain from the surface of near‐Earth asteroid (25143) Itokawa (abstract #1247). 45th Lunar and Planetary Science Conference.

Merrihue, C. & Turner, G. Journal of Geophysical. Research 71, 2852–2857 (1966).

Mittlefehldt, D.W., McCoy, T.J., Goodrich, C.A., Kracher, A., (1998). Non-chondritic meteorites from asteroidal bodies. In: Papike, J.J. (Ed.), Planetary Materials, vol. 36. pp. 4–1 to 4–195.

Mitchell, J.G. The argon-40/argon-39 method for potassium-argon age determinations Geochim. Cosmochim. Acta, 32 (1968), pp. 781-790

Mittlefehldt, D. W., Bogard, D. D., Berkley, J. L., & Garrison, D. H. (2003). Brachinites: Igneous rocks from a differentiated asteroid. Meteoritics and Planetary Science, 38, 1601-1625.

Mittlefehldt D. W. (2004) Achondrites. In Treatise on Geochemistry, Vol. 1: Meteorites, Comets and Planets (A. M. Davis, eds.), pp. 291–324. Elsevier, Oxford.

Mittlefehldt, D. W., & Hudon, P. (2004). Northwest Africa 1500: Not a basaltic ureilite; not even a ureilite. Meteoritics and Planetary Science, 39, A69.

Nakamura T., Noguchi T., Tanaka M., Zolensky M. E., Kimura M., Tsuchiyama A., Nakato A., Ogami T., Ishida H., Uesugi M., Yada T., Shirai K., Fujimura A., Okazaki R., Sandford S. A., Ishibashi Y., Abe M., Okada T., Ueno M., Mukai T., Yoshikawa M., and Kawaguchi J. (2011). Itokawa dust particles: A direct link between S‐type asteroids and ordinary chondrites. Science 333:1113–1116.

Nakamura E., Kunihiro T., Kitagawa H., Kobayashi K., Ota T., Sakaguchi C., Shimaki Y., Tanaka R., Tsujimori T., Yamanaka M., Bebout G. E.., Grokhovsky V. I., Koroleva O. 181

N., Litasov K. D., and Malkovets V. (2015). Recycling of an asteroid via a comet inferred from the Chelyabinsk meteorite (abstract #1865). 46th Lunar and Planetary Science Conference.

Nehru C. E., Prinz M. and Zucker S. M.(1979) Brachina: Origin, melt inclusions and relationship to Chassigny (abstract). Meteoritics 14, 493–494.

Nehru, C.E., Prinz, M., Delaney, J.S., Dreibus, G., Palme, H., Spettel, B., Wänke, H., (1983) Brachina: a new type of meteorite, not a chassignite. Journal of Geophysical. Research 88 B237–B244.

Nehru, C. E., Prinz, M., Weisberg, M. K., Ebihara, M. E., Clayton, R. N., & Mayeda, T.K. (1992). Brachinites: A new primitive achondrite group. Meteoritics, 27, 267.

Niedermann S (2002) Cosmic-ray-produced noble gases in terrestrial rocks as a dating tool for surface processes. Reviews in Mineral Geochemistry 47:731–784

Nishiizumi K., Caffee M. W., Huber L., Welten K. C., and Wieler R. (2013). Cosmogenic radionuclides and noble gases in Chelyabinsk meteorite (abstract #5260). Meteoritics & Planetary Science 48.

Okazaki, R., Nakamura, T., Takaoka, N., & Nagao, K. (2003). Noble gases in ureilites released by crushing. Meteoritics and Planetary Science, 38, 767-781.

Ott, U., Lohr, H. P., & Begemann, F. (1985). Noble gases and the classification of Brachina. Meteoritics, 20, 69-78.

Ott, U., Lohr, H. P., & Begemann, F. (1987). Noble gases in ALH 84025: Like Brachina, unlike Chassigny. Meteoritics, 22, 476-477.

Ott, U. (2002). Noble gases in meteorites -- Trapped components. Reviews in Mineral Geochemistry., 47, 71-100.

Ott, U., Löhr, H.-P., Begemann, F., (1993). Noble gases in Yamato-75097 inclusion: similarities to brachinites (only?) (abstract). In: Antarctic Meteorites, vol. 18, NIPR, Tokyo, pp. 236–239.

Park J., Herzog G. F., Haba M. K., Nagao K. (2014). Exposure Ages of Ureilites: Radionuclides and Noble Gases (abstract). Lunar and Planetary Science 45, 1777.

Petaev M. I., Barsukova L. D., Lipschutz M. E., Wang M.-S., Ariskin A. A., Clayton R. N. and Mayeda T. K. (1994) The Divnoe meteorite: petrology, chemistry, oxygen isotopes and origin. Meteoritics and Planetary Science 29, 182–199.

Patzer, A., & Schultz, L. (2001). Noble gases in enstatite chondrites I: Exposure ages, pairing and weathering effects. Meteoritics and Planetary Science 36, 947-961.

182

Patzer, A., et al. (2003). New noble gas data of primitive and differentiated achondrites including Northwest Africa 011 and Tafassasset. Meteoritics and Planetary Science 38: 1485-1497.

Petrova E. V., Grokhovsky V. I., and Muftakhetdinova R. F. (2016). Heat treatment of the different structure zones in the Chelyabinsk meteorite (abstract #6487). Meteoritics & Planetary Science 51.

Popova, O. P., Jenniskens, P., Emel'yanenko, V., Kartashova, A., Biryukov, E., Khaibrakhmanov, S., Mikouchi, T. (2013). Chelyabinsk airburst, damage assessment, meteorite recovery, and characterization. Science.

Povinec P. P., Laubenstein M., Jull A. J. T., Ferrière L., Brandstätter F., Sýkora I., Masarik J., Beňo J., Kováčik A., Topa D., and Koeberl C. (2015). Cosmogenic radionuclides and mineralogical properties of the Chelyabinsk (LL5) meteorites: What do we learn about the meteoroid? Meteoritics & Planetary Science 50:273–286.

Rai, V. K., Murty, S. V. S., & Ott, U. (2003). Noble gases in ureilites: Cosmogenic, radiogenic, and trapped components. Geochim. Cosmochim. Acta, 67, 4435-4456.

Renne P. R., Balco G., Ludwig K. R., Mundil R., and Min K. (2011). Response to the comment by W. H. Schwarz et al. on “Joint determination of 40K decay constants and 40Ar*/40K for the Fish Canyon sanidine standard, and improved accuracy for 40Ar/39Ar geochronology” by P. R. Renne et al. (2010). Geochimica et Cosmochimica Acta 75:5097–5100.

Riebe M. E. I., Welten K., Meier M. M. M., Wieler R., Bart M. I. F., Ward D., Laubenstein M., Bischoff A., Caffee M. W., Nishiizumi K., and Busemann H.Forthcoming. (2017) Cosmic‐ ray exposure ages of six chondritic Almahata Sitta fragments. Meteoritics & Planetary Science.

Righter, K., & Drake, M. J. (1996). Evidence for a deep terrestrial magma ocean: Prediction of siderophile element metal/silicate partition coefficients at P, T, fO2 and metallic and silicate melt composition. Lunar and Planetary Science XXVII, 1075-1076.

Righter K. and Neff K. E. (2007) Temperature and oxygen fugacity constraints on CK and R chondrites and implications for water and oxidation in the early solar system. Polar Science. 1, 25–44.

Righter K., Abell P., Agresti D., Berger E. L., Burton A. S., Delaney J. S., Fries M. D., Gibson E. K., Haba M. K., Harrington R., Herzog G. F., Keller L. P., Locke D., Lindsay F. N., McCoy T. J., Morris R. V., Nagao K., Nakamura‐Messenger K., Niles P. B., Nyquist L. E., Park J., Peng Z. X., Shih C.‐Y., Simon J. I., Swisher C. C. III, Tappa M. J., Turrin B. D., and Zeigler R. A. (2015). Mineralogy, petrology, chronology, and exposure history of the Chelyabinsk meteorite and parent body. Meteoritics & Planetary Science 50:1790–1819.

Roddick, J. C., Cliff, R. A. & Rex, D. C. Earth Planetary Science Letters 48, 185–208 (1980).

183

Rumble D. III, Irving A. J., Bunch T. E., Wittke J. H., and Kuehner S. M. (2008). Oxygen isotopic and petrological diversity among brachinites NWA 4872, NWA 4874, NWA 4882 and NWA 4969 (abstract #1974). 34th Lunar and Planetary Science Conference.

Rumble, D., Zolensky, M. E., Friedrich, J. M., Jenniskens, P., & Shaddad, M. H. (2010). The oxygen isotope composition of Almahata Sitta. Meteoritics and Planetary Science, 45, 1765- 1770.

Rumble D., Meteoritical Bulletin Database (2018)

Scherer, P., Loeken, T., & Schultz, L. (1994). Noble gases in chondrites found in hot deserts. Meteoritics, 29, 529.

Schultz, L., & Franke, L. (2004). Helium, neon, and argon in meteorites: A data collection. Meteoritics and Planetary Science, 39, 1889-1890.

Singerling, S.A., McCoy, T.J., Gardner-Vandy, K.G., (2013). Possible evidence for sulfidization reactions in the Miller Range brachinites (?) (abstract). Lunar and Planetary Science 44 # 1669.

Skublov S. G., Guseva N. S., Presnyakov S. L., Li X.‐H., Marin Yu. B., Sergeev S. A., Berezhnaya N. G., Tyuleneva N. V., and Alekseev V. I. (2015). U‐Pb age of zircon and the history of impact transformations of the Chelyabinsk meteorite. Doklady Earth Sciences 462:586–591.

Smith, J. V., I. M. Steele, C. A. Leitch, (1980) Mineral chemistry of the shergottites, nakhlites, Chassigny, Brachina, pallasites and urelites, Journal of Geophysical. Research., 88.

Steiger R. H. and Jäger E. (1977). Subcommission on geochronology: Convention on the use of decay constants in geo‐ and cosmochronology. Earth and Planetary Science Letters 36:359–362.

Stoenner, R. W., O. A. Schaeffer, S. Katcoff (1965) Half-lives of argon-37, argon-39, and argon- 42 Science, 148 (1965), pp. 1325-1328

Swindle, T. D., Kring, D. A., Burkland, M. K., Hill, D. H., & Boynton, W. V. (1998). Noble gases, bulk chemistry, and petrography of olivine-rich achondrites Eagles Nest and LEW88763: comparison to brachinites. Meteoritics and Planetary Science 33, 31-48.

Swindle, T. D.; Kring, D. A.; Olson, E. K.; Isachsen, C. E. (2006), Ar-Ar Dating of Shock- melted Ordinary Chondrites: Chronology of Asteroidal Impacts 37th Annual Lunar and Planetary Science Conference.

Swindle, T.D., Isachsen, C.E., Weirich, J.R., Kring, D.A., (2009). 40Ar–39Ar ages of H- chondrite impact melt breccias. Meteoritics and Planetary Science 44, 747–762. 184

Swindle, T. D., Isachsen, C. E.; Weirich, J. R.; Kimura, M. (2011). 40Ar-39Ar studies of shocked L6 chondrites Allan Hills 78003, Yamato 74445, and Yamato 791384. Lunar and Planetary Science Conference 42: Abstract #1897.

Swindle, T. D., Kring, D. A., & Weirich, J. R. (2014). 40Ar-39Ar ages of impacts involving ordinary chondrite meteorites. In F. Jourdan, D. Mark, & C. Verati (Eds.), 40Ar/39Ar dating: from geochronology to thermochronology, from archaeology to planetary sciences. (Vol. 378, pp. 333-347). London: Geological Society.

Swindle, T. D. and Weirich, J.R. (2017). The effect of Partial Thermal Resetting on 40Ar–39Ar “Plateaus”. 48th Lunar and Planetary Science Conference, Abs. #1265.

Takagami, Y. and I. Kaneoka (1987). Investigations of the effect of shock on the Antarctic meteorites by the 40Ar-39Ar method.Proc. NIPR Symp. Antarct. Met. 11: 133-143.

Trieloff, M., Korochantseva, E. V., A. I. Buikin, J. Hopp, M. A. Ivanova, A. V. Korochantsev, The Chelyabinsk meteorite: Thermal history and variable shock effects recorded by the 40Ar– 39Ar system, Meteoritics and Planetary Science, 2018.

Turner, G., Cadogan, P. H., & Yonge, C. J. (1973). Argon selenochronology. Proeedings Lunar and Planetary Science Conference, 4th, 1889-1914.

Vermeesch, P., (2018). IsoplotR: a free and open toolbox for geochronology Geoscience Frontiers.

Wadhwa, M., Shukolyukov, A., Lugmaier, G.W., (1998). 53 Mn–53 Cr systematics in Brachina: a record of one of the earliest phases of igneous activity on an asteroid (abstract). Lunar and Planetary Science 29.

Warren P. H. and Kallemeyn G. W. (1989). Allan Hills 84025: The second brachinite, far more differentiated than Brachina, and an ultramafic achondritic clast from L chondrite Yamato 75097. Proceedings, 19th Lunar and Planetary Science Conference. pp. 475–486.

Warren P. H. and Kallemeyn G. W. (1992). Explosive volcanism and the graphite‐oxygen fugacity buffer on the parent asteroid(s) of the ureilite meteorites. Icarus 100:110–126.

Warren P. H. and Rubin A. E. (2012). The Miller Range 090340 dunite: Not a uniquely ferroan ureilite, not even a ureilite (abstract #2528). 43rd Lunar and Planetary Science Conference.

Weber I. and Bischoff A. (1998) Mineralogy and chemistry of the ureilites Hammadah al Hamra 064 and Jalanash. In Lunar and Planetary Science XXIX, abstract #1365.

Weigel A, Eugster O, Koeberl C, Michel R, Krähenbühl U, Neumann S (1999) Relationships among lodranites and acapulcoites: Noble gas isotopic abundances, chemical composition, cosmic ray exposure ages, and solar cosmic ray effects. Geochim Cosmochim Acta 63:175–192 185

Weirich J. R., Isachsen C., Swindle T. D., and Kring D. A. (2009). Ar‐Ar impact ages of shocked LL chondrites (abstract #5368). Meteoritics & Planetary Science 44:A215.

Weisberg, M. K., McCoy, T. J., & Krot, A. N. 2006, Meteorites and the Early Solar System II, ed. D. S. Lauretta & H. Y. McSween Jr. (Tucson, AZ: Univ. Arizona Press), 19-52.

Wieler, R. (2002). Cosmic-ray-produced noble gases in meteorites. Reviews in Mineral Geochemistry 47: 125-170.

Ziegler K. (2018) Meteoritical Bulletin Database.