Pb isotopic age of the Allende chondrules

1321 © The Meteoritical Society, 2007. Printed in USA.

Meteoritics & Planetary Science 42, Nr 7/8, 1321–1335 (2007)Abstract available online at http://meteoritics.org

Pb isotopic age of the Allende chondrules

Yuri AMELIN1†* and Alexander KROT2

1Geological Survey of Canada, 601 Booth Street, Ottawa, Ontario, Canada, K1A 0E82Hawai‘i Institute of Geophysics and Planetology, School of Ocean, Earth Science and Technology,

University of Hawai‘i at Manoa, Honolulu, Hawai‘i 96822, USA†Present address: Research School of Earth Sciences, The Australian National University, 61 Mills Road, Canberra, ACT 0200, Australia

*Corresponding author. E-mail: [email protected]

(Received 20 September 2006; revision accepted 23 February 2007)

The appendix table for this article is available online at http://meteoritics.org.

Abstract–We have studied Pb-isotope systematics of chondrules from the oxidized CV3carbonaceous chondrite Allende. The chondrules contain variably radiogenic Pb with a 206Pb/204Pbratio between 19.5–268. Pb-Pb isochron regression for eight most radiogenic analyses yielded thedate of 4566.2 ± 2.5 Ma. Internal residue-leachate isochrons for eight chondrule fractions yieldedconsistent dates with a weighted average of 4566.6 ± 1.0 Ma, our best estimate for an average age ofAllende chondrule formation. This Pb-Pb age is consistent with the range of model 26Al-26Mg ages ofbulk Allende chondrules reported by Bizzarro et al. (2004) and is indistinguishable from Pb-Pb agesof Ca-Al-rich inclusions (CAIs) from CV chondrites (4567.2 ± 0.6 Ma) (Amelin et al. 2002) and theoldest basaltic meteorites. We infer that chondrule formation started contemporaneously with orshortly after formation of CV CAIs and overlapped in time with formation of the basaltic crust andiron cores of differentiated asteroids. The entire period of chondrule formation lasted from 4566.6 ±1.0 Ma (Allende) to 4564.7 ± 0.6 Ma (CR chondrite Acfer 059) to 4562.7 ± 0.5 Ma (CB chondriteGujba) and was either continuous or consisted of at least three discrete episodes. Since chondrules inCB chondrites appear to have formed from a vapor-melt plume produced by a giant impact betweenplanetary embryos after dust in the protoplanetary disk had largely dissipated (Krot et al. 2005), therewere possibly a variety of processes in the early solar system occurring over at least 4–5 Myr that wenow combine under the umbrella name of “chondrule formation.”

INTRODUCTION

Chondrules (millimeter- to submillimeter-sized spherulescomposed mostly of mafic silicates) are the most abundantcomponent in most chondrites (Zanda 2004). Chondrules inmost meteorites are believed to have formed by melting ofsolid precursors, including fine-grained dust, chondrules ofearlier generations and refractory inclusions, in theprotoplanetary disk (e.g., Jones et al. 2005 and referencestherein; Russell et al. 2005). Age relationships betweenchondrules and their more refractory counterparts, Ca-Al-richinclusions (CAIs), which are considered the earliest solidsformed in the solar system (Amelin et al. 2002), wouldtherefore provide an estimate for the lifespan of theprotoplanetary disk (Russell et al. 2006 and referencestherein). Although petrographic, mineralogical, chemical, andisotopic observations suggest that CAIs predate chondruleformation (Russell et al. 2005 and references therein), the

absolute age difference between these components remainspoorly constrained.

The relative ages of chondrules and CAIs have beendetermined in many studies, using mostly isotopic effectsproduced by decay of short-lived, now extinct radionuclides,including 26Al (t1/2 ~ 0.73 Myr), 53Mn (t1/2 ~ 3.7 Myr), and 129I(t1/2 ~ 16 Myr) (Kita et al. 2005 and references therein). Usingextinct-nuclide chronometers to measure the time differencebetween CAI and chondrule formation requires assumptionsof homogeneous initial distribution of the parent nuclides inthe CAI- and chondrule-forming regions, and both extinct-and extant-nuclide chronometers require the lack ofsubsequent disturbance of these isotope systematics duringparent body processing, such as metamorphism and aqueousalteration. Although the parent body effects on isotopicsystematics of chondrules and CAIs can be minimized bystudying mineralogically primitive chondrites, theassumption of homogeneous distribution of short-lived

1322 Y. Amelin and A. Krot

radionuclides in the CAI- and chondrule-forming regionsmay not be necessarily true, considering (i) variable and insome cases, large isotopic anomalies of many elements inCAIs (e.g., Ireland 1990), (ii) the possible irradiation origin of10Be (McKeegan et al. 2000) and 7Be (Chaussidon et al. 2006)in CAIs, and (iii) the proposed irradiation origin, sole orpartial, of other short-lived radionuclides (26Al, 41Ca, 53Mn)commonly used for relative chronology (Goswami et al. 2005and references therein). We note that, from bulk Mg isotopemeasurements of chondrites, Martian meteorites, andterrestrial samples, Thrane et al. (2006) inferred uniformdistribution of 26Al in the inner protoplanetary disk, whereCAIs and chondrules probably formed, but the matter needsfurther investigation.

Chronometers based on long-lived (extant) nuclides donot require the assumption of initial homogeneousdistribution of the parent nuclide. Instead, the distribution canbe verified by measuring the present-day abundance of thenuclide (high-precision U isotope data for Allende have beenreported by Stirling et al. 2006). Recent improvements in theprecision of Pb isotopic dating allow us to tackle the questionof the CAI-chondrule formation interval. Using modern Pb-isotope dating techniques, we resolved the difference in thetime of formation between CAIs from the reduced CV3chondrite Efremovka (4567.2 ± 0.6 Ma) and chondrules fromthe CR2 carbonaceous chondrite Acfer 059 (4564.7 ± 0.6 Ma)(Amelin et al. 2002). However, application of this result to theestimation of the time difference between CAI and chondruleformation involves general uncertainty related to the origin ofchondrules and CAIs in different chondrite groups (e.g., Shuet al. 2001; Russell et al. 2005; Scott and Krot 2005) and thelack of Pb-isotopic dating of CAIs from CR chondrites orchondrules from CV chondrites. For example, according tothe X-wind model, CAIs and chondrules in a chondrite groupformed contemporaneously near the X-point and weresubsequently transported to their accretion region(s) byX-wind (Shu et al. 2001). These uncertainties can beeliminated if Pb-isotopic dates are determined fromchondrules and CAIs extracted from the same meteorite, or atleast from meteorites of the same group, probably originatingfrom the same parent body.

Here we present the results of Pb isotopic dating ofchondrules from the oxidized CV chondrite Allende. Duringthis study, we have found that complete removal of common Pb,a condition for precise and accurate Pb-isotopic dating, fromAllende chondrules is hard to achieve (here all Pb other than insitu accumulated radiogenic Pb is referred to as common Pb).To circumvent this problem, we have developed an alternativeapproach to chondrule age determination based on the use ofcombined acid leachate and leaching residue Pb isotopiccompositions. We also discuss the consequences of possibleage heterogeneity in the Allende chondrule population,suggested by a recent 26Al-26Mg study of bulk Allendechondrules (Bizzarro et al. 2004), for Pb isotopic dating.

ALTERATION OF CHONDRULES FROM CV CHONDRITES

CV chondrites are currently subdivided into the reduced(CVred) and two oxidized subgroups, Allende-like (CVoxA)and Bali-like (CVoxB) (McSween 1977; Weisberg and Prinz1998), which largely reflect their complex alteration history(Krot et al. 1995, 1998a, 1998b, 2004). Chondrules in theCVoxB chondrites (e.g., Kaba, Bali) experienced aqueousalteration resulting in the replacement of primary mineralphases (mostly mesostasis and Fe,Ni metal; low-Ca pyroxeneis only slightly altered; olivine is almost unaltered) bysecondary phyllosilicates, magnetite, Fe,Ni sulfides, Fe,Nicarbides, fayalite (Fa>90), salite-hedenbergite pyroxenes(Fs10–50Wo45–50), and andradite. Chondrules in the CVoxAchondrites, including Allende, experienced iron-alkalimetasomatic alteration resulting in replacement of theirprimary minerals by nepheline, sodalite, salite-hedenbergitepyroxenes, andradite, Fe,Ni sulfides, magnetite, Ni-richmetal, and ferrous olivine (Fa~40–50). Most of the opaquenodules and mesostasis of the Allende chondrules arereplaced by secondary minerals; replacement of low-Capyroxene by ferrous olivine and enrichment of olivine in FeOdecrease toward chondrule cores. The CVred chondritesEfremovka and Leoville experienced alteration similar to thatof CVoxA, but to a smaller degree (see Krot et al. 1998a,1998b, 2004 for details).

From the petrographic and mineralogical observationsand isotopic data, including 26Al-26Mg, 53Mn-53Cr, and 129I-129Xe chronology of secondary phases and their oxygenisotopic compositions, it is inferred that secondarymineralization in CV chondrites resulted from prolonged(>10 Myr) fluid-assisted thermal metamorphism postdatingaccretion of the CV parent asteroid, and the CV subgroupsrepresent its different lithologies (Krot et al. 2006 andreferences therein).

Physico-chemical conditions (pressure, temperature, pH,fO2, water/rock ratio) of the CV alteration remain poorlyconstrained, which partly reflects its complex, multistagenature (aqueous alteration followed by thermalmetamorphism) (e.g., Krot et al. 1998a, 1998b; Bonal et al.2006 and references therein). For example, the petrographictype assigned to Allende varies from 3.2 based on thethermoluminescence studies of CV chondrites (Guimon et al.1995), to 3.6, suggested by Bonal et al. (2006) on the basis of aRaman spectroscopy study of structural order of organicmatter in CV chondrites. From thermodynamic analysis ofsecondary mineralization in CV chondrites, Krot et al. (1998a)and Zolotov et al. (2006) inferred that aqueous alteration of CVchondrites occurred below ~350 °C, at a total pressure below100 bars and with a relatively low water/rock ratio (<0.2).Allende may have then experienced thermal metamorphismwith peak metamorphic temperatures of ~400–550 °C(McSween 1977; Blum et al. 1989; Weinbruch et al. 1994).

Pb isotopic age of the Allende chondrules 1323

PREVIOUS ISOTOPIC STUDIES

Allende is arguably the most extensively studiedmeteorite. CV chondrites contain larger and more abundantCAIs than other chondrites, and Allende is by far the mostplentiful CV chondrite, therefore Allende CAIs in particularhave received a lot of attention from cosmochemists. Thenumber of radiogenic isotope studies of Allende chondrulesis much smaller. Early U-Th-Pb studies by Chen and Tilton(1976) and Tatsumoto et al. (1976), in which all componentsof the Allende meteorite (CAIs, chondrules, and matrix)were analyzed, yielded Pb isotopic dates of 4565 ± 4 Ma,and 4553 ± 4 Ma, respectively, from combined isochronregressions including chondrules, CAIs and matrix.Chondrules, CAIs and other components from Allende werealso analyzed for U-Pb by Arden and Cressey (1984), butthese authors did not report any age calculations. Rubidium-Sr data for Allende chondrules, reported by Gray et al.(1973) and Tatsumoto et al. (1976), form scattered arrayssuggesting late re-distribution of Rb and/or Sr. A morerecent Rb-Sr study by Shimoda et al. (2005) indicates twostages of disturbance, the first episode at 4.41 ± 0.12 Ga (theauthors use a 87Rb decay constant of 1.402 × 10−11 yr−1

following Begemann et al. 2001), followed by latercontinuous or episodic processes. A recent high-precisionMC-ICP-MS (multicollector inductively coupled plasmamass spectrometry) study of 26Al-26Mg systematics of bulkAllende chondrules (Bizzarro et al. 2004) suggests thatformation of Allende chondrules continued for at least1.2 Myr and the oldest chondrules may have formedcontemporaneously with CAIs. We note that model ages ofthe Allende chondrules based on bulk Mg isotopemeasurements may represent ages of chondrule precursorsrather than crystallization ages of chondrules; the latterrequires knowledge of internal Al-Mg systematics ofchondrules, which are typically measured with less precisionin situ by secondary ion mass spectrometry (SIMS).Preliminary SIMS data on Mg isotopic compositions ofmineral phases in chondrules from primitive chondritessuggest initial 26Al/27Al ratios in chondrules less than 1.5 ×10−5 so far (Kita et al. 2005 and references therein). Anindependent evidence for the minimum age interval ofAllende chondrule formation of 1 Myr is determinedobtained from an 26Al-26Mg SIMS study of a compoundchondrule, made of two chondrules of different ages (Akakiet al. 2007).

ANALYTICAL PROCEDURES

We analyzed chondrules from two specimens of theAllende meteorite. For this study, we processed an ~30 gspecimen, completely covered with fusion crust, from themeteorite collection of the Royal Ontario Museum. Thesample was crushed and chondrules were separated as

described previously for the Richardton chondrite (Amelinet al. 2005). Additional chondrules were picked from thesample earlier processed at the Washington University.

Data reported here have been acquired in four analyticalsessions between March 2002 and May 2004. Samplewashing, chemistry and mass spectrometry varied betweenthe sessions. Detailed descriptions of analytical proceduresused in each session are summarized in the Electronic AnnexEA-1. All chondrules were washed in acids to remove looselybound common Pb. Throughout the text, acid-washedmaterials are referred to as “residues” or simply “chondrules,”whereas the material extracted to acidic solutions duringwashing is referred to as “leachate.”

Isochrons, weighted means and medians and theiruncertainties were calculated using Isoplot/Ex version 3.00(Ludwig 2003). All isochrons, weighted mean and medianerrors have 95% confidence intervals, unless indicatedotherwise.

RESULTS

U-Th-Pb Systematics

Pb isotopic data, Pb, U and Th concentrations, and 238U/206Pb* and 207Pb*/206Pb* model dates are presented inTable 1. A complete form of this table, which contains allisotopic ratios with errors and error correlations necessary forisochron calculations, is available online at http://meteoritics.org. Uranium and Pb concentrations in acid-washed chondrules are lower and more variable than theconcentrations reported for Allende chondrules analyzedwithout acid leaching by Chen and Tilton (1976) andTatsumoto et al. (1976). For set 4, where acid washes wereanalyzed along with residues, U and Pb concentrations beforewashing have been calculated as the sums of concentrations inboth washes and the residue. Uranium concentrationsbetween 14–22 ppb and total Pb concentrations between 69–150 ppb are more consistent than the concentrations in acid-washed chondrules, indicating evidence for substantial andvariable Pb and U loss during acid washing of chondrules.

Lead isotopic ratios in chondrules are variablyradiogenic, with measured 206Pb/204Pb between 19.5–268 (or19.3–658 after correction for fractionation, spike, andanalytical blank). These ratios are less radiogenic than thevalues obtained from chondrules in CR chondrite Acfer 059and H5 chondrite Richardton (Amelin et al. 2002, 2005) usingsimilar techniques. Elevated content of common Pb inAllende chondrules, even after acid leaching, is a concern inisochron dating (Amelin 2006).

A complete set of leachate and residue analyses obtainedfor eight groups of chondrules during analytical session 4allows comparison of Pb isotopic compositions on acid-soluble and insoluble portions of chondrules. Acid leachatescontain less radiogenic Pb than residues of the same fractions:


Tabl

e 1.

U-T

h-Pb

dat

a.

Frac

tion

Set

Num

ber o

f ch

ondr

ules

Wei

ght

(g)

U (ppm

)Th (p

pm)

Pb (ppm

)Pb (p

g)

206 P

b/

204 P

b (r

aw)

206 P

b/

204 P

b(to

tal)

204 P

b/

206 P

b(to

tal)

204 P

b/

206 P

b%

err

or

207 P

b/

206 P

b (to

tal)

207 P

b/

206 P

b %

err

or

238 U

/ 20

6 Pb*

date

(Ma)

207 P

b*/

206 P

b*

date

(Ma)

207 P

b*/

206 P

b*da

te e

rror

Res

idue

s1

11

0.00

208

0.00

97nm

0.04

1887

47.3

647

.70

0.02

096

0.79

0.72

000

0.15

453

645

70.5

3.2

21

10.

0025

70.

0071

nm0.

0283

7342

.15

42.2

70.

0236

60.

680.

7318

50.

17 4

423

4570

.03.

63

11

0.00

417

0.01

37nm

0.04

8520

243

.42

43.5

40.

0229

70.

460.

7288

50.

18 4

536

4570

.24.

04

11

0.00

550

0.01

58nm

0.04

7726

364

.95

65.4

30.

0152

80.

670.

6920

50.

19 4

483

4563

.53.

75

11

0.00

319

0.01

20nm

0.03

4611

089

.80

92.2

10.

0108

50.

840.

6738

50.

11 4

614

4568

.01.

96

21

0.00

154

0.00

58nm

0.02

0031

50.8

551

.88

0.01

927

1.5

0.70

711

0.22

465

945

55.4

3.7

72

10.

0065

70.

0033

nm0.

0146

9630

.37

30.6

60.

0326

10.

500.

7689

20.

11 4

569

4560

.22.

18

21

0.00

464

0.00

07nm

0.00

6530

19.4

719

.34

0.05

171

0.49

0.85

276

0.23

552

645

50.3

9.5

92

10.

0065

50.

0028

nm0.

0151

9929

.68

29.9

40.

0334

00.

800.

7697

80.

25 5

147

4551

.27.

110

21

0.00

339

0.00

71nm

0.02

1172

87.4

293

.66

0.01

068

3.3

0.67

062

0.23

479

145

61.6

1.8

112

20.

0059

60.

0047

nm0.

0170

101

64.1

266

.30

0.01

508

1.9

0.69

172

0.35

518

045

65.0

6.5

122

10.

0050

90.

0130

nm0.

0714

363

171.

418

2.2

0.00

549

2.8

0.64

879

0.11

473

345

65.3

1.1

132

20.

0058

60.

0056

nm0.

0190

111

79.5

683

.04

0.01

204

2.0

0.67

785

0.16

487

545

64.5

1.6

142

20.

0069

60.

0054

nm0.

0172

120

60.1

161

.72

0.01

620

1.2

0.69

749

0.14

459

045

67.1

1.7

152

60.

0048

30.

0072

nm0.

0285

138

80.6

684

.11

0.01

189

1.9

0.67

651

0.20

482

445

62.7

2.6

162

40.

0059

80.

0056

nm0.

0178

106

71.7

374

.43

0.01

344

1.7

0.68

339

0.16

479

845

62.5

1.8

172

40.

0100

70.

0060

nm0.

0224

225

57.8

558

.73

0.01

703

0.83

0.69

896

0.13

485

045

61.0

2.4

183

Mul

tiple

1)

0.00

600

nmnm

0.00

7746

45.9

648

.04

0.02

082

3.6

0.71

782

0.38

4566

.23.

619

3M

ultip

le 2

)0.

0090

0nm

nm0.

0088

7933

.89

34.4

00.

0290

71.

230.

7571

70.

1745

73.6

3.1

20a

3M

ultip

le 3

)0.

0150

0nm

nm0.

0126

189

27.9

828

.12

0.03

557

0.44

0.78

448

0.10

4568

.13.

120

c3

Mul

tiple

3)

0.01

500

nmnm

0.00

5886

132.

714

6.5

0.00

683

9.1

0.65

646

0.48

4569

.47.

621

412

0.01

441

0.00

510.

0205

0.01

7124

716

3.5

180.

70.

0055

37.

70.

6491

00.

31 5

396

4565

.63.

322

48

0.00

335

0.00

470.

0140

0.01

1539

207.

966

1.5

0.00

151

165

0.63

127

1.52

487

945

66.1

4.5

234

190.

0073

30.

0061

0.02

260.

0201

147

174.

520

8.6

0.00

479

14.1

0.64

680

0.44

546

245

68.1

3.4

244

210.

0089

30.

0053

0.02

100.

0177

158

112.

012

3.4

0.00

811

7.4

0.66

004

0.35

534

545

64.1

2.3

254

160.

0068

00.

0051

0.02

060.

0165

112

137.

416

4.2

0.00

609

14.2

0.65

227

0.51

533

645

67.3

2.4

264

90.

0075

80.

0068

0.03

000.

0291

221

268.

531

9.9

0.00

313

13.5

0.63

862

0.27

694

545

66.3

1.8

274

140.

0066

40.

0073

0.03

110.

0252

168

177.

620

8.8

0.00

479

12.6

0.64

588

0.37

560

445

65.9

2.0

284

200.

0227

10.

0056

0.02

490.

0219

498

154.

616

1.8

0.00

618

3.4

0.65

199

0.15

595

645

65.6

1.7

295

Mul

tiple

0.01

353

nmnm

0.01

2516

812

2.4

131.

20.

0076

24.

50.

6574

80.

22

4563

.11.

730

5M

ultip

le0.

0114

3nm

nm0.

0192

219

71.7

73.7

0.01

357

1.8

0.68

349

0.16

4561

.32.

031

5M

ultip

le0.

0060

5nm

nm0.

0088

5310

4.3

124.

70.

0080

212

.90.

6599

90.

5945

65.0

2.8

326

Mul

tiple

0.00

684

nmnm

0.01

3290

159.

621

5.6

0.00

464

26.2

0.64

468

0.73

4564

.63.

033

6M

ultip

le0.

0101

5nm

nm0.

0303

308

105.

311

1.1

0.00

900

4.1

0.66

386

0.22

4563

.52.

235

6M

ultip

le0.

0084

0nm

nm0.

0322

271

248.

029

2.0

0.00

342

12.8

0.63

979

0.28

4565

.91.

7

Aci

d le

acha

tes

21 H

NO

34

120.

0144

10.

0087

nm0.

0878

1265

14.9

414

.93

0.06

696

0.29

0.92

077

0.22

452

545

42.3

13.0

21 H

Cl

412

0.01

441

0.00

24nm

0.00

6898

69.4

594

.19

0.01

062

310.

6727

11.

78 4

281

4567

.77.

322

HN

O3

48

0.00

335

0.01

26nm

0.07

8226

222

.21

22.3

90.

0446

60.

940.

8161

40.

13 4

950

4531

.87.

322

HC

l4

80.

0033

50.

0033

nm0.

0607

203

24.8

325

.25

0.03

960

1.74

0.75

014

0.48

10,0

6643

60.9

9.8

23 H

NO

34

190.

0073

30.

0132

nm0.

0618

453

23.6

523

.83

0.04

197

0.71

0.81

038

0.10

3 4

151

4557

.74.

923

HC

l4

190.

0073

30.

0030

nm0.

0091

6674

.29

122.

820.

0081

458

0.66

164

2.68

484

245

67.7

11.0

24 H

NO

34

210.

0089

30.

0072

nm0.

0488

436

18.3

418

.34

0.05

453

0.60

0.86

740

0.15

439

945

59.0

9.0

24 H

Cl

421

0.00

893

0.00

10nm

0.00

2523

39.9

972

.17

0.01

386

960.

6945

56.

61 3

926

4587

.135

.6


25 H

NO

34

160.

0068

00.

0089

nm0.

0640

435

16.5

616

.52

0.06

054

0.48

0.89

524

0.18

409

245

62.9

8.2

25 H

Cl

416

0.00

680

0.00

17nm

0.00

5135

52.2

795

.93

0.01

042

850.

6753

74.

73 4

756

4576

.720

.126

HN

O3

49

0.00

758

0.01

00nm

0.04

3232

722

.48

22.6

40.

0441

70.

800.

8191

20.

12 3

855

4552

.96.

526

HC

l4

90.

0075

80.

0054

nm0.

0124

9484

.95

123.

540.

0080

938

0.66

120

1.76

417

545

67.1

6.5

27 H

NO

34

140.

0066

40.

0113

nm0.

0581

386

21.4

921

.60

0.04

630

0.52

0.83

179

0.10

412

545

65.3

5.4

27 H

Cl

414

0.00

664

0.00

31nm

0.00

8255

54.7

180

.62

0.01

240

440.

6834

12.

86 4

208

4574

.813

.228

HN

O3

420

0.02

271

nmnm

0.06

0013

6317

.38

17.4

00.

0574

86.

20.

8713

53.

97 1

799

4511

.220

0.5

28 H

Cl

420

0.02

271

0.00

17nm

0.00

4399

59.1

674

.23

0.01

347

220.

6857

41.

56 3

986

4568

.47.

4nm

= n

ot m

easu

red.

Num

bers

of c

hond

rule

s-m

ultip

le c

hond

rule

frac

tions

: 1 =

oliv

ine-

rich

light

-col

ored

; 2 =

dar

ker,

mos

tly p

yrox

ene-

rich;

3 =

dar

k w

ith su

lfide

s and

mat

rix o

verg

row

ths.

Wei

ghts

bef

ore

leac

hing

. U, T

h, a

nd P

b co

ncen

tratio

ns a

re c

alcu

late

d us

ing

thes

e w

eigh

ts.

Isot

opic

ratio

s den

oted

“ra

w”

are

mea

sure

d ra

tios w

ithou

t any

cor

rect

ions

.Is

otop

ic ra

tios d

enot

ed “

tota

l” a

re c

orre

cted

for f

ract

iona

tion,

bla

nk, a

nd sp

ike.

Erro

rs a

re 2

sigm

a of

the

mea

n.Is

otop

es m

arke

d w

ith a

ster

isk

are

“rad

ioge

nic”

: cor

rect

ed fo

r fra

ctio

natio

n, sp

ike,

bla

nk, a

nd p

rimor

dial

Pb

isot

opic

com

posi

tion

from

Tat

sum

oto

et a

l. (1

973)

as i

nitia

l Pb.

Age

s cal

cula

ted

usin

g th

e pr

imor

dial

Pb

isot

opic

com

posi

tion

from

Tat

sum

oto

et a

l. (1

973)

.“R

ho”

is e

rror

cor

rela

tion.

Tabl

e 1.

Con

tinue

d. U

-Th-

Pb d

ata.

Frac

tion

Set

Num

ber o

f ch

ondr

ules

Wei

ght

(g)

U (ppm

)Th (p

pm)

Pb (ppm

)Pb (p

g)

206 P

b/

204 P

b (r

aw)

206 P

b/

204 P

b(to

tal)

204 P

b/

206 P

b(to

tal)

204 P

b/

206 P

b%

err

or

207 P

b/

206 P

b (to

tal)

207 P

b/

206 P

b %

err

or

238 U

/ 20

6 Pb*

date

(Ma)

207 P

b*/

206 P

b*

date

(Ma)

207 P

b*/

206 P

b*da

te e

rror


nitric acid leachates (first washing cycle) have measured206Pb/204Pb between 14.9–23.7, hydrochloric leachates(second cycle) have measured 206Pb/204Pb between 24.8–84.9, whereas the residues contain Pb with 206Pb/204Pbbetween 112–268.

Residue U-Pb and Pb-Pb Isochrons

The results of the U-Pb and Pb-Pb isochron calculationsare shown in Table 2. We have calculated conventional Pb-Pband U-Pb isochrons and three-dimensional linear regressions(Ludwig 1998). U-Pb isochrons, both conventional and three-dimensional, show large scattering of points (MSWDbetween 31–571), suggesting that U-Pb systems weredisturbed by incoherent migration of U and Th in naturalenvironments and/or during laboratory treatment, and cannotbe used for age determination.

Pb-Pb isochrons show more coherent behavior. We use“inverse” isochrons, which provide direct reading of the datefrom the y-axis intercept, and have much smaller errorcorrelations than “normal” isochrons. Isochron dates forindividual analytical sessions are between 4563 ± 13 Ma and4568 ± 27 Ma, and agree with each other within error.Isochron regression of all residue analyses yields 4565.3 ±1.5 Ma (MSWD = 6.0). If analyses with low 206Pb/204Pb,potentially more affected by variations in common Pbcomposition, are excluded, then the data scattering decreases.The MSWD value decreases from 6.0 for the isochronincluding all analyses, to 0.60 (no scattering outside ofanalytical errors) for the isochron including only data pointswith 206Pb/204Pb > 200 (two of the isochrons are shown, asexamples, in Fig. 1).

This pattern is consistent with the presence of more thanone component in common Pb. The decrease in the datadispersion, however, is not matched by the decrease in theerror of the date, which remains, almost unchanged, between±1.5–2.5 Ma, and even increases for the regression based onthe most radiogenic data only (206Pb/204Pb > 200). Therelatively large errors of ±2.5 to ±4.2 Ma for the isochron withno excess scattering is produced by expansion of the isochronerror envelope when extrapolated to the y-axis. A more“precise” isochron date with the error of less than ±1 Ma canbe obtained by excluding the most deviant points, irrespectiveto their 206Pb/204Pb ratio, from the data set. However, we haveno a priori reasons for excluding these data points fromregression, and would consider such an exercise invalid, and“high precision” obtained this way unfounded. Therefore weneed to look for alternative ways of data treatment that mighthelp us to improve the precision of the dates.

Pb-Pb Model Dates

We have calculated weighted averages of single-stagemodel dates, based on primordial Pb of Tatsumoto et al.

(1973), for the same data sets as the Pb-Pb isochronregressions. The data are presented in Table 2. The average ofmodel dates for each data set agrees within error with theisochron date for the same set, and the averages of modeldates are more precise than the isochron dates. Agreementbetween model dates and isochron dates suggests that theisotopic composition of prevailing common Pb componentsin Allende chondrules is not very different from theprimordial Pb. For example, common Pb in the chondrulescan be dominated by Pb redistributed from the Allende matrix(206Pb/204Pb between 9.7–10.1) (Chen and Tilton 1976;Tatsumoto et al. 1976), which is only slightly more radiogenicthan the primordial Pb. However, there is no guarantee thatthe common Pb in chondrules is identical to the primordial Pbor to the matrix Pb, and that the isotopic composition of thechondrule common Pb is homogeneous. Therefore, there is noreason to assume that the model dates are necessarily accuratewithin their random errors. On the contrary, the lack ofreproducibility between the average model dates measured indifferent analytical sessions (using different leachingprocedures) suggests that at least two common Pbcomponents are present, and one is more easily removed byacid leaching than the other. Despite the agreement betweenisochron dates and model dates and apparently high precisionof the latter, model dates may be biased to an uncertaindegree. For the highest measured 206Pb/204Pb ratios, about200–300, the potential bias in model dates due to inaccuratecommon Pb assumption can be as large as 4–6 Ma, and aslarge as 15 Ma for a sample with 206Pb/204Pb = 100 if thecommon Pb varies between primordial and average modernterrestrial isotopic compositions (Amelin 2006). Even thedifference between matrix Pb and primordial Pb can cause asignificant bias for less radiogenic samples. Model dates aretherefore suitable only for reconnaissance age determination.

Residue—Leachate Isochrons

The precision of Pb-isotopic dating can be potentiallyimproved by including Pb isotopic composition of acidleachates in age calculations, using an approach similar to thesingle mineral Pb-Pb dating of Frei and Kamber (1995). If achondrule contains two or more common Pb components,then it can be expected that some of these components, forexample those associated with surface contamination or withmore soluble minerals, are more easily leached by acids thanthe others. Sequential acid leaching of such a chondrulewould produce a series of fractions with variable common Pbcontent and isotopic composition. The common Pbcomposition in any of these fractions can also differ fromcommon Pb in the residue: therefore, two-point internalresidue-leachate isochrons including any leaching fraction orcombined leachates might yield inaccurate dates. Data formultiple acid leachates from the same chondrule wouldproduce a scattered or curved array in the isochron diagram if


the common Pb-isotopic composition varies between theleaching steps. Co-linearity of the points, on the other hand,would be proof of the uniform common Pb composition andwould suggest that the date is accurate. If multiple leachatesare analyzed from several chondrules or chondrule groupsfrom the same meteorite, then internal residue-leachateisochrons without excess scattering would yield precise (andpresumably accurate) dates, whereas the dates from scatteredisochrons would be less precise. Calculation of the age ofchondrules as a weighted average of internal residue-leachateisochrons thus favors more accurate internal isochron dates.

Another method of age determination from residue-leachate isochrons is based on the assumption that the moresoluble common Pb component is removed by the firstleaching step, whereas the second leachate and the residuecontain only the second, relatively insoluble component.Two-point isochrons for “residue-second wash” pairs forindividual fractions would then give accurate and consistentdates. If this assumption is wrong, then the residue-secondwash isochron dates would be irreproducible.

The validity of both approaches will be confirmed ifinternal residue-leachate isochrons for several chondrulesgive consistent dates. The first method is more stringentbecause it has two tests for consistency: between leachingsteps and between chondrules. In contrast to isochron datingusing multiple chondrule residues combined in one isochron,internal residue-leachate isochrons do not require uniformcommon Pb composition among chondrules. This is animportant advantage if we study a chondrule population withvariable mineralogy and/or degree of alteration (which maybe the case for the Allende chondrules).

Two-point and three-point isochron dates for Allendechondrules analyzed during analytical session 4 are presentedin Table 3. Five out of eight three-point isochrons have no

dispersion outside of analytical errors (MSWD range between0.0024 and 0.85), whereas three other isochrons showsubstantial excess scattering. All eight three-point isochronsyield consistent dates (Fig. 2), with the weighted average of4566.6 ± 1.0 Ma.

Two-point isochron residue-first wash and residue-second isochron dates have more uniform precision, but theirweighted average of 4566.6 ± 1.0 Ma and 4566.8 ± 2.8 Ma,respectively, are identical to the average of three-pointisochron dates. The residue-second isochron dates are lessprecise, because most second leachates contained very littlePb and their analyses are relatively imprecise. All residue-leachate isochron dates and their averages agree with theisochron dates for residues only (Table 2). Agreementbetween the dates calculated using various methods residue-only isochrons, model dates, and leachate-residue isochronssupports the above suggestion that variation in common Pbcomposition between Allende chondrules is relatively small.

DISCUSSION

Advantages and Drawbacks of Acid Leaching in PbIsotopic Analysis of Chondrules

Analyzing acid-leached chondrules has two advantagescompared to analysis of untreated chondrules. Removal ofcommon Pb is the main virtue of acid leaching. There are tworeasons why reducing common Pb content to a negligiblelevel is important. First, it eliminates the major, nonanalyticaluncertainty in 207Pb/206Pb age calculations. The possibleinaccuracy of model dates and isochron dates (and theisochron date uncertainty due to excess scatter) decreases asthe ratio of radiogenic Pb to common Pb increases (Allègreet al. 1995; Amelin et al. 2005; Amelin 2006). The second

Fig. 1. Pb-Pb isochrons for acid-washed individual chondrules and chondrule fractions from Allende. a) Ten fractions containing mostradiogenic Pb with measured 206Pb/204Pb > 150. b) 27 fractions with measured 206Pb/204Pb > 50.


reason is that it is easier to recognize the patterns of multi-stage evolution in the systems containing radiogenic Pb butno common Pb. The importance of complete removal ofcommon Pb for precise and accurate Pb-isotopic dating canhardly be overestimated. Reduction of common Pb content isthe main reason of improved precision of meteorite ages inrecent Pb isotopic studies.

The second advantage of leaching is the possibility of Pbisotopic dating using internal residue-leachate isochrons.These isochrons can yield accurate dates for individualchondrules and are indispensable for studying chondrulepopulations with heterogeneous age and/or common Pbisotopic composition. For chondrule populations withuniform age but variable common Pb, precision of the agedetermination can be increased by averaging internalisochron dates for individual chondrules. In addition, residue-leachate isochrons can help, if remaining common Pb in theresidues compromises the precision of dating with aconventional “residue-only” isochron—a case described inthis paper.

The drawbacks of acid leaching are related to the loss ofuranium and radiogenic Pb. Incongruent extraction of U andradiogenic Pb during leaching steps (Fig. 3) makes itimpossible to study natural discordance in the U-Pb systemand to determine the timing of secondary processes usingconcordia diagram.

It is interesting that acid leaching caused disturbance ofthe U-Pb system in chondrules from CV3 chondrite Allende(this study), but chondrules from H5 chondrite Richardton,leached using a similar procedure, have concordant U-Pbsystems (Amelin et al. 2005). This difference in chondrule

behavior may be a result of more intensive recrystallization ofRichardton chondrules during high-temperaturemetamorphism.

Another downside of leaching is the reduction of the totalamount of radiogenic Pb available for analysis, with resultingdeterioration of precision. The distribution of Pb between twowashes and the residue, illustrated by Fig. 3b, shows that theresidues contain between 8–34% of the total Pb content, and66–92% of Pb is in the leachates. Of course, a large part of Pbextracted during acid leaching is common Pb, but the loss ofradiogenic Pb also occurs, as indicated by relativelyradiogenic Pb isotopic compositions in the second leachates.In order to compensate for the loss of radiogenic Pb and toachieve reasonable precision of analyses, we have to analyzefractions of several chondrules together rather than individualchondrules, thus losing the ability to detect potential agedifferences between the chondrules.

Although the drawbacks of leaching are important, theadvantages clearly exceed them. The drawbacks of leaching

Fig. 2. Weighted average of three-point internal Pb-Pb isochrons(HNO3 wash, HCl wash, residue) for each of eight fractions ofAllende chondrules (fractions 21–28 in Table 1). The data point forthe fraction 22, with large error bar, is omitted from the figure forclarity, but is included in the weighted average calculation. Omissionof this point from calculation has trivial effect on the weightedaverage value.

Fig. 3. Distribution of U (a) and total Pb (b) between HNO3 leachates(Wash 1), HCl leachates (Wash 2), and residues of eight chondrulefractions analyzed during analytical session 4. The distributions forU and Pb are not identical, suggesting decoupled extraction of U andPb. These results show that over 50% of U and Pb in chondrules arerelatively poorly bound and are extracted during the first leachingstep.


Tabl

e 2.

Sum

mar

y of

Pb-

Pb is

ochr

on a

nd m

odel

dat

e ca

lcul

atio

ns (r

esid

ues

only

).

Frac

tions

Num

ber o

f fr

actio

ns P

b-Pb

204 P

b/20

6 Pb-

207 P

b/20

6 Pb

isoc

hron

dat

e (M

a)M

SWD

Wei

ghte

d av

erag

e of

m

odel

dat

es ca

lcul

ated

us

ing

prim

ordi

al P

bM

SWD

Num

ber o

f fr

actio

ns U

-Pb

3-di

men

sion

al to

tal

Pb/U

isoc

hron

dat

e (M

a)In

itial

206 P

b/20

4 Pb

Initi

al20

7 Pb/

204 P

bM

SWD

Set 1

(fra

ctio

ns 1

–5)

5 4

563

± 13

2.5

4568

.3 ±

2.9

2.7

5 4

567

± 23

8.68

± 0

.73

9.95

± 0

.80

31Se

t 2 (f

ract

ions

6–1

7)12

4565

.6 ±

3.3

5.4

4563

.5 ±

1.8

8.2

1245

65.6

± 8

.612

.3 ±

1.2

12.1

1 ±

0.82

67Se

t 3 (f

ract

ions

18a

–20c

)4

4568

± 2

74.

445

69.6

± 5

.63.

8nd

ndnd

ndnd

Set 4

(fra

ctio

ns 2

1–28

)8

4567

.3 ±

2.4

0.74

4565

.94

± 0.

790.

858

457

4 ±

18 6

5 ±

16 4

4 ±

1057

Sets

5+6

(fra

ctio

ns)

645

66.9

± 1

.90.

4945

63.8

± 1

.82.

8nd

ndnd

ndnd

All

35 4

565.

3 ±

1.5

6.0

4564

.8 ±

1.0

6.5

25 4

564

± 19

11.0

± 1

.911

.3 ±

1.3

503

206 P

b/20

4 Pb_

tota

l>40

3045

65.2

± 1

.84.

845

64.8

7 ±

0.90

4.9

22 4

563

± 29

11.1

± 2

.211

.4 ±

1.8

571

206 P

b/20

4 Pb_

tota

l>50

2645

66.6

± 1

.73.

445

64.6

5 ±

0.89

4.5

19 4

565

± 30

16.1

± 3

.814

.4 ±

2.7

485

206 P

b/20

4 Pb_

tota

l>70

2145

67.1

± 1

.62.

245

64.7

5 ±

0.85

3.5

14 4

570

± 35

20.6

± 6

.217

.1 ±

4.3

519

206 P

b/20

4 Pb_

tota

l>10

015

4566

.8 ±

1.6

0.86

4565

.32

± 0.

671.

49

457

3 ±

21 5

1 ±

18 3

6 ±

1117

220

6 Pb/

204 P

b_to

tal>

150

1045

66.2

± 2

.50.

6245

65.8

1 ±

0.60

0.58

8 4

573

± 22

52

± 20

37

± 13

191

206 P

b/20

4 Pb_

tota

l>20

06

4565

.8 ±

4.2

0.60

4566

.07

± 0.

910.

504

457

1 ±

1811

1 ±

69 7

4 ±

4331

MSW

D =

mea

n sq

uare

of w

eigh

ted

devi

ates

.

Tabl

e 3.

Sum

mar

y of

resi

due-

was

h Pb

-Pb

isoc

hron

cal

cula

tions

for i

ndiv

idua

l fra

ctio

ns.

Thre

e-po

int i

soch

ron

Res

idue

+ H

NO

3 was

hR

esid

ue +

HC

l was

h

Frac

tion

206/

204

resi

due

207*

/206

* m

odel

dat

e(M

a)20

7*/2

06*

mod

el d

ate

erro

r D

ate

2σ e

rror

(Ma)

MSW

D D

ate

2σ e

rror

(Ma)

Dat

e2σ

err

or(M

a)

2118

145

65.6

3.3

4566

.9 3

.10.

5845

66.4

3.4

4563

.011

2266

245

66.1

4.5

3348

.066

,000

1045

4566

.73.

545

70.8

4.7

2320

945

68.1

3.4

4568

.9 3

.40.

0024

4568

.93.

645

69.0

1624

123

4564

.12.

345

64.6

2.6

1.50

4564

.52.

645

35.0

110

2516

445

67.3

2.4

4567

.6 2

.50.

8545

67.5

2.5

4555

.051

2632

045

66.3

1.8

4567

.0 1

.80.

3545

66.9

1.8

4565

.84.

927

209

4565

.92.

045

66.0

2.2

1.7

4565

.92.

245

61.0

1228

162

4565

.61.

745

64.7

5.5

0.68

4569

.010

4563

.56.

9

Wei

ghte

d m

ean

4565

.94

4566

.63

4566

.58

4566

.80

2-si

gma%

err.

of m

ean

0.80

0.98

0.96

2.80

MSW

D0.

850.

830.

790.

86Pr

obab

ility

of f

it0.

550.

560.

600.

5420

7 Pb*

/206 P

b* m

odel

dat

es a

re c

alcu

late

d us

ing

the

prim

ordi

al P

b is

otop

ic c

ompo

sitio

n fr

om T

atsu

mot

o et

al.

(197

3).

MSW

D =

mea

n sq

uare

of w

eigh

ted

devi

ates

.Pr

obab

ility

of f

it is

the

prob

abili

ty th

at, i

f the

onl

y re

ason

for s

catte

r is t

he a

naly

tical

err

ors a

ssig

ned

to th

e da

ta p

oint

s, th

e sc

atte

r of t

he d

ata

poin

ts w

ill e

xcee

d th

e am

ount

obs

erve

d fo

r the

dat

a (L

udw

ig20

03).


can be minimized by developing more selective leachingprocedures, which would extract common Pb without movingradiogenic Pb and U. This can be achieved by a systematicstudy of the effects of acid leaching on chondrule mineralogyand chemical composition, currently underway.

Residue-Leachate Isochrons—One Possible Solution forthe Deficiency of Conventional Isochron Model

The “conventional” Pb-Pb isochron approach (forexample, linear fitting using the algorithm of York [1969],with error propagation using the maximum-likelihoodestimation algorithm of Titterington and Halliday [1979]),such as used in Isoplot) has a fundamental flaw: anunfavorable shape of the isochron error envelope, producedby currently available isochron regression models, results inmany cases in very inefficient use of analytical data, and inunreasonably large isochron date errors. An isochron basedon data points with low but almost invariable 204Pb/206Pb cangive a very imprecise y-axis intercept (i.e., 207Pb/206Pb date),whereas an addition of one or more much less radiogenicpoint radically changes the shape of the error envelope andincreases precision of the isochron date. With the currentisochron regression models, it often appears advantageous toinclude unradiogenic data points into regressions. Thisconclusion may be counterintuitive, but it is confirmed bymany regressions of both simulated and observational datasets, and by considerations of isochron formalism. In the caseof a small spread of the 204Pb/206Pb ratios, the isochron datecan be many times less precise than the model date for any ofthe points, even for a true isochron with no excess scatter. Thecause of this problem is that the assumptions of the isochronmodels do not reflect the geochemical reality of the U-Pbisotopic systems in meteorites.

There are three possible solutions to this problem. Thefirst is separating and analyzing fractions that contain nocommon Pb other than analytical blank. Such data do notrequire isochron regression—the age is calculated as aweighted average of 207Pb/206Pb dates of individual fractions.The validity of the age is further verified by checkingconcordance of the U-Pb isotopic systems. This is the bestapproach, both straightforward and model-independent,identical to the approach used in terrestrial U-Pb (e.g., zircon)geochronology in the cases where high precision andaccuracy are required, for example, in time scale studies.Complete removal of common Pb can be achieved for somematerials, for example, CAIs and angritic pyroxenes (Amelin2007), but unfortunately this is not possible for chondruleswith currently available techniques.

If fractions free from common Pb cannot be obtained,then the other two approaches can be used to optimize theshape of the isochron error envelope and to use the precisionof analyses more efficiently. One approach is developing ageochemically realistic isochron regression model, which

would account for the possible presence of two or morecommon Pb components in meteoritic materials. In such amodel, the weight of analysis in isochron regression woulddepend on the ratio of radiogenic Pb to common Pb, as well ason the analytical error. One of the authors (Y. A.) has recentlyinitiated a project aimed at development of such a model.

The final approach is using leachate analyses to constrainthe common Pb isotopic composition in the studied materials.The rationale of using three-point internal isochronsconsisting of a residue analysis and two step leachate analyseswas discussed by Krot et al. (2005), and in the Residue-Leachate Isochrons section. Co-linear position of residue andtwo (or more) leachate data points in the isochron diagramindicates that the isotopic composition of common Pb in allthree fractions is identical within error, i.e., that the three-point residue-leachate isochron is a true isochron. If theassumption of the uniform isotopic composition of commonPb, released during leaching, is not satisfied, then we canexpect scattering both among the data points in the internalisochron, and among internal isochron dates for variousfractions. These scattered data receive low weights in thecalculations of weighted averages of the dates, and arediscriminated against unscattered, and therefore more precise,internal isochron dates.

Pb-Pb Isochrons for a Chondrule Population withHeterogeneous Age

An assumption that the samples combined in an isochronregression formed simultaneously is a basic condition ofconstructing an isochron. 26Al-26Mg dating of bulk Allendechondrules (Bizzarro et al. 2004) has revealed variations inthe initial 26Al/27Al ratio, which correspond to the range ofages of 1.2 Ma. A possible bias of Pb-Pb isochron dates forthe Allende chondrules due to age heterogeneity has beenexplored by applying a distribution of 26Al-26Mg agesreported by Bizzarro et al. (2004) to a simulated Pb-Pbisochron data set (adapted from Amelin et al. 2005) with orwithout random scatter caused by analytical uncertainties. Forconsistency with real chondrule analyses, we used model Pb-Pb isochrons in 207Pb/206Pb versus 204Pb/206Pb coordinates.Each model isochron consists of 11 data points with eachpoint being a mixture between primordial Pb and radiogenicPb, with the fraction of primordial Pb between 0.5% and 10%.Errors assigned to each point are similar to the typical errorsin thermal ionization mass spectrometry (TIMS) analysis ofsmall fractions of Pb with external normalization: a fixederror of 0.08% is assigned to 207Pb/206Pb, whereas the error of204Pb/206Pb varies between 4.3% for the most radiogenicpoint with 206Pb/204Pb = 1861 and 1.2% for the leastradiogenic point with 206Pb/204Pb = 143, due to changingcounting statistics on 204Pb (the error is inversely proportionalto the square root of 204Pb abundance). Error correlationbetween 204Pb/206Pb and 207Pb/206Pb is 0.9.


Model isochron calculations are shown in Table 4.Model 1 represents an “ideal” isochron calculated with fixedradiogenic Pb isotopic composition corresponding to4567.2 Ma (the age of the CV CAIs) (Amelin et al. 2002). Theerror of the isochron date of 0.47 Ma is propagated from theassigned errors.

In model 2, radiogenic Pb isotopic compositions aremodified to match the age differences between the CV CAIsand Allende chondrules found by Bizzarro et al. (2004). Theaverage of 26Al-26Mg chondrule ages, 0.52 Ma after CAIs, isconsidered the true average chondrule age. The results of Pbisochron regressions in this model depend on whether theages of chondrules are correlated with the 206Pb/204Pb ratio. Ifwe apply the 26Al-26Mg chondrule ages in the same sequenceas they are presented in the Bizzarro et al. (2004) paper(presumably random sequence, run 1), then the isochron dateagrees, within error, with the true average chondrule age. Ifthe 26Al-26Mg chondrule ages positively correlate with the206Pb/204Pb ratio, then the isochron date is too old, and vice-versa (runs 2 and 3, denoted “ordered up” and “ordereddown,” respectively). Model isochrons for “ordered”chondrule ages (runs 2 and 3) have smaller excess scatteringthan the “random sequence” isochron.

In model 3, random scattering, simulating mass biasvariations in Pb isotopic analysis, is applied to 207Pb/206Pbratios (the effect on 204Pb/206Pb is negligible). Five runs withdifferent sets of random numbers produced isochrons with theages within error of 4567.2 Ma, and probabilities of fitbetween 0.32–0.86.

In model 4, mass bias variations (with the same sets ofrandom numbers as in model 3) and age variations (as inmodel 2) are applied together. This superposition of agevariations and mass bias variations produces isochrons withincreased scattering compared to the isochrons with only onesource of variations. Combined random age variations andmass bias variations (runs 1_1 to 5_1) yield isochrons withthe largest scattering and low probability of fit between0.006–0.13 and thus lower precision. Combined ordered agevariations and mass bias variations (runs 1_2 to 5_3) produceless scattered (and thus more precise) isochrons with biaseddates.

The models described above show that the effect of agevariations in a chondrule population on Pb-Pb dating dependson the range of age variations and precision of Pb-isotopicanalysis, and, more importantly, on possible correlationbetween chondrule chemistry and age. If 206Pb/204Pb ratios

Table 4. Model Pb-Pb isochron calculations for Allende chondrules. Model description

Model RunMass bias variations

Age variations

Isochron age(Ma)

Error(Ma) MSWD

Probability of fit

1 1 No No 4567.20 0.47 0.00 1.000

2 1 No Random 4566.23 0.47 1.16 0.3102 2 No Ordered up 4567.43 0.48 0.10 1.0002 3 No Ordered down 4566.00 0.47 0.19 0.995

3 1 Yes No 4567.07 0.47 1.16 0.3203 2 Yes No 4567.26 0.47 0.60 0.790 3 3 Yes No 4566.77 0.47 0.53 0.8603 4 Yes No 4567.16 0.47 0.60 0.8003 5 Yes No 4567.31 0.47 1.06 0.390

4_1 1_1 Yes Random 4566.03 0.81 2.60 0.0064_1 2_1 Yes Random 4566.27 0.68 1.70 0.0924_1 3_1 Yes Random 4565.87 0.74 2.20 0.0224_1 4_1 Yes Random 4566.20 0.63 1.50 0.1304_1 5_1 Yes Random 4566.25 0.75 2.40 0.010

4_2 1_2 Yes Ordered up 4567.31 0.48 0.84 0.5804_2 2_2 Yes Ordered up 4567.49 0.48 0.59 0.8004_2 3_2 Yes Ordered up 4567.00 0.47 0.44 0.9104_2 4_2 Yes Ordered up 4567.39 0.48 0.80 0.6204_2 5_2 Yes Ordered up 4567.55 0.48 1.02 0.420

4_3 1_3 Yes Ordered down 4565.70 0.60 1.80 0.0694_3 2_3 Yes Ordered down 4566.06 0.47 0.97 0.4604_3 3_3 Yes Ordered down 4565.57 0.47 0.64 0.7704_3 4_3 Yes Ordered down 4565.96 0.47 0.93 0.5004_3 5_3 Yes Ordered down 4565.92 0.60 1.60 0.110MSWD = mean square of weighted deviates.Probability of fit is the probability that, if the only reason for scatter is the analytical errors assigned to the data points, the scatter of the data points will

exceed the amount observed for the data.


and chondrule ages are not correlated, then age variations,combined with analytical dispersion, can produce a scatteredand less precise isochron. If, however, chondrule ages andchondrule chemistry (and thus 206Pb/204Pb) are correlated oranti-correlated, then the isochron date can be biased. Thisproblem can be solved in several ways. Getting highlyradiogenic Pb (ideally, pure radiogenic Pb) and 26Al-26Mgand/or 53Mn-53Cr ages for the same chondrules would help todistinguish variations in chondrule formation ages frompossible multi-stage evolution. Another possibility is to dateeach chondrule using an internal residue-leachate isochron.The 26Al-26Mg and 53Mn-53Cr dating can be used in this caseas well, either by analysis of separate unleached fragments ofchondrules, or by re-combining column washes from Pb-isotopic separations of residues and leachates.

The Age of Allende Chondrules

Ages of eight individual groups of Allende chondrules(fractions 21–28), determined with leachate-residueisochrons, agree within error. We can therefore regard our setof dates as homogeneous, and average individual dates. Weconsider the weighted average of three-point residue-leachateisochrons of 4566.6 ± 1.0 Ma as the best estimate for theaverage age of Allende chondrules. All other estimates, fromtwo-point internal isochrons and from combined residue-onlyisochrons, agree with this number.

The agreement between Pb-isotopic ages of individualgroups of chondrules may seem to contradict the agevariability among Allende chondrules demonstrated by thebulk 26Al-26Mg data (Bizzarro et al. 2004). In fact there is nodisagreement between these data sets. Each of our fractionsanalyzed with step leaching was composed of multiplechondrules (between 8 and 21), and thus the age variations areaveraged. In addition, the entire range of age variationsresolved with the 26Al-26Mg method is 1.2 Ma, smaller thanthe confidence interval of the most precise internal Pb-Pbisochron. The relative ages after CAI formation, measuredwith 26Al-26Mg and Pb-Pb, agree very well: the average ofeleven 26Al-26Mg chondrule ages is 0.52 Ma after CAIs,whereas the average of eight Pb-isotopic ages of 4566.6 Mapost-dates CAIs by 0.6 Ma. This agreement confirms earliersuggestions that 26Al-26Mg and Pb-Pb chronometers giveconsistent readings of time in the early solar system (Zinnerand Göpel 2002; Amelin et al. 2002).

The Time Span of Chondrule Formation

According to Pb-Pb and 26Al-26Mg systematics, Allendechondrules are the oldest currently known chondrules. Theirformation might have started contemporaneously with the CVCAI formation or shortly after, and lasted for 1–2 millionyears. Chondrules in Allende are distinctly older than those inCR (4564.7 ± 0.6 Ma) and CB chondrites (4562.7 ± 0.5 Ma),

which were dated by Pb-Pb systematics as well (Amelin et al.2002; Krot et al. 2005). These observations indicate thatchondrules formed during several episodes (minimum twoepisodes, if we exclude CB meteorites, or minimum threeepisodes if we consider them), or possibly continuouslybetween ~4567–4562 Ma. We note, however, that chondrulesin CB chondrites may have formed from a vapor-melt plumeproduced by a giant impact between planetary embryos afterdust in the protoplanetary disk had largely dissipated (Krotet al. 2005), and hence may not be considered as typicalproducts of the protoplanetary disk.

The timing of formation of Allende chondrules overlapswith the timing of formation of the oldest basaltic crust andmetal-silicate segregation on differentiated asteroids—thesources of eucrites and angrites, as indicated by Pb-Pb ages ofthe angrites Sahara 99555 (between 4566.18 ± 0.14 Ma[Baker et al. 2005] and 4564.41 ± 0.65 Ma [Amelin 2007]; theage needs to be verified) and D’Orbigny of 4564.48 ±0.24 Ma (Amelin 2007), and eucrite Asuka-881394 of4566.52 ± 0.33 Ma (Wadhwa et al. 2005; Amelin et al. 2006),as well as 182Hf-182W ages of magmatic irons relative to theAllende CAIs (Kleine et al. 2005; Markowski et al. 2006).Assuming that 26Al was the major heating source ofdifferentiated asteroids (consistent with the lack of evidencefor presence of 60Fe in these asteroids) (Bizzarro et al. 2007),theoretical modeling suggests that these asteroids must haveaccreted within 0.7 Myr after formation of the CV CAIs(Bizzarro et al. 2005). Since chondrule formation lasted overa period of at least 4–5 Myr, the latest generations ofchondrules postdate accretion and early differentiation ofsome asteroids. However, the old Al-Mg and Pb-Pb ages ofthe Allende chondrules and CAIs may provide an indirectevidence that the early differentiated asteroids could haveconsisted of typical chondritic components as well.

It has been recently concluded that desegregation of earlydifferentiated asteroids was a common phenomena in theprotoplanetary disk (e.g., Asphaug et al. 2006; Yang et al.2006). As a result, one may expect that fragments ofdifferentiated or extensively thermally metamorphosedasteroids could have been present among chondruleprecursors. Reported discoveries of annealed differentiatedclasts inside chondrules (Libourel et al. 2006; Sokol andBischoff 2006) may support this hypothesis.

Based on the emerging picture of early accretion (i.e.,within ~1 Myr of CAI formation) of differentiatedplanetesimals and the existence of undifferentiated chondriteparent asteroids, it seems inevitable that the latter accretedlate (Kleine et al. 2005; Bizzarro et al. 2005; Baker et al.2005). This is consistent with the presence of relatively youngchondrules in primitive chondrites (e.g., Kunihiro et al.2004). How primitive materials such as presolar grains, CAIs,chondrules, and low-temperature minerals found in chondritematrices could survive a few million years in theprotoplanetary disk (i.e., until they were accreted into


chondrite parent bodies) alongside the accreting anddifferentiating asteroids is still unclear. A betterunderstanding of the dynamics within a protoplanetary diskwould help to clarify this.

CONCLUSIONS

1. Allende chondrules leached in acids contain variablyradiogenic Pb with measured 206Pb/204Pb ratios between19.5–268 (or 19.3–658 after correction for fractionation,spike, and blank). These ratios are less radiogenic than insome previously studied chondrules in other meteorites.Pb-Pb isochron regression for ten most radiogenicanalyses with 206Pb/204Pb > 150 yielded the date of4566.2 ± 2.5 Ma.

2. Acid leaching is a crucial step in Pb-isotopic dating ofchondrules.

3. Acid leachates have been analyzed, along with residuesafter leaching, from eight chondrule fractions. Internalresidue-leachate isochrons yielded consistent dates witha weighted average of 4566.6 ± 1.0 Ma, our best estimatefor the age of Allende chondrules. These are the oldestchondrules studied so far.

4. Measurements of the time interval between formation ofCAIs and Allende chondrules using Pb isotopes (thisstudy) and 26Al-26Mg (Bizzarro et al. 2004) agree verywell.

5. Modeling Pb isochron for a set of chondrules withvariable ages, as suggested for Allende chondrules by26Al-26Mg data, shows that the Pb-Pb isochron date isaccurate, even if less precise, if there is no correlationbetween 206Pb/204Pb and the age for individualchondrules. The isochron date may be biased if 206Pb/204Pb and the age are correlated. This problem does notexist if the age is determined from internal residue-leachate isochrons.

6. Formation of Allende chondrules may have startedsimultaneously with, or shortly after formation of the CVCAIs, and possibly overlapped in time with formation ofthe oldest basaltic crust and iron cores of differentiatedasteroids.

7. The entire period of chondrule formation continued forabout 4–5 Myr, and included at least three discreteepisodes, or possibly was continuous. There wereprobably a variety of processes in the early solar systemthat we now call “chondrule formation.”

Acknowledgments–The sample of Allende was provided bythe Royal Ontario Museum. Additional chondrules werepicked from the sample earlier processed at the WashingtonUniversity (courtesy O. Pravdivtseva). Reviews byM. Bizzarro, C. Göpel, and N. Kita, and editorial commentsby C. Floss, helped to improve the paper. Analytical work atthe Jack Satterly Lab and the Geological Survey of Canada

was supported by NSERC research grant to Y. A., andCanadian Space Agency contract 9F007-010128/001/SR.This work was supported by NASA grants NAG5-10610 (A.N. Krot, P. I.), and NAG5-4212 (K. Keil, P. I.). This isHawai‘i Institute of Geophysics and Planetology publicationno. 1494 and School of Ocean and Earth Science andTechnology publication no. 7162.

Editorial Handling—Dr. Christine Floss

REFERENCES

Akaki T., Nakamura T., Noguchi T., and Tsuchiyama A. 2007.Multiple formation of chondrules in the early solar system:Chronology of a compound Al-rich chondrule. The AstrophysicalJournal 656:L29–L32.

Allègre C. J., Manhès G., and Göpel C. 1995. The age of the Earth.Geochimica et Cosmochimica Acta 59:1445–1456.

Amelin Y. 2006. The prospect of high-precision Pb isotopic dating ofmeteorites. Meteoritics & Planetary Science 41:7–17.

Amelin Y. 2007. The ages of angrites (abstract #1669). 38th Lunarand Planetary Science Conference. CD-ROM.

Amelin Y., Krot A. N., Hutcheon I. D., and Ulyanov A. A. 2002. Leadisotopic ages of chondrules and calcium-aluminum-richinclusions. Science 297:1678–1683.

Amelin Y., Ghosh A., and Rotenberg E. 2005. Unraveling theevolution of chondrite parent asteroids by precise U-Pb datingand thermal modeling. Geochimica et Cosmochimica Acta 69:505–518.

Amelin Y., Wadhwa M., and Lugmair G. 2006. Pb-isotopic dating ofmeteorites using 202Pb-205Pb double-spike: Comparison withother high-resolution chronometers (abstract #1970). 37th Lunarand Planetary Science Conference. CD-ROM.

Arden J. W. and Cressey G. 1984. Thallium and lead in the AllendeC3V carbonaceous chondrite—A study of the matrix phase.Geochimica et Cosmochimica Acta 48:1899–1912.

Asphaug E., Agnor C. B., and Williams Q. 2006. Hit-and-runplanetary collisions. Nature 439:155–160.

Baker J., Peate D., Waight T., and Meyzen C. 2004. Pb isotopeanalysis of standards and samples using a 207Pb-204Pb doublespike and thallium to correct for mass bias with double-focusingMC-ICP-MS. Chemical Geology 211:275–303.

Baker J., Bizzarro M., Wittig N., Connelly J., and Haack H. 2005.Early planetesimal melting from an age of 4.5662 Gyr fordifferentiated meteorites. Nature 436:1127–131.

Begemann F., Ludwig K. R., Lugmair G. W., Min K., Nyquist L. E.,Patchett P. J., Renne P. R., Shih C.-Y., Villa I. M., and WalkerR. J. 2001. Call for an improved set of decay constants forgeochronological use. Geochimica et Cosmochimica Acta 65:111–121.

Bizzarro M., Baker J. A., and Haack H. 2004. Mg isotope evidencefor contemporaneous formation of chondrules and refractoryinclusions. Nature 431:275–278.

Bizzarro M., Baker J. A., Haack H., and Lundgaard K. L. 2005. Rapidtime scales for accretion and melting of differentiatedplanetesimals inferred from 26Al-26Mg chronometry. TheAstrophysical Journal 632:L41–L44.

Bizzarro M., Ulfbeck D., Trinquier A., Thrane K., Connelly J. N., andMeyer B. S. 2007. Evidence for a late supernova injection of 60Feinto the protoplanetary disk. Science 316:1178–1181.

Blum J. D., Wasserburg G. J., Hutcheon I. D., Beckett J. R., andStolper E. M. 1989. Origin of opaque assemblages in CV3


meteorites: Implications for nebular and planetary processes.Geochimica et Cosmochimica Acta 53:543–556.

Bonal L., Quirico E., and Bourot-Denise M. 2004. Petrologic type ofCV3 chondrites as revealed by Raman spectroscopy of organicmatter (abstract #1562). 35th Lunar and Planetary ScienceConference. CD-ROM.

Chen J. H. and Tilton G. R. 1976. Isotopic lead investigations of theAllende carbonaceous chondrite. Geochimica et CosmochimicaActa 40:635–643.

Chen J. H. and Wasserburg G. J. 1981. The isotopic composition ofuranium and lead in Allende inclusions and meteoriticphosphates. Earth and Planetary Science Letters 52:1–15.

Chaussidon M., Robert F., and McKeegan K. D. 2006. Li and Bisotopic variations in an Allende CAI: Evidence for the in situdecay of short-lived 10Be and for the possible presence of theshort-lived nuclide 7Be in the early solar system. Geochimica etCosmochimica Acta 70:224–245.

Frei R. and Kamber B. S. 1995. Single mineral Pb-Pb dating. Earthand Planetary Science Letters 129:261–268.

Gerstenberger H. and Haase G. 1997. A highly effective emittersubstance for mass spectrometric Pb isotope ratiodeterminations. Chemical Geology 136:309–312.

Goswami J. N., Marhas K. K., Chaussidon M., Gounelle M., andMeyer B. S. 2005. Origin of short-lived radionuclides in the earlysolar system. In Chondrites and the protoplanetary disk, editedby Krot A. N., Scott E. R. D., and Reipurth B. San Francisco:Astronomical Society of the Pacific. pp. 485–527.

Gray C. M., Papanastassiou D. A., and Wasserburg G. J. 1973. Theidentification of early condensates from the solar nebula. Icarus20:213–239.

Guimon R. K., Symes S. J. K., Sears D. W. G., and Benoit P. H. 1995.Chemical and physical studies of type 3 chondrites XII: Themetamorphic history of CV chondrites and their components.Meteoritics & Planetary Science 30:704–714.

Ireland T. R. 1990. Presolar isotopic and chemical signatures inhibonite-bearing refractory inclusions from the Murchisoncarbonaceous chondrite. Geochimica et Cosmochimica Acta 54:3219–3237.

Jones R. H., Grossman G. N., and Rubin A. E. 2005. Chemical,mineralogical, and isotopic properties of chondrules: Clues totheir origins. In Chondrites and the protoplanetary disk, editedby Krot A. N., Scott E. R. D., and Reipurth B. San Francisco:Astronomical Society of the Pacific. pp. 251–286.

Kita N. T., Huss G. R., Tachibana S., Amelin Y., Nyquist L. E., andHutcheon I. D. 2005. Constraints on the origin of chondrules andCAIs from short-lived and long-lived radionuclides. InChondrules and the protoplanetary disk, edited by Krot A. N.,Scott E. R. D., and Reipurth B. San Francisco: AstronomicalSociety of the Pacific. pp. 558–587.

Kleine T., Mezger K., Palme H., Scherer E., and Münker C. 2006.Early core formation in asteroids and late accretion of chondriteparent bodies: Evidence from 182Hf-182W in CAIs, metal-richchondrites, and iron meteorites. Geochimica et CosmochimicaActa 69:5805–5818.

Krot A. N., Scott E. R. D., and Zolensky M. E. 1995. Mineralogic andchemical variations among CV3 chondrites and theircomponents: Nebular and asteroidal processing. Meteoritics 30:748–775.

Krot A. N., Petaev M. I., Scott E. R. D., Choi B.-G., Zolensky M. E.,and Keil K. 1998a. Progressive alteration in CV3 chondrites:More evidence for asteroidal alteration. Meteoritics & PlanetaryScience 33:1065–1085.

Krot A. N., Zolensky M. E., Keil K., Scott E. R. D., andNakamura K. 1998b. Secondary Ca-Fe-rich minerals in theBali-like and Allende-like oxidized CV3 chondrites and

Allende dark inclusions. Meteoritics & Planetary Science 33:623–645.

Krot A. N., Petaev M. I., and Bland P. A. 2004. Multiple formationmechanisms of ferrous olivine in CV3 carbonaceous chondritesduring fluid-assisted metamorphism. Antarctic MeteoriteResearch 17:154–172.

Krot A. N., Amelin Y., Cassen P., and Meibom A. 2005. Youngchondrules in CB chondrites from a giant impact in the earlysolar system. Nature 436:989–992.

Krot A. N., Hutcheon I. D., Brearley A. J., Pravdivtseva O. V., PetaevM. I., and Hohenberg C. M. 2006. Time scales and settings foralteration of chondritic meteorites. In Meteorites and the earlysolar system II, edited by Lauretta D. S. and McSween H. Y., Jr.Tucson, Arizona: The University of Arizona Press. pp. 525–555.

Kunihiro T., Rubin A. E., McKeegan K. D., and Wasson J. T. 2004.Initial 26Al/27Al in carbonaceous-chondrite chondrules: Too little26Al to melt asteroids. Geochimica et Cosmochimica Acta 68:2947–2957.

Libourel G., Krot A. N., and Chaussidon M. 2006. Olivines inmagnesian porphyritic chondrules: Mantle material of earliergenerations of differentiated planetesimals (abstract)?Meteoritics & Planetary Science 41:A106.

Ludwig K. R. 1998. On the treatment of concordant uranium-leadages. Geochimica et Cosmochimica Acta 62:665–676.

Ludwig K. R. 2003. Isoplot/Ex version 3.00, A GeochronologicalToolkit for Microsoft Excel. Berkeley: Geochronology Center.Special Publication 4. 71 p.

Markowski A., Leya I., Quitté G., Ammon K., Halliday A. N., andWieler R. 2006. Correlated helium-3 and tungsten isotopes iniron meteorites: Quantitative cosmogenic correlations andplanetesimal formation times. Earth and Planetary ScienceLetters 250:104–115.

McKeegan K. D., Chaussidon M., and Robert F. 2000. Incorporationof short-lived 10Be in a calcium-aluminum-rich inclusion fromthe Allende meteorite. Science 289:1334–1337.

McSween H. Y., Jr. 1977. Petrographic variations amongcarbonaceous chondrites of the Vigarano type. Geochimica etCosmochimica Acta 41:1777–1790.

Rock N. M. S., Webb J. A., McNaughton N. J., and Bell G. D. 1987.Nonparametric estimation of averages and error for small data-sets in isotope geoscience: A proposal. Chemical Geology 66:163–177.

Russell S. S., Krot A. N., MacPherson G. J., Huss G. R., Itoh S.,Yurimoto H., and Keil K. 2005. Genetic relationship betweenrefractory inclusions and chondrules. In Chondrites and theprotoplanetary disk, edited by Krot A. N., Scott E. R. D., andReipurth B. San Francisco: Astronomical Society of the Pacific.pp. 317–353.

Russell S. S., Hartmann L., Cuzzi J., Krot A. N., Gounelle M., andWeidenschilling S. 2006. Time scales of the solar protoplanetarydisk. In Meteorites and the early solar system II, edited byLauretta D. S. and McSween H. Y., Jr. Tucson, Arizona: TheUniversity of Arizona Press. pp. 233–253.

Scott E. R. D. and Krot A. N. 2005. Chondritic meteorites and thehigh-temperature nebular origins of their components. InChondrites and the protoplanetary disk, edited by Krot A. N.,Scott E. R. D., and Reipurth B. San Francisco: AstronomicalSociety of the Pacific. pp. 15–54.

Shimoda G., Nakamura N., Kimura M., Nohda S., and Yamomoto K.2005. Evidence from the Rb-Sr system for 4.4 Ga alteration ofchondrules in the Allende (CV3) parent body. Meteoritics &Planetary Science 40:1059–1072.

Shu F. H., Shang S., Gounelle M., Glassgold A. E., and Lee T. 2001.The origin of chondrules and refractory inclusions in chondriticmeteorites. The Astrophysical Journal 548:1029–1050.


Sokol A. K. and Bischoff. A. 2006. Simultaneous accretion ofdifferentiated of metamorphosed asteroidal clasts and chondrules(abstract). Meteoritics & Planetary Science 41:A164.

Stirling C. H., Halliday A. N., Potter E.-K., Andersen M. B., andZanda B. 2006. A low initial abundance of 247Cm in the earlysolar system and implications for r-process nucleosynthesis.Earth and Planetary Science Letters 251:386–397.

Tatsumoto M., Unruh D. M., and Desborough G. A. 1976. U-Th-Pb and Rb-Sr systematics of Allende and U-Th-Pbsystematics of Orgueil. Geochimica et Cosmochimica Acta 40:617–634.

Thirlwall M. F. 2002. Multicollector ICP-MS analysis of Pb isotopesusing a 207Pb-204Pb double spike demonstrates up to 400 ppm/amu systematic errors in Tl-normalization. Chemical Geology184:255–279.

Thrane K., Bizzarro M., and Baker J. A. 2006. Extremely briefformation interval for refractory inclusions and uniformdistribution of 26Al in the early solar system. The AstrophysicalJournal 646:L159–L162.

Titterington D. M. and Halliday A. N. 1979. On the fitting of parallelisochrons and the method of maximum likelihood. ChemicalGeology 26:183–195.

Wadhwa M., Amelin Y., Bogdanovski O., Shukolyukov A., LugmairG. W., and Janney P. 2005. High-precision relative and absoluteages for Asuka-881394, a unique and ancient basalt (abstract

#2121). 36th Lunar and Planetary Science Conference.CD-ROM.

Weinbruch S., Armstrong J., and Palme H. 1994. Constraints on thethermal history of the Allende parent body as derived fromolivine-spinel thermometry and Fe/Mg interdiffusion in olivine.Geochimica et Cosmochimica Acta 58:1019–1030.

Weisberg M. K. and Prinz M. 1998. Fayalitic olivine in CV3chondrite matrix and dark inclusions: A nebular origin.Meteoritics & Planetary Science 33:1087–1111.

Yang J., Goldstein J. I., and Scott E. R. D. 2006. Evolution ofdifferentiated asteroids as inferred from cooling rates ofmagmatic iron meteorites (abstract #5146). Meteoritics &Planetary Science 41:A193.

York D. 1969. Least-squares fitting of a straight line with correlatederrors. Earth and Planetary Science Letters 5:320–324.

Zanda B. 2004. Chondrules. Earth and Planetary Science Letters224:1–17.

Zinner E. and Göpel C. 2002. Aluminum-26 in H4 chondrites:Implications for its production and its usefulness as a fine-scalechronometer for early solar system events. Meteoritics &Planetary Science 37:1001–1013.

Zolotov M. Y., Mironenko M. V., and Shock E. L. 2006.Thermodynamic constraints on fayalite formation on parentbodies of chondrites. Meteoritics & Planetary Science 41:1775–1796.

Documents

Pb isotopic age of the Allende chondrules