author's personal copyharrison/harrison... · author's personal copy thermal constraints...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Thermal constraints on the early history of the H-chondriteparent body reconsidered

Keith P. Harrison *, Robert E. Grimm

Southwest Research Institute, 1050 Walnut St., Ste 300, Boulder, CO 80302, USA

Received 16 March 2010; accepted in revised form 26 May 2010; available online 25 June 2010

Abstract

Reconstructions of the early thermal history of the H-chondrite parent body have focused on two competing hypotheses.The first posits an undisturbed thermal evolution in which the degree of metamorphism increases with depth, yielding an“onion-shell” structure. The second posits an early fragmentation–reassembly event that interrupted this orderly cooling pro-cess. Here, we test these hypotheses by collecting a large number of previously published closure age and cooling rate data andcomparing them to a suite of numerical models of thermal evolution in an idealized parent body. We find that the onion-shellhypothesis, when applied to a parent body of radius 75–130 km with a thermally insulating regolith, is able to explain 20 ofthe 21 closure age data and 62 of the 71 cooling rates. Furthermore, six of the eight meteorites for which multiple data (atdifferent temperatures) are available, can be accounted for by onion-shell thermal histories. We therefore conclude that nocatastrophic disruption of the H-chondrite parent body occurred during its early thermal history. The relatively small numberof data not explained by the onion-shell hypothesis may indicate the formation of impact craters on the parent body which,while large enough to excavate all petrologic types, were small enough to leave the parent body largely intact. Impact eventsfulfilling these requirements would likely have produced transient crater diameters at least 30% of the parent body diameter.� 2010 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

The H-chondrite meteorites are thought to have origi-nated in a single, undifferentiated parent body (e.g.,Wasson, 1972). The parent body underwent varying degreesof metamorphism as a result of heat released internally,probably by the radioactive decay of 26Al (Minster andAllegre, 1979). The degree of metamorphism is inferredfrom petrologic type, which ranges from type 3 (least meta-morphosed) to 6 (most metamorphosed; Van Schmus andWood, 1967). Petrologic type has thus been used as a proxyfor peak metamorphic temperatures (Dodd, 1969, 1981).The peak temperatures of Dodd (1981) were derived fortypes 3 and 6 only (the respective ranges are 400–600 and750–950 �C), with peak temperatures for types 4 and 5 cal-culated by interpolation. Newer thermometric techniqueshave yielded temperature ranges of 865–926 �C for type 6

meteorites (Slater-Reynolds and McSween, 2005), 675–750 �C for the lower bound on peak temperatures for types4–6 (Wlotzka, 2005; Kessel et al., 2007), and temperaturesanywhere from 260 to 600 �C for the different subclassesof type 3 (Huss et al., 2006, and references therein).

The relationship between peak temperature and petro-logic type has allowed broad constraints to be placed onthe early thermal history of the H-chondrite parent body.The most straightforward approach, arising from a simplethermal model of internal heating in a sphere, is the“onion-shell” model. Peak temperatures decrease monoton-ically away from the center of the body, producing layers ofprogressively lower petrologic type (Wood, 1967; Minsterand Allegre, 1979; Pellas and Storzer, 1981). The lowerthe peak temperature, the shorter the cooling time, and arange of methods (described in further detail below) havebeen employed in recent years to infer such times (Trieloffet al., 2003; Amelin et al., 2005; Bouvier et al., 2007).

Additional constraints are available in the form of cool-ing rates: samples that originated near the center of the

0016-7037/$ - see front matter � 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2010.05.034

* Corresponding author. Tel.: +1 720 240 0112.E-mail address: [email protected] (K.P. Harrison).

www.elsevier.com/locate/gca

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 74 (2010) 5410–5423


parent body, where temperature gradients were low, likelycooled slowly, while samples from shallow depths, wheretemperature gradients were steeper, likely cooled rapidly.The onion-shell model might thus be confirmed if petro-logic type (i.e., peak temperature) were observed to corre-late inversely with cooling rate.

Some of the first cooling rate measurements made on H-chondrites (Pellas and Storzer, 1981; Lipschutz et al., 1989)seemed to confirm the onion-shell hypothesis, while othersdid not (Scott and Rajan, 1981). The analysis of a largenumber of samples by Taylor et al. (1987) again appearedto contradict the onion-shell model, leading these workersto invoke the hypothesis, first suggested by Grimm(1985), that the early hot parent body was shattered by alarge impact, followed by the haphazard reassembly offragments of various temperatures. Subsequent cooling ofsuch a body would be expected to record no correlation be-tween petrologic type and cooling rate.

There are, however, problems with some of the coolingrate data collected by Scott and Rajan (1981) and Tayloret al. (1987). The highest H-chondrite cooling rates re-corded by these workers are for metals in the fine-grainedmatrices of regolith breccias. These matrices appear to haveformed by the comminution of larger clasts as a result ofshallow regolith development on the surface of the parentbody (Bischoff et al., 1983). If this process occurred duringthe early metamorphic heating of the parent body, the com-minution of hot clasts into finer grains would have acceler-ated cooling rates in the material, explaining the very widerange of values up to a few thousand �C/Ma (Scott andRajan, 1981; Taylor et al., 1987; Williams et al., 2000). Alter-natively, these materials may have been reheated followingearly metamorphism (e.g., Kessel et al., 2007). These cool-ing rates may therefore have little relevance to global-scaleparent body thermal history. While this does not necessarilyput to rest the fragmentation–reassembly hypothesis, thereare additional concerns about the Taylor et al. (1987) datathat raise significant doubts, and these are addressed later.

Some cooling rate data published after those of Tayloret al. (1987) appear to support an onion-shell model onceagain (Gopel et al., 1994; Trieloff et al., 2003), leaving a some-what conflicting picture of the early H-chondrite parent bodythermal history. We attempt to address this problem here bybringing together a large sample of the available data andcomparing them to numerical models of onion-shell thermaldevelopment in an idealized parent body.

2. METHODOLOGY

Our methodology, which is similar to that of Bennettand McSween (1996), consists of the following steps: (1)Initialize a numerical model of parent body thermal historywith thermal parameter values drawn from a plausiblerange (Section 2.2). (2) Run the model iteratively, each timeadjusting the heat source (via the initial 26Al/27Al ratio) un-til the peak temperature attained at the center of the body isapproximately 1000 �C (Section 2.2). (3) Rerun the modelwith different parent body radii until an optimal fit is foundto closure time and cooling rate data from the literature(Section 2.1).

We proceed first with a detailed description of the liter-ature data used, followed by a full characterization of ourthermal model.

2.1. Data analysis

Our study integrates two types of data from the literaturethat are often reported separately. The first is the closuretime corresponding to a particular radiometric dating tech-nique, namely the time elapsed before a meteorite passesthrough a given temperature during metamorphic coolingof the parent body. The second type of data gives the rateat which the sample is thought to have cooled through a par-ticular closure temperature. Critical to the comparison be-tween these data and numerical results is the choice ofpeak temperatures used to distinguish each petrologic type:data from the literature are also used to motivate this choice.We discuss our various data sources below.

2.1.1. Closure times

Closure time is defined as the difference between cal-cium–aluminum-rich inclusion (CAI) age and the closureage of the sample under consideration. We use a recentlycalculated CAI age of 4568.5 ± 0.5 Ma (Bouvier et al.,2007) for all data except those of Kleine et al. (2008), whoseclosure times use the statistically indistinguishable CAI ageof 4568.3 ± 0.7 Ma (Lugmair and Shukolyukov, 1998).

(1) 40Ar–39Ar closure times: 40Ar produced by the radio-genic decay of K is retained in oligoclase feldsparbelow about 280 �C (Turner et al., 1978). Theamount of 40Ar and, therefore, the time taken forthe sample to reach the closure temperature, is mea-sured relative to artificially produced 39Ar. Trieloffet al. (2003) performed Ar–Ar measurements on sev-eral samples (Table 1), and we include their results inour analysis. We have followed the recommendationof Trieloff et al. (2003) and subtracted 30 Ma fromtheir closure times to account for recent recalibra-tions of the Ar–Ar age scale (Renne, 2000; Begemannet al., 2001; Trieloff et al., 2001).

(2) Pb–Pb closure times: Production of 207Pb by thedecay of 235U early in the solar system allows agemeasurements of very old samples to be madethrough measurements of the lead ratio 207Pb/206Pb.The reliability of this method for a particular samplecan be gauged by comparing its results against thoseof the 238U–206Pb system. Pb–Pb closure times, whichhave a closure temperature in the range 430–530 �C,were used by Gopel et al. (1994) to date phosphatesin ordinary chondrites. Their results are widely cited(e.g., Ganguly and Tirone, 2001; Trieloff et al., 2003;Bouvier et al., 2007), and we include them in ouranalysis. We also include related Pb–Pb measure-ments, made by Bouvier et al. (2007) and Amelinet al. (2005), that have closure temperatures of 680–880 �C. (A 480 �C Pb–Pb closure age of 77 ± 16 Mafor the Estacado meteorite is also available; Blinovaet al., 2007. We acquired knowledge of this datumtoo late for it to be fully integrated into our analysis,

History of the H-chondrite parent body 5411


Table 1Closure time and cooling rate data.

Meteorite Type Closure timea (Ma) Temperature (�C) Method Reference

Closure times

Forest Vale H4 7.6 ± 0.7 480 ± 50 Pb–Pb Gopel et al. (1994)16.5 ± 8.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Ste. Marguerite H4 5.8 ± 1.1 480 ± 50 Pb–Pb Gopel et al. (1994)6.8 ± 1.7 730 ± 50 Pb–Pb Bouvier et al. (2007)6.5 ± 16.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

ALH 84069 H5 5.9 ± 1.4 825 ± 75 Hf–W Kleine et al. (2008)

Allegan H5 18.3 ± 0.8 480 ± 50 Pb–Pb Gopel et al. (1994)27.5 ± 11.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Nadiabondi H5 12.9 ± 3.9 480 ± 50 Pb–Pb Gopel et al. (1994)9.6 ± 2.8 730 ± 50 Pb–Pb Bouvier et al. (2007)33.5 ± 10.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Richardton H5 5.6 ± 1.1 825 ± 75 Hf–W Kleine et al. (2008)17.8 ± 3.1 480 ± 50 Pb–Pb Amelin et al. (2005)43.5 ± 11.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Estacado H6 10.0 ± 1.7 825 ± 75 Hf–W Kleine et al. (2008)103.5 ± 5.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Guarena H6 64.1 ± 0.5 480 ± 50 Pb–Pb Gopel et al. (1994)84.5 ± 6.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Kernouve H6 9.4 ± 1.1 825 ± 75 Hf–W Kleine et al. (2008)46.0 ± 2.0 480 ± 50 Pb–Pb Gopel et al. (1994)69.5 ± 6.5 280 ± 20 Ar–Ar Trieloff et al. (2003)

Meteorite Type Cooling rate Temperatureb (�C) Method Reference

Cooling rates

Dhajala H3 50c 500 Met.d Taylor et al. (1987)

Tieschitz H3 2c 500 Met. Willis and Goldstein (1981)Conquista H4 25 500 Met. Taylor et al. (1987)

Kesen H4 20 500 Met. Taylor et al. (1987)

Sete Lagoas H4 15 500 Met. Williams et al. (2000)

Ste. Marguerite H4 84 ± 77 200 244Pu Trieloff et al. (2003)

Wellman H4 8 500 Met. Willis and Goldstein (1981)Allegan H5 3.25 ± 0.55 200 244Pu Trieloff et al. (2003)

15 500 Met. Taylor et al. (1987)

Ehole H5 4 500 Met. Willis and Goldstein (1981)

Fayetteville H5 20 to >1000 (3)e 500 Met. Williams et al. (2000)

Ipiranga H5 5–20 (3) 500 Met. Williams et al. (2000)

Malotas H5 8 500 Met. Willis and Goldstein (1981)

Nuevo Mercurio H5 15 500 Met. Taylor et al. (1987)

Nulles H5 25–140 (11) 500 Met. Williams et al. (2000)

Richardton H5 2.95 ± 0.45 200 244Pu Trieloff et al. (2003)20 500 Met. Taylor et al. (1987)

Sena H5 2.45 ± 0.35 200 244Pu Trieloff et al. (2003)

Sete Lagoas H5 5–25 (2) 500 Met. Williams et al. (2000)

Sutton H5 4 500 Met. Willis and Goldstein (1981)

Cangas de Onis H6 5–40 (12) 500 Met. Williams et al. (2000)

Estacado H6 2.55 ± 0.35 200 244Pu Trieloff et al. (2003)10 500 Met. Taylor et al. (1987)

5412 K.P. Harrison, R.E. Grimm / Geochimica et Cosmochimica Acta 74 (2010) 5410–5423


but we have determined that it, and an associatedcooling rate of about 6 �C/Ma from the sameauthors, are consistent with our conclusions).

(3) Hf–W closure times: 182W produced by the decay of182Hf early in solar system history can, when com-pared to other W isotopes, be used to determine agesfor the relatively high closure temperature range of750–900 �C (e.g., Harper and Jacobsen, 1996). Kleineet al. (2008) applied this system to ordinary chon-drites, and we include their results in our analysis.

2.1.2. Cooling rates

(1) Metallographic cooling rates: As Fe–Ni alloys in theparent body began to cool from high temperatures(>700 �C), the phase diagram of the binary Fe–Ni systemindicates that they consisted of pure taenite (Wood, 1967).As cooler temperatures were reached, however, precipita-tion of kamacite began, and the Ni content of the two phaseschanged via diffusion from taenite to kamacite. Because therate of diffusion in both phases decreases with cooling, thesystem must eventually have moved out of equilibrium. Evi-dence for this is found in the heterogeneous distribution ofNi in the grains which, together with idealized models ofNi diffusion, can be used to infer the cooling rate experi-enced by the grains as they passed through about 500 �C.

Most of the cooling rate data available in the literatureare derived with the metallographic method. We use therates listed in Taylor et al. (1987), which include data ac-quired by previous workers (e.g., Wood, 1967; Scott andRajan, 1981) and later revised by Willis and Goldstein(1981). Interestingly, most of the measurements taken di-rectly by Taylor et al. (1987) appear to be consistent withthe onion-shell model, as pointed out by Gopel et al.(1994). Of the remaining data, we omit the following:

(a) Bath meteorite: The cooling rate of 80 �C/Ma may beproblematic due to scatter in the data, as discussed byWillis and Goldstein (1981).

(b) A cooling rate of approximately 140 �C/Ma isreported for two meteorites (types 4 and 5) in the

key figure (Fig. 5) of Taylor et al. (1987), but thesource of these data is not clear. Candidates arethe Weston, Fayetteville, and Leighton meteorites.Weston and Fayetteville are regolith breccias, andbecause the high cooling rates initially reported byScott and Rajan (1981) were measured in the matri-ces of these breccias, we regard them as irrelevantto global-scale thermal history, as discussed in Sec-tion 1. The Leighton cooling rate of 200 �C/Ma, orig-inally measured by Wood (1967) and modifiedupwards by Willis and Goldstein (1981), was derivedfrom a relatively incoherent data plot. Wood (1967)suggests that “nothing can be said about [its] coolingrate”. We therefore discard this datum also.

Finally, we include a large set of metallographic coolingrates estimated for equilibrated clasts in four H-chondriteregolith breccias (Williams et al., 2000). Once again, we omitwidely variable cooling rates for matrix material also esti-mated by these authors. Williams et al. (2000) used the samemethodology as Taylor et al. (1987) and we therefore assignto all metallographic cooling rates the uncertainty of a factorof 2.5 suggested by the earlier authors. There may also be asystematic error associated with the metallographic method:Taylor et al. (1987) suggest that the method may producecooling rates up to a factor of 3 too small. A more recent revi-sion to the method indicates that cooling rates may be up toan order of magnitude too small, but this analysis was ap-plied to iron meteorites and mesosiderites only (Hopfe andGoldstein, 2001). Although these suggestions are worth not-ing, the lack of reliable data constraining systematic coolingrate errors for ordinary chondrites compels us to assume neg-ligible values for the time being.

(2) 244Pu cooling rates: Fission tracks produced by 244Pufission will fail to anneal below a certain mineral-dependenttemperature, thereby remaining in the rock indefinitely.Fission track density can thus be related to the time atwhich this temperature was reached (Pellas and Storzer,1981; Pellas et al., 1997). Trieloff et al. (2003) analyzed fis-sion tracks in two minerals (orthopyroxene and merrillite)

Table 1 (continued)

Meteorite Type Cooling rate Temperatureb (�C) Method Reference

Guarena H6 75 ± 25 775 Ol–Spf Kessel et al. (2007)2.95 ± 0.45 200 244Pu Trieloff et al. (2003)

Ipiranga H6 5–25 (11) 500 Met. Williams et al. (2000)

Kernouve H6 2.65 ± 0.35 200 244Pu Trieloff et al. (2003)10 500 Met. Taylor et al. (1987)

Sete Lagoas H6 5–75 (7) 500 Met. Williams et al. (2000)

a Closure times are relative to CAI formation. A CAI age of 4568.5 ± 0.5 Ma (Bouvier et al., 2007) is used to calculate closure times fromages when necessary.

b Uncertainties for temperatures associated with cooling rates are not generally available; we use ±25 �C.c Cooling rates from Taylor et al. (1987) and Willis and Goldstein (1981) are, at the former authors’ suggestion, assumed to have an

uncertainty of a factor of 2.5.d Met. = Metallographic cooling rate method.e Numbers in parentheses indicate the number of clasts for which independent cooling rate measurements were made by Williams et al.

(2000). For brevity, we provide the only the range of cooling rates obtained, and refer the reader to Williams et al. (2000) for details.f Ol–Sp = Olivine–spinel cooling rate method.



in ordinary chondrites, yielding two closure times andtherefore a cooling interval between the two closure temper-atures (280 and 120 �C, respectively). We use this interval toinfer an approximate cooling rate at the intermediate tem-perature (200 �C). To compute the uncertainty in thesemeasurements, we use individual and systematic uncertain-ties in the cooling interval, and an uncertainty in closuretemperature (Trieloff et al., 2003).

(3) Olivine–spinel cooling rate: Fe–Mg exchange betweenolivine and spinel can be analyzed in a fashion similar to themetallographic method in order to obtain cooling rates attemperatures of about 700–850 �C (Kessel et al., 2007).Kessel et al. (2007) provide a single measurement only,which we include in our analysis.

2.1.3. Peak temperature

A crucial part of the modeling effort is the delineation ofpredicted results according to petrologic type. This exerciserelies on the assumption introduced earlier that peak tem-perature and petrologic type are correlated. Although thisassumption is still thought valid, recent work suggests thatit ought to be quantified somewhat differently. Early workby Dodd (1981) yielded peak temperature estimates of400–600 �C for type 3 and 750–950 �C for type 6. Types 4and 5 were assigned intermediate peak temperaturesthrough interpolation. Further work on type 3 meteoriteshas revealed a wide range of peak temperatures from 260to 600 �C (Huss et al., 2006, and references therein). Con-cerning type 6 meteorites, Kleine et al. (2008) refer to morerecent work using two-pyroxene thermometry, which indi-cates peak temperatures of 865–926 �C (Slater-Reynoldsand McSween, 2005), although a higher limit of �1000 �Cis possible (this is where melting in the FeNi–FeS system be-

gins). Kleine et al. (2008) also note that olivine–spinel ther-mometry indicates a 675–750 �C lower bound on peaktemperatures for types 4–6 (Wlotzka, 2005; Kessel et al.,2007), which dovetails reasonably well with the 600 �Cupper limit for type 3. Thus, while types 3 and 6 appearto be easily distinguishable by peak temperature, types 4and 5 do not. In our thermal modeling work we thereforeconsider types 4 and 5 to be part of one group with peaktemperatures ranging from 675 to 865 �C. Subtypes withintype 3s appear to be quite well correlated with peak temper-ature (e.g., Wlotzka, 2005), but there is little point in delin-eating these subtypes in thermal models since there are sofew type 3 age and cooling rate data to compare againstthe predicted regions. Accordingly then, we delineate modelparameter spaces into three regions corresponding to alltype 3s (peak temperatures <675 �C), types 4 and 5 together(675–865 �C), and type 6s (865–1000 �C).

2.2. Numerical model

As stated above, the purpose of the current work is tocompare closure time and cooling rate data with predictionsmade from onion-shell numerical models. The numericalmodel allows us to estimate which parts of the tempera-ture-vs.-time and cooling rate-vs.-temperature parameterspaces are expected to be occupied by meteorites of differentpetrologic types. It also allows us to determine if multipledata from a single meteorite are consistent with the coolinghistory of a single parcel of material in the parent body.

Our thermal model, implemented with the COMSOL3.5a finite-element code (www.comsol.com) is a 1D,spherically symmetric abstraction of a parent body withuniform internal heating due to 26Al decay, and a radiative

Table 2Parameter values for our optimal model with parent body radius 100 km. Optimal results can be obtained for other radii by varying porosityor thermal diffusivity as indicated in Fig. 1.

Thermal and geometric properties

Thermal diffusivitya A + B/TA 1.56 � 10�7 m2 s�1

B 8.88 � 10�5 m2 K s�1

Thermal conductivityb (A + B/T)q cP(T)(1 � 1.13/0.333)

Interior Megaregolithc Regolithc

Bulk density (q, kg m�3) 3250 2500 1500Porosity (/, %) 6 25 50Initial temperature 170 KParent body radius 100 km

Internal heating

Initial 26Al/27Al 7.45 � 10�6 (2.2 Ma after CAIs)Mass fraction 27Al 0.011726Al decay constant 9.63 � 10�7 year�1

Radiative boundary condition

Semi-major axis 3 AUEmissivity 0.8Albedo 0.05

a After Yomogida and Matsui (1983).b After Akridge et al. (1998). Specific heat in this expression is modeled after Yomogida and Matsui (1983).c Megaregolith and regolith are both 410 m thick. This value was obtained through the process of optimizing the model.



boundary (Table 2). We use the finite-element methodinstead of simpler finite-difference or analytical solutionsbecause we were originally working toward a model of 3Dfragmentation and reassembly (Grimm et al., 2005). Ourmodel assumes that parent body accretion occurred on atime scale short enough to be thermally insignificant.Certainly, if accretion was relatively sedate (on the order of1 Ma), thermal evolution during this time would be important(Ghosh et al., 2003), however recent transient-overdensitymodels suggest that accretion may have occurred much morerapidly (within several orbits, Johansen et al., 2007).

We ran two types of model: homogeneous and insulated.The homogeneous model has spatially uniform thermalproperties, while the insulated model has three distinct re-gions: interior, megaregolith, and regolith, modeled afterAkridge et al. (1998).

In all models, thermal diffusivity is temperature depen-dent. We use j(T) = A + B/T, where A and B are constantsand T is temperature. This relationship, first put forward byYomogida and Matsui (1983), is commonly used in othermodels of ordinary chondrite parent body thermal history(e.g., Bennett and McSween, 1996; Akridge et al., 1998).We ran models with three different sets of values for A

and B, all from Yomogida and Matsui (1983) (Table 2).The first two are those considered by Bennett and McSween(1996) to be appropriate for uncompacted (A = 1.56 �10�7, B = 8.88 � 10�5) and compacted (A = 3.66 � 10�7,B = 2.91 � 10�4) parent bodies, respectively. The third(A = 4.47 � 10�7, B = 1.32 � 10�4) is that used in the insu-lated model of Akridge et al. (1998).

Thermal conductivity is given by k(T) = j(T)q cP(T),where q is bulk density and cP is specific heat. The modelof Bennett and McSween (1996) uses a fixed thermalconductivity, implying that specific heat must vary withtemperature so as to compensate for the temperature-dependence of the diffusivity. We found that this specificheat variation is in good agreement with the temperature-dependence suggested by Yomogida and Matsui (1983),which we use in our models.

For our insulated model, we add a porosity term to theconductivity, in keeping with Akridge et al. (1998):

kðT Þ ¼ jðT ÞqcP ðT Þð1� 1:13/0:333Þ ð1Þ

This approach models the lower bulk thermal conductivi-ties associated with higher porosities in the megaregolithand regolith. We considered megaregolith and regoliththicknesses in the same range (0.5–3 km) as Akridge et al.(1998). Thicknesses were varied within this range so as toimprove the fit between model and data.

We ran models with the above variety of thermal param-eters in combination with six alternative parent body radii:55, 70, 85, 100, 125, and 150 km, which encompass theapproximate range of values suggested by other authors(�75–100 km; Grimm et al., 2005, and references therein).

3. RESULTS AND DISCUSSION

We find a range of models that produce the same opti-mal fit to the literature data. These are insulated models

with a unique combination of parent body size and thermalconductivity. Changes to parent body size require propor-tionate changes to conductivity, which can be madethrough modifications to either the thermal diffusivity orthe porosity as indicated in Fig. 1. All models in the optimalsuite accrete (instantaneously) at 2.2 Ma after CAI forma-tion, corresponding to an initial 26Al/27Al of 7.45 � 10�6.We discuss in detail the 100 km radius parent body(Table 2), since this is the smallest body with a thermal dif-fusivity computed from parameters drawn from the chosenparameter space. However, we note that identical resultscan be obtained in even smaller bodies by increasing theporosity in Eq. (1) such that thermal conductivity decreasesby the same factor as the parent body volume withoutchanging the diffusivity. Such porosity increases are notunreasonable for bodies down to about 75 km radius (atthis point, the interior porosity reaches that of the overlyingmegaregolith). We compare graphically the optimal modelresults with the collected literature data in Figs. 2 and 3.(In Fig. 3, the lateral spread of the data is made for clarity,and does not represent real temperature variations.)

Our measure of model success has two components.First, we determine the number of data from our literaturesurvey that fall within the region predicted (by our model)for their petrologic type. Second, we determine if there existmodeled thermal histories that can explain multiple data-points available for a single meteorite. We begin with adiscussion of the first, broader measure of model success,looking in turn at closure times, and then cooling rates.

Fig. 1. Parameter space yielding optimal model results. The dashedline indicates the factor by which the interior porosity of thenominal model (0.06) must be multiplied in order to obtain optimalresults for a different parent body radius. For radii below about75 km, the interior porosity reaches that of the overlying megar-egolith, and higher values are therefore unlikely. For radii above120 km, porosity reaches zero and can no longer be used tooptimize the model. Optimal results can alternatively be obtainedby multiplying nominal thermal diffusivity parameters A and B bythe factor indicated by the solid line. For parent body radii between100 and 130 km, this adjustment yields diffusivities within the rangeconsidered in most studies.



3.1. Closure times

For our optimal model suite, only 1 of the 21 data-points lies outside the region predicted for its petrologictype (Fig. 2). This point (closure time = 5.8 Ma; tempera-ture = 480 �C) is from the type 4 Ste. Marguerite meteorite(Gopel et al., 1994), and is statistically identical to closuretimes estimated for this meteorite for two other tempera-tures (280 and 730 �C, Trieloff et al., 2003; Bouvier et al.,2007). Since the meteorite cannot have cooled through mul-tiple temperatures at the same time, one or more of thesedata must be erroneous. The uncertainties in the two lattertemperatures are such that a coherent cooling history maybe derived for Ste. Marguerite if the Gopel et al. (1994) da-tum is ignored (see Section 3.3).

3.2. Cooling rates

Cooling rate fits are shown in Fig. 3. Most of the coolingrates in our dataset are metallographic, and therefore haverelatively large uncertainties derived from the cooling ratecurves used and from the scatter of the data plotted overthese curves (Wood, 1967; Taylor et al., 1987; Williamset al., 2000). Thus few points can, with statistical confi-dence, be said to lie outside the appropriate predicted re-gion. Of the 33 type 4 and 5 meteorites and the 36 type 6meteorites, only 7 and 2 suffer this fate, respectively. Onlytwo type 3 cooling rates are available, and both fall withintheir predicted region. In order to further investigate theinfluence of the assumed uncertainty in metallographiccooling rates, we re-evaluated the fit for a range of uncer-tainty factors from 1 (no uncertainty) to 5, which includesour nominal value of 2.5 (Taylor et al., 1987). We find thatthe proportion of data fitting model predictions remains rel-atively high (>70%) for uncertainty factors as low as 1.7(Fig. 4). Even the assumption of no uncertainty yields afit greater than 50%. We thus retain our nominal uncer-tainty factor of 2.5. As described above, there may be sys-tematic errors in the metallographic method thatartificially reduce cooling rates by about a factor of 3(Taylor et al., 1987) or possibly even 10 (Hopfe and

Goldstein, 2001). We find that no uniform upward revisionof the metallographic cooling rates in our study producesany improvement to their overall fit with the thermalmodel. It may be that adjustments to the model couldpreserve the fit if a systematic increase in cooling rateswas found to be necessary, but until such revisions arebetter constrained, we omit them from our analysis.

Including non-metallographic cooling rates, then, wefind that 87% of all the cooling rate data collected (62 of71) conform to the predictions of our optimal onion-shellmodel suite. The nine offending data do not originate fromone particular study, they are not confined to a single mete-orite, nor are they exclusively metallographic (Fig. 3).

3.3. Thermal constraints on individual meteorites

Our second method of assessing the thermal model is toconsider the meteorites in our study for which multiple data(at different closure temperatures) are available. We wish todetermine if there exist model thermal histories that can ex-plain all measurements within a single meteorite. To begin,we consider closure time and cooling rate data indepen-dently. Starting with closure times, we seek the set of com-puted thermal histories that pass within the error bars ofthe closure time measurements for a given meteorite. Thereare eight meteorites for which multiple closure time data areavailable (Allegan, Estacado, Forest Vale, Guarena,Kernouve, Nadiabondi, Richardton, and Ste. Marguerite).There exist thermal histories that successfully connect theclosure age data for all of these meteorites except Ste. Mar-guerite (H4; Fig. 4 and Table 3). As discussed above, one ofthe Ste. Marguerite data does not lie within the field pre-dicted for type 4/5 meteorites (Fig. 2). If we eliminate thisdatum, thermal histories with peak temperatures of 704–837 �C are observed to connect the remaining two points.

Next, we consider meteorites with multiple coolingrate data (Allegan, Estacado, Guarena, Kernouve, andRichardton). Cooling rate histories can be found (Fig. 6)that connect the available for all of these meteorites,although we emphasize that only two data are availablein each case.

Fig. 2. (a) Closure times from the literature plotted against regions (shaded, with solid borders) for each petrologic type predicted by ouroptimal onion-shell numerical model. Numbered circles depict petrologic type. The five coolest types 4 and 5 data all correspond to atemperature of 280 �C, and have been spread vertically for clarity.



Plausible thermal histories connecting either closure timesor cooling rates for a given meteorite are presented quantita-tively in Table 3. The histories are denoted by their range ofpeak temperature and their corresponding radial positions inthe parent body. In all cases except one, plausible thermalhistories overlap with the range corresponding to the petro-logic type of the meteorite. The exception is Guarena (H6):the range of thermal histories derived from its two coolingrates have peak temperatures between 762 and 793 �C, belowthe minimum peak temperature chosen for type 6s (865 �C).

Having found thermal histories for closure time andcooling rate data independently, we now seek thermal histo-

ries that satisfy both sets of data. In principle, the two setsof histories for each meteorite should overlap. This is whatwe observe for four of the six relevant meteorites (Allegan,Estacado, Kernouve, and Richardton). Guarena and Ste.Marguerite are the two exceptions. As noted already, theGuarena cooling rate histories do not fall in the type 6range, making it impossible for these histories to coincidewith the very narrow type 6 range derived from closure timedata. For Ste. Marguerite, the lack of plausible historiesconnecting its closure time data makes moot the inclusionof the cooling rate datum. However if, once again, we omitthe Ste. Marguerite closure time that does not fall within

Fig. 3. Cooling rates at 200 �C for (a) types 4 and 5 and (b) type 6; at 500 �C for (c) type 3, (d) types 4 and 5, and (e) type 6; at 780 �C for (f)type 6 meteorites. The abscissa of each panel represents the single closure temperature indicated, with the data spread laterally for clarity. Asin Fig. 1, the solid shaded area in each panel denotes the region predicted by our optimal thermal model for the petrologic type(s) indicated.Note the varying ordinate scales.



the type 4/5 region, we find thermal histories that connectthe remaining two closure times and the cooling rate. Thesehistories are the same as those derived from the two closuretimes alone: 704–837 �C, equivalent to normalized radialpositions of 0.952–0.975.

3.4. Additional constraints

We briefly consider other available constraints on earlyH-chondrite parent body thermal evolution. First is thetime of parent body accretion as measured from CAI for-

mation. This has been estimated in other modeling effortsto be approximately 2 Ma (Miyamoto et al., 1981;Grimm, 1985; Grimm and McSween, 1993), and has beenconstrained by the Pb–Pb chronology of Gopel et al.(1994) to be 3.0 ± 2.6 Ma. Our value of 2.2 Ma agreeswell with these constraints. Second is the radius of theparent body, which has been estimated to be between75 and 100 km (Grimm et al., 2005 and references there-in). As shown in Fig. 1, reasonable adjustments to poros-ity or thermal diffusivity allow our optimal results to berepeated for an even wider range of parent body radii(75–130 km).

Next, we consider the formation time interval, defined asthe time taken for the center of the body to cool throughthe Rb–Sr blocking temperature of about 130 �C.Miyamoto et al. (1981) used a formation time interval of100 Ma, although more recent work by Bennett andMcSween (1996) argues for a smaller value of about60 Ma. Our optimal models do not conform well to thisconstraint, producing formation time intervals on the orderof 400 Ma (although the most shallow type 6 material takesonly 70 Ma). This is because some closure time age dataused in our analysis are in direct conflict with a shortformation time interval. Specifically, Ar–Ar closure timedata for type 6 meteorites (Trieloff et al., 2003) indicate thatsome samples were still at temperatures more than twice theRb–Sr blocking temperature at times as late as 100 Ma(Table 1 and Fig. 2). Further analysis of both the Rb–Srresults used to determine formation time interval, and theAr–Ar ages of Trieloff et al. (2003), must be carried outin order to resolve this problem, and is beyond the scopeof the current work.

Fig. 4. The proportion of metallographic cooling rates that fit withmodel predictions, as a function of the uncertainty factor assignedto these rates. The dot marks the nominal factor of 2.5 used in ouranalysis.

Table 3Plausible thermal histories for those meteorites with multiple data available (see Figs. 4 and 5).

Name Type Range of peaktemperatures (�C)

Normalized radialdistancea

Thermal histories derived from closure time data

Forest Vale H4 553–707 0.975–0.991Ste. Marguerite H4 No plausible historiesAllegan H5 831–909 0.929–0.953Nadiabondi H5 773–886 0.938–0.965Richardton H5 804–923 0.922–0.959Estacado H6 984–990 0.848–0.862Guarena H6 991–993 0.836–0.842Kernouve H6 971–985 0.861–0.885

Thermal histories derived from cooling rate data

Allegan H5 717–998 0.804–0.973Richardton H5 762–988 0.853–0.966Estacado H6 826–1000 0–0.954Guarena H6 762–793 0.961–0.966Kernouve H6 815–1000 0–0.957

Thermal histories derived from data of both types

Ste. Marguerite H4 No plausible historiesAllegan H5 831–909 0.929–0.953Richardton H5 804–923 0.922–0.959Estacado H6 984–990 0.848–0.862Guarena H6 No plausible historiesKernouve H6 971–985 0.861–0.885

a Distances are normalized to a radius of 100 km.



Finally, we consider the proportion (by volume) of eachpetrologic type thought to contribute to the parent body.Values for our nominal optimal model are 84%, 10%, and6% for type 6, type 4/5, and type 3, respectively. Of themodels run by Bennett and McSween (1996), the one withinitial conditions most closely matching those of our ownyields proportions of 70%, 10%, and 20%, respectively.The Akridge et al. (1998) model yields proportions of88%, 6%, and 6%. That our proportions are closer to those

of Akridge et al. (1998) reflects our use of insulating megar-egolith and regolith layers. Ultimately, though, such pro-portions have limited applicability: Akridge et al. (1998)and Bennett and McSween (1996) note that processes suchas parent body ejection and atmospheric passage, com-pounded by likely changes in flux over time, have almostcertainly introduced significant bias to the proportion ofpetrologic types in the meteorite collection, rendering theseproportions unsuitable for constraining thermal modeling.

Fig. 5. Onion-shell thermal histories (shaded) can be found that pass within the error bars of all available closure time data for the meteoritesindicated. Error bars are replaced with boxes surrounding each data-point (“x”). The range of peak temperatures corresponding to each set ofthermal histories is given in Table 3.



3.5. A perturbed onion-shell model

A broad assessment of the data presented above indi-cates that 95% of closure times and 87% of cooling ratesare consistent with our optimal onion-shell interpretationof early H-chondrite parent body thermal history. Coolingrates derived by fitting theoretical cooling curves to closuretime data (Ganguly and Tirone, 2001; Kleine et al., 2008),also support this interpretation.

In reality, an onion-shell thermal history will not haveproduced a perfect set of closure times and cooling rates.Our thermal model is highly idealized, and the regions itpredicts for the various petrologic types should not be re-garded as having fixed, sharply defined boundaries. Alsouncertain are the peak temperatures we have chosen to dis-tinguish between petrologic types. However, alternativechoices were found to yield no improvements to the fit.

We therefore need to consider processes that can perturbthe onion-shell thermal structure, thereby explaining thesmall group of ill-fitting data without requiring complete

disruption of the parent body (Grimm et al., 2005;Schwartz et al., 2006; Scott et al., 2010). The most likelyprocess is impact cratering due to collisions between theparent body and smaller asteroids. Although it is not cer-tain that all petrologic types need to be excavated (for in-stance, Scott et al., 2010, argue for excavation of type 4sonly), we adopt the conservative endmember assumptionthat all types must be included. Such an impact would haveto excavate to a depth of at least 5.6 km in our 100 km ra-dius optimal model. We employ a commonly used scalinglaw for gravity-dominated impacts (Housen et al., 1983;Asphaug, 1997) to confirm that such an impact is plausible:

rp ¼ 0:41D1:28g0:28v�0:56i :

Here, rp is the projectile radius, D is the transient craterdiameter, g is surface gravitational acceleration, and vi isimpact velocity. We set D = 56 km, since the excavationdepth (which must be 5.6 km) is thought to be about a tenthof the transient crater diameter (Melosh, 1989). For a100 km diameter target with mean density of 3250 kg m�3

Fig. 6. Onion-shell thermal histories (shaded) can be found that pass within the error bars of all available cooling rate data for the meteoritesindicated. Error bars are replaced with boxes surrounding each data-point (“x”). The range of peak temperatures corresponding to each set ofthermal histories is given in Table 3.



(equivalent to the interior density of our model, whichdominates the mean), we have g = 0.091 m s�2. Given D

and g, we seek an impact with a constant value of theexpression rpvi

0.56 = 2.5 � 105. For a wide range of impactvelocities (0.1–10 km s�1, corresponding to projectile radiiof 19 km down to 1.3 km), the impact energy per unit targetmass varies from about 30 to 150 J kg�1. This range isabout an order of magnitude lower than the energy re-quired to shatter the body (�103 J kg�1, defined as the en-ergy which results in fragments that remain gravitationallybound, with the largest fragment equal to half the totalmass of the body; Nolan et al., 2001). The range is also or-ders of magnitude less than the energy required to perma-nently disrupt the body (�105 J kg�1; Housen et al., 1991;Benz and Asphaug, 1999; Nolan et al., 2001).

We can also obtain an upper bound on excavation depthby finding the most energetic impact that just avoids shat-tering the body (i.e., an impact with energy �103 J kg�1).For the range of projectile velocities considered above,maximum excavation depths for such an impact are foundto lie between 9.2 and 13.5 km, or about twice the minimumexcavation depth. An impact that excavated 13.5 km ofmaterial would produce anomalous cooling rates in the fol-lowing approximate volumetric proportions of material(not accounting for the three-dimensional shape of theexcavation zone): 15%, 27%, and 58% for type 3, type4/5, and type 6, respectively. An impact that excavated onlythe minimum 5.6 km of material required to reach type 6swould result in proportions of 36%, 64%, and �0%.

The minimum required transient crater diameter of56 km is significant compared to the target body diameter(200 km). However, even larger impacts have been observedto leave their target bodies intact (e.g., Moore et al., 2004).The largest crater on Vesta, for instance, is estimated tohave had a transient diameter of �300 km (Asphaug,1997), about 60% of the Vesta diameter (530 km). In ourcase, the transient diameter is only 30% of the body diam-eter. A brief look at other parent body radii in our optimalmodel suite (i.e., 75–130 km) indicates that the excavationof type 6 material would require minimum transient craterdiameters ranging from about 40 to 70 km. In all cases, theenergy produced by such impacts is too low to shatter theparent body.

We note briefly that deep excavation by impacts is sug-gested by the observation of all petrologic types in regolithbreccia clasts. However, cooling rates in these clasts (unlikein the matrix material) are generally consistent with theonion-shell model (as seen in the data from Williamset al., 2000, for example), suggesting that impacts excavatedthese particular samples after the body cooled.

4. CONCLUSIONS

In order to discriminate between the traditional onion-shell hypothesis for ordinary chondrite parent body thermalevolution, and the alternative fragmentation–reassemblyhypothesis, we have synthesized available H-chondrite clo-sure time and cooling rate data in order to compare themwith predictions from a spherically symmetric thermalmodel of the parent body. Our approach is to vary model

parameters such as parent body radius, thermal conductiv-ity, and the effectiveness of an insulating regolith, in orderto find an optimal fit between model and data.

Our optimal fit suggests that the data can be explainedwell by a thermally insulated onion-shell thermal historyin a parent body with reasonable thermal and physicalproperties. Not only do 20 of the 21 closure age data,and 62 of the 71 cooling rates fall within the predictedranges for their petrologic types but, in 6 out of 8 cases,plausible thermal histories can be found that pass throughmultiple data from an individual meteorite.

The characteristics of our nominal optimal thermalmodel include a parent body radius of 100 km, a thermaldiffusivity of j(T) = 1.56 � 10�7 + 8.88 � 10�5 T, an inte-rior porosity of 0.06, insulating regolith and megaregolithlayers 410 m thick, and an initial 26Al/27Al of 7.5 � 106.However, the same optimal fit to the data can be achievedby scaling the parent body radius together with eitherporosity or thermal diffusivity. The range of parent bodyradii accessible via this scaling is 75–130 km.

Those meteorite samples with closure times or coolingrates that do not conform to our model may have beenmoved by impact processes to different depths in the parentbody during cooling. Impacts capable of excavating mate-rial of even the highest (i.e., deepest) petrologic types arenot likely to have shattered the parent body, even thoughthey would have produced transient crater diameters of atleast 30% of the parent body diameter. Larger craters (rel-ative to parent body diameter) have been observed to formon certain bodies (such as Vesta) without causing them todisrupt.

ACKNOWLEDGMENTS

Funding for this work was provided by the NASA Outer SolarSystem program. The authors thank Hap McSween and an anon-ymous reviewer for insightful and helpful reviews.

REFERENCES

Akridge G., Benoit P. H. and Sears D. W. G. (1998) Regolith andmegaregolith formation of H-chondrites: thermal constraintson parent body. Icarus 32, 185–195.

Amelin Y., Ghosh A. and Rotenberg E. (2005) Unraveling theevolution of chondrite parent asteroids by precise U–Pb datingand thermal modeling. Geochim. Cosmochim. Acta 69, 505–518.

Asphaug E. (1997) Impact origin of the Vesta family. Meteorit.

Planet. Sci. 32, 965–980.

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. andWalker R. J. (2001) Call for an improved set of decay constantsfor geochronological use. Geochim. Cosmochim. Acta 65, 111–

121.

Bennett M. E. and McSween H. Y. (1996) Revised modelcalculations for the thermal histories of ordinary chondriteparent bodies. Meteorit. Planet. Sci. 31, 783–792.

Benz W. and Asphaug E. (1999) Catastrophic disruptions revisited.Icarus 142, 5–20.

Bischoff A., Rubin A. E., Keil K. and Stoffler D. (1983)Lithification of gas-rich chondrite regolith breccias by grainboundary and localized shock melting. Earth Planet. Sci. Lett.

66, 1–30.



Blinova A., Amelin Y. and Samson C. (2007) Constraints on thecooling history of the H-chondrite parent body from phosphateand chondrule Pb-isotopic dates from Estacado. Meteorit.

Planet. Sci. 42, 1337–1350.

Bouvier A., Blichert-Toft J., Moynier F., Vervoort J. D. andAlbarede F. (2007) Pb–Pb dating constraints on the accretionand cooling history of chondrites. Geochim. Cosmochim. Acta

71, 1583–1604.

COMSOL AB (2008) COMSOL Multiphysics User’s Guide.COMSOL AB, Stockholm, Sweden.

Dodd R. T. (1969) Metamorphism of the ordinary chondrites: areview. Geochim. Cosmochim. Acta 33, 161–203.

Dodd R. T. (1981) Meteorites. Cambridge University Press, NewYork.

Ganguly J. and Tirone M. (2001) Relationship between coolingrate and cooling age of a mineral: theory and applications tometeorites. Meteorit. Planet. Sci. 36, 167–175.

Ghosh A., Weidenschilling S. J. and McSween H. Y. (2003)Importance of the accretion process in asteroid thermalevolution: 6 Hebe as an example. Meteorit. Planet. Sci. 38,

711–724.

Gopel C., Manhes G. and Allegre C. J. (1994) U–Pb systematics ofphosphates from equilibrated ordinary chondrites. Earth

Planet. Sci. Lett. 121, 153–171.

Grimm R. E. (1985) Penecontemporaneous metamorphism, frag-mentation, and reassembly of ordinary chondrite parent bodies.J. Geophys. Res. 90, 2022–2028.

Grimm R. E. and McSween H. Y. (1993) Heliocentric zoning ofthe asteroid belt by aluminum-26 heating. Science 259, 653–

655.

Grimm R. E., Bottke W. F., Durda D., Enke B., Scott E. R. D.,Asphaug E. and Richardson D. (2005) Joint thermal andcollisional modeling of the H-chondrite parent body. LunarPlanet. Sci., #1798 (abstr.).

Harper C. L. and Jacobsen S. B. (1996) Evidence for Hf-182 in theearly solar system and constraints on the timescale of terrestrialaccretion and core formation. Geochim. Cosmochim. Acta 60,

1131–1153.

Hopfe W. D. and Goldstein J. I. (2001) The metallographic coolingrate method revised: application to stone meteorites andmesosiderites. Meteorit. Planet. Sci. 36, 135–154.

Housen K. R., Schmidt R. M. and Holsapple K. A. (1983) Craterejecta scaling laws: Fundamental forms based on dimensionalanalysis. J. Geophys. Res. 88, 2485–2499.

Housen K. R., Schmidt R. M. and Holsapple K. A. (1991)Laboratory simulations of large scale fragmentation events.Icarus 94, 180–190.

Huss G. R., Rubin A. E. and Grossman J. N. (2006) Thermalmetamorphism in chondrites. In Meteorites and the Solar

System II (eds. D. S. Lauretta and H. Y. McSween)). University

of Arizona Press, Tucson, Arizona, pp. 567–586.

Johansen A., Oishi J. S., Low M. M., Klahr H., Henning T. andYoudin A. (2007) Rapid planetesimal formation in turbulentcircumstellar disks. Nature 448, 1022–1035.

Kessel R., Beckett J. R. and Stolper E. M. (2007) The thermalhistory of equilibrated ordinary chondrites and the relationshipbetween textural maturity and temperature. Geochim. Cosmo-

chim. Acta 71, 1855–1881.

Kleine T., Touboul M., Van Orman J. A., Bourdon B., Maden C.,Mezger K. and Halliday A. N. (2008) Hf–W thermochronom-etry: closure temperature and constraints on the accretion andcooling history of the H chondrite parent body. Earth Planet.

Sci. Lett. 270, 106–118.

Lipschutz M. E., Gaffey M. E. and Pellas P. (1989) Meteoric parentbodies: nature, number, size, and relation to present-dayasteroids. In Asteroids (eds. R. P. Brinzel, T. Gehrels and M.

S. Mathews). University Arizona Press, Tucson, Arizona, pp.

740–778.

Lugmair G. W. and Shukolyukov A. (1998) Early solar systemtimescales according to 53Mn–53Cr systematics. Geochim. Cos-

mochim. Acta 62, 2863–2886.

Melosh H. J. (1989) Impact Cratering. Oxford University Press,New York.

Minster J. F. and Allegre C. J. (1979) 87Rb/87Sr chronology of Hchondrites: constraint and speculations on the early evolutionof their parent body. Earth Planet. Sci. Lett. 42, 333–347.

Miyamoto M., Fujii N. and Takeda H. (1981) Ordinary chondriteparent body: an internal heating model. In Proc. Lunar Planet.

Sci., vol. 12B. Lunar and Planetary Institute, Houston. pp.1145–1152.

Moore J. M., Schenk P. M., Bruesch L. S., Asphaug E. andMcKinnon W. B. (2004) Large impact features on middle-sizedicy satellites. Icarus 171, 421–443.

Nolan M. C., Asphaug E., Greenberg R. and Melosh H. J. (2001)Impacts on asteroids: fragmentation, regolith transport, anddisruption. Icarus 153, 1–15.

Pellas R. and Storzer D. (1981) 244Pu fission track thermometryand its application to stony meteorites. Proc. R. Soc. Lond. Ser.

A 374, 253–270.

Pellas P., Fieni C., Trieloff M. and Jessberger E. K. (1997) Thecooling history of the Acapulco meteorite as recorded by the244Pu and 40Ar–39Ar chronometers. Geochim. Cosmochim. Acta

61, 3477–3501.

Renne P. R. (2000) 40Ar/39Ar age of plagioclase from Acapulcometeorite and the problem of systematic errors in cosmochro-nology. Earth Planet. Sci. Lett. 175, 13–26.

Schwartz W. H., Trieloff M., Korochantseva E. V. and Buikin A. I.(2006) New data on the cooling history of the H-chondriteparent body. Meteorit. Planet. Sci. 41, Supplement, 69th Ann.

Meteorit. Soc. Meet. p. 5132.Scott E. R. D. and Rajan R. S. (1981) Metallic minerals, thermal

histories, and parent bodies of some xenolithic, ordinarychondrites. Geochim. Cosmochim. Acta 45, 53–67.

Scott E. R. D., Mandell D., Yang J., Goldstein J. I., Krot T. andTaylor G. J. (2010) Metamorphism and impacts on the parentasteroid of H-chondrites. Lunar Planet. Sci., #1529 (abstr.).

Slater-Reynolds V. and McSween H. Y. (2005) Peak metamorphictemperatures in type 6 ordinary chondrites: an evaluation ofpyroxene and plagioclase geothermometry. Meteorit. Planet.

Sci. 40, 745–754.

Taylor G. J., Maggiore P., Scott E. R. D., Rubin A. E. and Keil K.(1987) Original structures, and fragmentation and reassemblyhistories of asteroids: evidence from meteorites. Icarus 69, 1–13.

Trieloff M., Jessberger E. K. and Fieni C. (2001) Comment on‘40Ar/39Ar age of plagioclase from Acapulco meteorite and theproblem of systematic errors in cosmochronology’ by Paul R.Renne. Earth Planet. Sci. Lett. 190, 267–269.

Trieloff M., Jessberger E. K., Herrwerth I., Hopp J., Fieni C.,Ghelis M., Bourot-Denise M. and Pellas P. (2003) Structureand thermal history of the H-chondrite parent asteroid revealedby thermochronometry. Nature 422, 502–506.

Turner G., Enright M. and Cadogan P. H. (1978) The early historyof chondrite parent bodies inferred from 40Ar–39Ar ages. InProc. Lunar Planet. Sci., vol. 9. Lunar and Planetary Institute,Houston. pp. 989–1025.

Van Schmus W. R. and Wood J. A. (1967) A chemical-petrologicclassification for the chondritic meteorites. Geochim. Cosmo-

chim. Acta 31, 747–765.

Wasson J. T. (1972) Formation of ordinary chondrites. Rev.

Geophys. Space Phys. 10, 711–759.

Williams C. V., Keil K., Taylor G. J. and Scott E. R. D. (2000)Cooling rates of equilibrated clasts in ordinary chondrite



regolith breccias: implications for parent body histories. Chem.

Erde 59, 287–305.

Willis J. and Goldstein J. I. (1981) A revision of metallographiccooling rate curves for chondrites. Lunar Planet. Sci. 12B,

1135–1143.

Wlotzka F. (2005) Cr spinel and chromite as petrogenetic indica-tors in ordinary chondrites: equilibration temperatures ofpetrologic types 3.7 to 6. Meteorit. Planet. Sci. 40, 1673–1702.

Wood J. A. (1967) Chondrites: their metallic minerals, thermalhistories, and parent planets. Icarus 6, 1–49.

Yomogida K. and Matsui T. (1983) Physical properties of ordinarychondrites. J. Geophys. Res. 88, 9513–9533.

Associate editor: Christian Koeber


author's personal copyharrison/harrison... · author's personal copy thermal constraints...

Documents