Condrulos

7/30/2019 Condrulos

1/13

ELSEVIER Earth and Planetary Science Letters 131 (1995) 27-39EPSL

Chondrule formation, metamorphism, brecciation, an importantnew primary chondrule group, and the classification of

chondrulesDerek W.G. Sears, Huang Shaoxiong, Paul H. Benoit

Cosmochemi stt y Group, Department of Chemi stry and Biochemi stry , U niv ersit y of Ar kansas, Fayett euil le, AR 72701, USAReceived 17 May 1994; accepted after revision 9 January 1995

AbstractThe recently proposed compositional classification scheme for meteoritic chondrules divides the chondrules into

groups depending on the composition of their two major phases, olivine (or pyroxene) and the mesostasis, both ofwhich are genetically important. The scheme is here applied to discussions of three topics: the petrographicclassification of Roosevelt County 075 (the least-metamorphosed H chondrite known), brecciation (an extremelyimportant and ubiquitous process probably experienced by > 40% of all unequilibrated ordinary chondrites), andthe group A5 chondrules in the least metamorphosed ordinary chondrites which have many similarities tochondrules in the highly metamorphosed equilibrated chondrites. Since composition provides insights into bothprimary formation properties of the chondrules and the effects of metamorphism on the entire assemblage it ispossible to determine the petrographic type of RC075 as 3.1 with unique certainty. Similarly, the new scheme can beapplied to individual chondrules without knowledge of the petrographic type of the host chondrite, which makes itespecially suitable for studying breccias. Finally, the new scheme has revealed the existence of chondrules notidentified by previous techniques and which appear to be extremely important. Like group Al and A2 chondrules(but unlike group Bl chondrules) the primitive group A.5 chondrules did not supercool during formation, but unlikegroup Al and A2 chondrules (and like group Bl chondrules) they did not suffer volatile loss and reduction duringformation. It is concluded that the compositional classification scheme provides important new insights into theformation and history of chondrules and chondrites which would be overlooked by previous schemes.

1. IntroductionBy virtue of their great age, solar composition,unusual isotopic properties and relatively unal-

tered textures chondrites provide unique insightsinto early solar system processes [l]. One of themajor components of chondrites are the chon-drules [2], whose presence implies that pulseheating was a common phenomenon in the earlysolar system [1,3]. However, the origin of chon-

drules, and their possible relevance to the forma-tion of the major chondrite classes, has provedhighly contentious [4,51. Some of the controversyis due to the great diversity and complexity ofchondrules. The classification of chondrules istherefore an essential step in understanding thiscomplexity and the questions surrounding theseobjects. As part of our efforts in this direction werecently proposed a new classification scheme formeteoritic chondrules [6,7].

0012-821X/95/$09.50 0 1995 Elsevier Science B.V. All rights reservedSSDZ 0012-821X(95)00007-0

7/30/2019 Condrulos

2/13

28 D. WG. Sears et al . Eart h and Pl anetar y Science Lett ers 131 (1995) 27-39

In a subject plagued with a great many classifi-cation schemes it is essential to show restraint inproposing yet another. We certainly endorseWoods comment that bad classification schemescan obscure relationships and even cloud obser-vations [8]. Our reasons for proposing a newscheme were prompted by a very pragmatic issue:In attempting to understand the thermolumines-cence (TL) and cathodoluminescence (CL) prop-erties of chondrules we needed a way of quicklysorting them into meaningful groups, each withreasonably uniform properties, independent ofthe host meteorite. We found that with existingschemes an uncomfortable fraction seemed to beanomalous. Our new scheme for chondrule clas-sification relies on the compositions of the twomajor components, the olivine grains (or pyrox-ene grains in the very few chondrules withoutolivine) and the surrounding mesostasis. This en-ables classification of > 95% of the chondrules ina meteorite. However, having developed the newscheme we quickly realized that there were sev-eral significant advantages to the scheme whichare of a more fundamental kind. This is becauseboth the olivine (or pyroxene) and the mesostasisof chondrules display major compositional trendswhich reflect differences in original formationprocesses and the subsequent overprint of meta-morphism.In a recent paper Scott et al. 191 pointed outseveral ways in which our compositional classifi-cation scheme could be improved, although theyfavored the retention of their previous system.They did not discuss the main strengths of thenew scheme which are that, unlike its predeces-sors, it is simple, comprehensive and, most impor-tantly, lacks assumptions about the original na-ture of a given chondrule and its subsequentalteration. In this paper we briefly summarize ourchondrule classification scheme, including somerefinements added to incorporate the improve-ments suggested by various researchers. We thendiscuss three applications of the new schemewhich demonstrate its utility. Throughout thispaper we attempt to distinguish between (1) themechanical (but very important) procedure ofclassifying chondrules based on their mineralchemistry and (2) interpretations derived from

these data concerning the processes of metamor-phism, brecciation and chondrule formation.1.1. Summary of the composit ional classifi cati on ofchondrules and compari son w it h previ ous schemes

We define eight chondrule classes in terms ofthe compositional boundaries for the chondruleolivines shown in Fig. 1 and the compositionalboundaries for the mesostasis shown in Fig. 2.We do this because (1) these are normally thetwo major components of the chondrules and (2)it has been shown that the chemistry of both are

0.4

0.3

0.23E 0.1.Pg00 0.50"

0.4

0.3

0.2

0, i

00 10 20 30 10 20 30Fe0 (Weight %)Fig. 1. Fields used to define compositional classes for chon-drules using the CaO-Fe0 plot. The fields proposed by Searset al. and DeHart et al. [6,7] are shown (a) together with (b)slightly simplified, and perhaps more realistic, versions. Alsoshown are the fields for equilibrated H, L and LL chondrites.Because chondrules display such large compositional ranges itwould be meaningless to define separate fields for chondrulesfrom unequilibrated H, L and LL chondrites, but minormodification might be required in order to apply these fieldsto enstatite and carbonaceous chondrites.

7/30/2019 Condrulos

3/13

D. W.G. Sears et al . Eart h and Pl anetar y Science Lett ers 131 (1995) 27-39 29

of considerable importance in deciphering chon-drule history. Rather than document thepainstaking and slow process by which chondruleclassification has evolved we have attempted toexplain how previous chondrule classificationschemes relate to the present proposals as foot-notes to Table 1. The present scheme picks up onall of the previously recognized groups, and addsothers, but does so in a simple way and in a waythat is not compromised by metamorphism andbrecciation.

lor [17] showed that these two distinctions couldbe recognized among the chondrules of ordinarychondrites, and they further distinguished be-tween type IA and type IB chondrules, thelatter meaning FeO-poor chondrules in whicholivine was poikilitically enclosed in pyroxene.More recently, Jones [l&19] introduced typeIAB for FeO-poor pyroxene-rich chondrules.Subdivision of the type II chondrules along simi-lar lines is a likely future development.

The redox state of the environment and loss ofvolatiles during chondrule formation have bothbeen discussed by many authors (e.g., [6,7,10-151).Laboratory heating experiments and thermody-namic calculations show that considerable loss ofvolatiles and reduction of Fe0 to Fe should occurat or near the melting point if chondrules wereheated in a gas of even approximately cosmiccomposition, and yet many chondrules show littleor no volatile loss or reduction (e.g., [12]). Therehave been a great many ideas proposed wherebychondrules could be heated to near the meltingpoint in order to obtain their observed texturesand yet not undergo compositional changes.However, McSween [16] showed that a significantnumber of chondrules in carbonaceous chon-drites did appear to be both reduced and poor involatiles. These were termed type I, while theterm type II was reserved for chondrules whichwere not volatile-poor or reduced. Scott and Tay-

Chondrule mesostasis is important not onlybecause it is the major depository of the volatilesbut also because it provides unique insights intothe crystallization history of the chondrules [20-231. Some chondrules (which are termed groupsAl, A2 and A5 in the compositional classificationscheme) have mesostases which are feldspar nor-mative with moderate to low amounts of norma-tive quartz. Other chondrules (which are termedgroup Bl chondrules) have appreciable quartz intheir mesostasis norms even though they containolivine, which is indicative of considerable super-cooling.As many authors have pointed out, both olivineand mesostasis compositions show major changesduring metamorphism [6,7,24-261. Understandingthe effects of metamorphism has proved impor-tant in our efforts to understand both meteoriteformation and history on the parent body andthus the compositional changes to the chondrulesaccompanying metamorphism must also be in-

Table 1An interpretation of the formation of the major chondrule groups in terms of the intensity of the heating event * and subsequentcrystallizationChondrule group Intensity of heating event Cooling rateAl 5 High, reduction and loss of volatiles during crystallization Slow, phenocrysts and mesostasis at equilibriumA2# High, but slightly less reduction and volatile loss than Al Same as AlA5 {a Low, no reduction or mass loss Same as AlBl + Low, no reduction or mass loss Fast, considerable supercooling* Intensity of heating event refers to the chemical changes caused by the chondrule formation event. The major factor responsiblefor these was almost certainly temperature, but time spent at the peak temperature during formation, cooling rate, and absence ofinsulating materials (especially oxygen-rich dust) will also influence the effect of the heat pulse. Includes some of the dropletchondrules of Kieffer [41], some of the non-porphyritic pyroxene chondrules of Gooding and Keil [42], the type I chondrules ofMcSween [16], metal-rich microporhyritic chondrules of Dodd [43], and the type IA chondrules of Scott and Taylor [17]. # Includesthe poikilitic pyroxene and type IB chondrules of Scott and Taylor [17] and many of the type IAB chondrules of Jones [l&19].@ There appear to be no previous observations of this chondrule group in unmetamorphosed meteorites. + Dodds [44] lithic orelastic chondrules and Dodds [43] metal-poor microporphyritic chondrules are included in this group, as are the type IIchondrules of McSween [161,Scott and Taylor [17] and Jones [30].

7/30/2019 Condrulos

4/13

30 D. W.G. Sears et al. Earth and Planetary Science Letters 131 (1995) 27-39eluded in the chondrule classification. Placing thefield boundaries so as to be of maximum value isvery difficult and will no doubt improve as workprogresses, especially as we extend it to enstatiteand carbonaceous chondrites.1.2. Chondrule classificati on: Ol iv ine composit ion

The primary compositional parameters forchondrule olivine are the CaO and Fe0 contents(Fig. 1). The compositional boundaries we usedfor olivine in our earlier papers [6,7] (Fig. la)were placed so as to enclose existing data for asmall group of meteorites, but an alternative andpossibly preferable approach is to draw straightlines across the plot in the fashion shown in Fig.lb. This results in a simpler scheme and fills theavailable CaO-Fe0 space, especially the gap be-tween groups A3 and A4, a region which, how-ever, we note is poorly populated on the basis ofcurrently available databases. We emphasize thatthe simplifications to our original graph do notresult in the wholesale changing of classificationsfor previously classified chondrules; nor do theychange the luminescence distinctions we havenoted between the various chondrule groups.We do not know the exact trajectory of theolivine compositions across the diagram duringmetamorphism, because compositional zoning aswell as chondrule size and grain-size effects resultin unique paths for each individual grain. Thecurved arrows in Fig. lb describe some possibili-ties. The major trajectories seem to be Al + A3-+A4-+A5, A2-+A5, Al-+A2-+A5 and Bl+B2*B3+A5, but even Al+A2+A3+A4 -+ A5 is possible, although the latter trajec-tory is probably the least likely given the verynarrow range of overlap in calcium contents be-tween these two chondrule groups.1.3. Chondrule classi fi cati on: Mesostasis composi -t ion

Fig. 2 shows the available data for Semarkona(type 3.0) chondrule mesostasis compositions. Wehave now completed three major studies of chon-drule compositions for type 3 ordinary chondrites[26-281. Each project was performed with differ-

DeHart - JSC microprobe HutchisonlAlexander- OUiBM microarobe

Ab An Ab AnLu Jie - AMNH microprobe

Ab An Ab

Jones/Scott - UNM microprobe

0 1 . . . $A$Ab An

Fig. 2. Comparison of mesostasis compositions for Semarkonachondrules determined by a variety of analysts using differentelectron microprobes. J. DeHart with the Johnson SpaceCenter microprobe [6,25], R. Hutchison and C. Alexanderwith the British Museum and Open University microprobes[28] and L. Jie with the American Museum of Natural Historymicroprobe 112,261produced very similar data, while the dataobtained by R. Jones and E. Scott with the University of NewMexico microprobe are quite different, reflecting their use ofa broader analysis beam 122,281.

ent collaborators, and thereby different analystsand different electron microprobes, and each haseither been published [6,7], is in press [28], or isabout to be submitted [29]. The number of chon-drules we have studied in detail is now nearly 200and includes a wide selection of those present inour research samples, rather than a select subset

7/30/2019 Condrulos

5/13

D. KG . Sears et al . Eart h and Pfanefary Science Lett ers I 31 (1995) 27-39 31

picked on the basis of a variety of textural andmineralogical criteria. All our studies show thatmany chondrules in Semarkona (the group Blchondrules) contain over (sometimes well over)50 mol% normative quartz when the data areprojected onto the normative quartz-albite-anorthite ternary diagram. The Scott et al. [9]Semarkona data are also shown in Fig. 2 andclearly differ from the data from the other labo-ratories. The differences largely reflect differ-ences in analytical technique, in turn reflectingdifferences in intended application. The litera-ture data shown by Scott et al. [9] for 29 chon-drules were obtained in an attempt to determinethe composition of residual melt and were thuscollected using relatively large beam diameters inorder to integrate glass and quench crystals[23,30]. In contrast, the other analysts use smallerbeam diameters confined to the glass (includingcrystal microlites which are not apparent in opti-cal microscopy). We note that, under the latterapplication, the mesostasis to be analyzed in-evitably exhibits CL, which can be used to iden-

tify appropriate regions in the electron micro-probe prior to analysis [6,7,26,27].For the few instances of glassy or cryptocrys-talline chondrules a meaningful classification canusually be obtained by plotting a defocussed elec-tron beam analysis of the entire chondrule, or anaverage of several regions of the chondrule, onthe mesostasis plot. Such an application of thescheme cannot distinquish between some groupsof chondrules, especially for the Al-A4 serieswhich differ in their definitions largely in themineral chemistry of olivine phenocrysts. Furtherwork is needed on these rare chondrules, but wefeel that it is appropriate to at least initiallygroup them with other chondrules with whichthey at least share some chemical similarity.1.4. Facility of the compositional classificationscheme

It is worth detailing briefly the relative easewith which the various chondrule classificationschemes can be applied. By applying the miner-

I Metamorphism 16 -1Samarkona RC 075 Ktymka Chainpur ALHA 4 Dhajala Bremoworde 5arwell

3.0 3.1 I 3.4 3.5 3.8 3.9 5

Fig. 3. The distribution (relative abundance by number) of chondrules over the compositional classes for six type 3 ordinaryordinary chondrites and Barwell (L5). The classifications were made from low-power photomicrographs of sections, using the CL todetermine the compositions of the major phases in the chondrules. It is possible to assign virtually all the chondrules in the sectionsto compositional groups and, since composition and therefore the chondrule group varies with metamorphism (as well as formationconditions), it is possible to assign the especially significant Roosevelt County 075 to a petrographic type of 3.1.

7/30/2019 Condrulos

6/13

32 D. W G. Sears et al . Eart h and Planetary Science Lett ers 131 (1995) 27-39

alogical-textural type scheme to the chondrulesseparated from their host meteorite during one ofour projects we were able to classify 5 out of 23chondrules [12,31], and in the light of recentchanges introduced by Jones [l&19] could stillonly classify a further 4 chondrules. The remain-ing chondrules simply did not display the largearray of textural and mineralogical properties nowincluded in the definitions of types IA to IIAB,whether the chondrules were examined by us orby our collaborators. When the compositionalclassification scheme is applied to chondrules inthin section > 95% of the chondrules were suc-cessfully classified in a number of different type 3ordinary chondrites (Fig. 3).

Most of our studies have involved chondruleswhich have been physically removed from themeteorite and represent more challenging tests ofthe schemes since these are intended to be ran-dom samples of chondrules rather than subsetshomogeneous in one or more properties. Wehave examined data for 107 chondrules from thestudies of DeHart [26] and Lu [27]. Only 17chondrules were either devoid of olivine or theolivine plotted outside the fields, while for only 8chondrules were there no mesostasis data. It waspossible to classify 100 chondrules when bothparameters were used. In about two-thirds of thecases mesostasis alone would have enabled classi-fication, while in about one-third of the casesolivine compositions are alone sufficient.

Perhaps the strongest difference between thetwo classification schemes lies in their major em-phasis. Our proposed scheme classifies chon-drules on the basis of major element chemistry ofmajor chondrule phases (i.e., olivine and mesosta-sis). As such, it can be applied with little or nodisagreement by anyone, including those with onlylimited experience with chondrules or with expe-rience of the more qualitative aspects of chon-drule petrographic description. The previous clas-sification scheme was originally based largely onpetrographic textures and was thus subjective andrequired a fair degree of previous knowledge ofrelative textures before it could be successfullyapplied. Later work added mineral chemistry tothis scheme, but only as an adjunct to texture andonly in an individual meteorite (i.e., Semarkona).

A key assumption in this scheme is that texture,to which petrographers were necessarily biased inthe days before the wide availability of electronmicroprobes, must reveal more about the historyand origin of chondrules than their major ele-ment chemistry. We argue that the availability ofelectron microprobes and the progressively in-creasing ease of their use, as well as increasinguncertainty in the basis of laboratory studies ofthe significance of chondrule texture, are goodreasons to move to a more quantitative classifica-tion scheme.

1.5. Pyroxene compositionAlthough Huang et al. [29] were able to useolivine for the classification of 90 of their 107chondrules, a few chondrules contain only olivine

which is too fine grained to analyze or containonly pyroxene. Obviously one cannot simply plotpyroxene data on the olivine mineral chemistrygraph (Fig. la) although, in analogy to glassychondrules discussed above, the mesostasis com-positions of these chondrules can be used toobtain an approximate classification in most cases.We anticipate that a pyroxene analog to Fig. 1will eventually be added to the classificationscheme, although the literature data on which toproduce such a plot are at present scarce. Ourpresent applications, however, are examinationsof chondrule populations in the broadest sense,and for such general applications the scheme inits present form is sufficient.

2. The petrographic classification of RooseveltCounty 075While LL chondrites of petrologic type < 3.4

are relatively common it has been only recentlythat a few H chondrites of type < 3.4 have beenreported. One of them is the heavily weatheredRoosevelt County 075 [32]. Weathering and thelack of equilibration make classification uncer-tain, but it is probably an H chondrite. Weather-ing also makes it very difficult to assign a petro-logic type. For example, removal of the weather-

7/30/2019 Condrulos

7/13

D. W.G. Sears et al. /Earth and Planetaty Science Letters 131 (1995) 27-39 33ing products by acid-washing increased the ther-moluminescence sensitivity of RC075 by a factorof N 7, equivalent to a change in petrologic typeestimate from 3.0 to 3.3-a major difference.Since the compositional classification scheme re-flects metamorphic effects as well as primarychondrule differences and can be applied to vir-tually all the chondrules present, it is well suitedto the assignment of a petrologic type to RC07.5by comparison with the chondrule populations ofother meteorites of known petrologic type.Cathodoluminescence properties of the phasesin chondrules are simply related to composition,so that it is possible to assign all the chondrulesin our 7 x 5 mm section of RC07.5 to composi-tional classes using their CL. While not as fine-tuned as the quantitative classification by mineralchemistry, the use of luminescence permits rapid,reproducible classification of virtually a ll chon-drules in a given thin section, and thus permitsstatistically significant sampling of chondrulepopulations. The apparatus used in the presentapplication is a relatively simple commerciallyavailable attachment (a nuclide luminoscope) foran optical microscope and our operating condi-tions (e.g., for the electron beam) are typical forroutine petrographic applications. Contrary to anapparently misunderstood personal communica-tion cited by Scott et al. [9], we know of no casewhere minerals luminesce blue in the lu-minoscope CL apparatus and red in the electronmicroprobe, and we know of no physical way toachieve such a change in the CL activator. Steelehas performed a detailed series of studies on theCL properties of meteoritic minerals using anelectron microprobe [33] and we know of in-stances where his data differ from our own obser-vations with a luminoscope.The results of applying our classificationscheme to the chondrules in RC07.5 are shown inFig. 3, along with similar data for seven otherordinary chondrites [6,7]. The relative abundanceof group B chondrules in RC075 is less than inthe type 3.0 ordinary chondrite Semarkona, andcomparable with the higher types, while theabundance of A5 chondrules is comparable tothat in the type 3.1 chondrite Krymka and inter-mediate between that in Semarkona and Chain-

pur (type 3.4). Most significantly the fraction ofAl chondrules is very large and comparable(within error) to that of Semarkona, while thelarge number of group A3 chondrules is compa-rable only to Krymka. Although RC075 is a heav-ily weathered meteorite we do not believe thatweathering has affected our observations; ourprevious observations on chondrules from othermeteorite finds indicate that their chondruleshave TL properties similar to those of unweath-ered meteorites of similar petrologic type. On thebasis of these data RC075 would appear to beintermediate to Semarkona and Chainpur andcomparable to Krymka in its petrologic type.

McCoy et al. [32] report means ranging from0.07 to 7.2 mol% Fa and 0.11 to 0.36 wt% CaOfor olivine in six type IA chondrules and 12.3-20.2mol% Fa for five type II chondrules in RC075.Four of the type IA chondrules resembled thoseof Semarkona in olivine composition (< 2 mol%Fa). Unlike the compositional classificationscheme, which leads fairly simply to unambiguouspetrologic type assignment, it seems difficult toassign RC075 to a petrologic type on the basis ofolivine compositions and texture alone. It is quitedifficult to learn anything from the relative abun-dance of type I and type II chondrules usingpurely textural definition of types in the fashionof McCoy et al. [25]-even if these could bedetermined for a significant proportion of thechondrules present-since they are independentof metamorphism. The compositional classifica-tion scheme provides one of the best methods forassigning the weathered and highly unequili-brated RC075 meteorite to a specific petrologictype, and suggests that it is type 3.1. Thermolumi-nescence sensitivity, which is directly related toCL intensity (blue wavelengths in particular), pro-duces a similar result, but, as a bulk technique, ismore affected by weathering. In any case, it isapparent that RC075 is the lowest petrologic typeH chondrite presently known.

Some points to stress in favor of adopting thecompositional classification scheme are that (1)the technique is feasible because major changesin the composition of chondrules occur with smallchanges in metamorphism, and (2) being compre-hensive it is possible to make meaningful state-

7/30/2019 Condrulos

8/13

34 D. W.G. Sears et al . Eart h and Planetar y Science Lett ers 131 (1995) 27-39ments about the proportion of various chondrulegroups.

3. Brecciation in type 3 ordinary chondritesMore than 40% of type 3 ordinary chondrites

are known or strongly suspected to be breccias[34]. This figure is almost certainly an underesti-mate because research samples and sections tendto be very small, especially in comparison withnormal hand specimens. It is possible to use somebulk properties, such as carbon content, olivineheterogeneity and thermoluminescence sensitiv-ity, to estimate the mean degree of metamor-phism experienced by the meteorite or at least toestimate the range of metamorphism exhibited bysubregions of the meteorite. To assume that anindividual chondrule in a breccia has been meta-morphosed to exactly the same degree as themean meteorite is, however, a dangerous fallacy,and to assume that all meteorites are not brec-ciated until proven otherwise is a still greaterfallacy.

Ngawi is a spectacular example of a brecciatedchondrite with bulk properties (measured, forexample, by induced thermoluminescence, carboncontent and olivine heterogeneity) equivalent topetrologic type - 3.7. It is apparent that thismeteorite is brecciated on the thin-section (< 1cm) scale (Fig. 4). Scott et al. [35] publishedolivine composition data for both the brecciatedALHA 77278 and Ngawi and these data are su-perimposed on the fields which are supposed tocorrespond to the petrologic type of the hostchondrite [9] in Fig. 5. Clearly the data for bothmeteorites spread across the entire diagram, indi-cating that a knowledge of the average petro-graphic type is a poor guide to expected chon-drule composition in brecciated meteorites.We obtained a CL mosaic of a thin section ofNgawi and classified all the chondrules in thesection on the basis of their luminescence. Asketch of this section is shown in Fig. 4. It isapparent, as was also apparent in the chemicaldata (Fig. 5), that there is a great diversity ofchondrules in this single section. The presence ofA4, A5 and B3 chondrules is expected consider-

Fig. 4. Sketch of the texture of Ngawi with the chondrulesassigned to compositional classes using their CL properties.Ngawi is heavily brecciated, containing low petrographic typeclasts in a host matrix which is largely of high petrographictype but with chondrules of all compositional groups. Accord-ing to most conventional criteria Ngawi is type 3.7.

ing the average high petrologic type of the hostmeteorite, in analogy to the Dhajala meteorite(Fig. 3). However, the other chondrule groups,including group Al, A2 and B1,2 chondrules, arealso present in significant numbers. In this themeteorite resembles the most unequilibratedchondrites, such as Krymka, Semarkona, and RC075 which was discussed above (Fig. 3).

On the basis of these data we find that settingapparent metamorphic boundaries for chondrules(Fig. 5) vastly oversimplifies the complexity ofchondrules, especially in brecciated meteorites.On the basis of Fig. 5 one might expect all ormost chondrules in a meteorite described as ofsome average petrologic type to exhibit very re-stricted chondrule compositions. In reality this isnot necessarily the case in many meteorites. Weprefer to describe chondrules only on the basis of

7/30/2019 Condrulos

9/13

D. W.G. Sears et al. /Earth and Planeta ry Science Lett ers 131 (1995) 27-39 35

i o 15FE&Vt%

25 30

Fig. 5. CaO-Fe0 for chondrule olivines in Ngawi and ALHA77278 [33]. The fields are those proposed by Scott et al. [9],the upper figure describing how the composition of type IAchondrules varies with petrographic type; the lower diagramrefers to type II chondrules. Clearly, knowing the type (IA,IB, IAB, II) of a chondrule and the petrologic type of the hostmeteorite does not describe the chondrule. At least 40%, andprobably most, of the type 3 chondrites are breccias and it isessential to characterize chondrules in terms of direct obser-vations of the present chondrules.

the chemical compositions (or, as in this case, onthe basis of the luminescence properties whichderive from the mineral chemistry). Such a sys-tem makes no assumptions about the response ofindividual chondrules to bulk rock metamor-phism.

4. The group AS chondrulesThe presence of group A5 chondrules in highlyunequilibrated meteorites such as Semarkona isof great interest 1361. The group A5 chondrulesmake up essentially 100% of the chondrules in

equilibrated ordinary chondrites (petrologic types5 and 6) and their abundance increases as afunction of petrologic type within the unequili-brated type 3 ordinary chondrites (Fig. 3) and onthis basis one might expect that they are a prod-uct of the high-temperature metamorphism ofchondrules. Their presence in meteorites whichhave experienced little or no metamorphism andin significant numbers (about 15% of all chon-drules in these meteorites) is thus an entirely newand unexpected discovery.

The group A5 chondrules stand out ratherspectacularly in low-magnification CL images be-cause of their bright blue mesostases and non-luminescent mineral grains. (By contrast the othercommon chondrules in the most unequilibratedchondrites have quite different luminescenceproperties, group Al,2 chondrules having brightyellow CL and group Bl chondrules being non-luminescent.) The CaO and Fe0 of the olivine inthe A5 chondrules of several LL chondrites cov-ering a range of average petrologic types areshown in Fig. 6. Like the other chondruIe groups,the olivines of group A5 chondrules appear todecrease in CaO and increase in Fe0 with in-creasing metamorphism, but much less than forthe other group A chondrules and somewhatsimilar to the group B chondrules. The olivineheterogeneity (expressed as the coefficient of

0 5 10 15 20 25 30 35Fe0 (wt%)Fig. 6. The CaO-Fe0 composition of olivine in the new groupA5 chondrules. With metamorphism, CaO decreases and Fe0increases. It would be a simple matter to subdivide the groupto distiquish metamorphically derived subgroups. However,we refrain from doing so pending an examination of A5chondrules in further chondrites. (Data are from [26,27,45,46].)

7/30/2019 Condrulos

10/13

36 D. W.G. Sears et al . Eart h and Planetar y Science Lett ers 131 (1995) 27-39

Semarkona 3.0

ALHA 4 3.4

Fig. 7. Mesostasis compositions for group A.5 chondrules. Thedata change little with petrographic type of the host mete-orite. ALHA 77214 contains numerous chondrules which arecompositionally zoned, being B3 at the center and A5 on theoutside. Previous chondrule classification schemes wouldoverlook this important property of this type 3.4 chondrite.

variation) increases from 5-20% in types 3.0-3.1to about 50% in type 3.5, and then decreases to< 2% in types 4-6. Low-Ca pyroxenes increase inCaO and Fe0 with metamorphic type in A5chondrites. Mesostasis compositions (Fig. 7) arevery similar in all A5 chondrules, with few readilydelineated metamorphic trends. Our INAA datafor group A5 chondrules from unequilibratedchondrites show them to have unfractionated bulkcompositions relative to homogenized CI chon-drite.

The compositional classification scheme forchondrules does not, at present, differentiate be-tween metamorphosed and unmetamorphosed A5chondrules. Small differences certainly exist.Metamorphism causes homgenization and a smallincrease in the Fe0 content of the olivines, forinstance. The olivines in group Bl and B3 chon-drules show similar differences. However, theseare small compared to the very large changes inmesostasis composition observed for group Blchondrules but not observed for group A5 chon-drules. Future work might make it possible tosubdivide the A5 group, but this should be donewith caution. It currently seems to us that thesimilarities between the group A5 chondrules inunequilibrated and equilibrated chondrites aremore important than their differences.

The changes in compositional heterogeneityobserved among the olivines and pyroxenes ofgroup A5 chondrules suggests that the group A5chondrules in Semarkona were not metamor-phosed prior to emplacement in the meteorite,and the presence of A5 chondrules in Semarkonais not due to brecciation. Rather, we interpretthis as meaning that the chondrule-forming pro-cess was capable of producing chondrules withmineral phase compositions resembling those ofchondrules in equilibrated chondrites. Based onmineral chemistry we can make a number ofinterpretations about the chondrule-forming pro-cess. There are several lines of evidence (sum-marized by [37] and 1381)which suggest that groupAl chondrules experienced higher temperaturesduring formation than other chondrules. We havepreviously argued that this also resulted in theirvolatile-poor and highly reduced compositions[12,29]. Conversely, the volatile-rich, oxidizedgroup Bl chondrules apparently experiencedlower peak temperatures during formation. Thecomposition of the mesostasis relative to theolivine/pyroxene grains allows us to infer super-cooling in the case of group B chondrules but notin the case of group A chondrules, including A5[20-23,291. We suggest that the four chondrulegroups in Semarkona represent three very differ-ent thermal histories. The group Al and A2chondrules appear to have formed at relativelyhigh temperatures and did not experience super-

7/30/2019 Condrulos

11/13

D. +? . Sears et al . Eart h and Planeta ry Science Lett ers 131 (I 995) 27-39 31

cooling during subsequent crystallization, whilethe group Bl chondrules formed at lower temper-atures but with supercooling. The group A5 chon-drules may have formed at low temperatures,similar to group Bl chondrules, but did not expe-rience supercooling (Table 1). We suggest thatthe cooling history of chondrules must have beenindependent of the intensity of the heating pulsewhich formed them and probably depended onlocal gas and dust density. Zoned mesostasis andsulfide- and metal-rich rims around group Achondrules might imply an environment suitablefor recondensation of volatiles for group A [39,40].We note, however, that these are interpretationsand we do not suggest classifying chondrules ac-cording to the presumed nature of their forma-tion process or subsequent cooling history.

5. The classification of chondrules: Some con-cluding remarks

Classifications should be based on objectivedescriptions of the objects being classified, con-centrating on the most fundamental properties ofthe objects (i.e., those of greatest importance withrespect to origin and history). Most certainly theyshould not be model-dependent or dependent onassumptions about the original properties of theobjects or what has happened to them since theirformation, regardless of how sensible those as-sumptions may seem at any given moment intime. We suggest that the most fundamental wayof describing a chondrule is in terms of the com-position of its two major components, both ofwhich reflect initial formation and metamorphismand both of which have been shown by a greatmany studies to provide important insights intochondrule history. It seems to us that to definegroups of chondrules in terms of the compositionof one phase and texture and then to adjust thecompositional parameter depending on the petro-graphic type of the host chondrite is to makeunreasonable generalizations about the originalnature of the chondrule and its response to meta-morphism. It also seems a mistake to ignorebrecciation, the influence of which is probablygreatly underestimated. The response to meta-

morphism of a given chondrule is, of course, verycomplicated and will depend on a great manyfactors, such as chondrule texture, chondrule size,the nature of adjacent phases, and the presenceof cracks through which diffusion or other modesof transport could occur. Most serious of all,brecciation will cause chondrules of very differentmetamorphic histories to appear adjacent to eachother in the same meteorite, and we believe thatmost, and probably all, type 3 ordinary chondritesare breccias.It is certainly true that the compositional clas-sification scheme needs further work. It took 150years to arrive at the present situation, with con-siderable redundancy in each step and a greatmany small steps. Some of the initial criticisms ofthe compositional classification scheme are pro-viding the first contribution to polishing thescheme, and are reflected in adjustments to thecompositional boundaries shown in Fig. lb.

We end our paper by returning to the ideathat a bad classification scheme can hinder ratherthan help our research efforts. McSween [16]introduced type I and type II during his studiesof C2-3 chondrites, meaning essentially chon-drules with FeO-rich and FeO-poor olivines. Herightly recognized the importance of the redoxstate of chondrules in our efforts to understandchondrules and chondrites. However, in applyingthe scheme to ordinary chondrites, olivine com-position and texture have become confused, sincetextures change little but compositions changeconsiderably during metamorphism. Thus we havepotentially confusing anomalies, such as type Ichondrules being originally defined, and nowwidely known as, FeO-poor when in fact FeO-poor type I chondrules are very rare and presentin only a handful of ordinary chondrites. Thuswhen a serious study of metamorphism of ordi-nary chondrite chondrules was undertaken byMcCoy et al. [25] they had to substitute thewell-known and commonly understood definitionof type I with a purely textural definition. Thispoint alone underlines the need for a betterclassification scheme. We argue instead that weshould anchor chondrule studies (and classifica-tion) in the most fundamentally important andmost objectively determined properties of the

7/30/2019 Condrulos

12/13

38 D. W. G. Sears et al. Ear th and Planetary Science Letters 131 (1995) 27-39chondrules, namely the chemical composition oftheir major phases as determined by electronmicroprobe or luminescence techniques.

Acknowledgements [lOIWe are grateful to John Wood, Robert

Hutchison, and Rhian Jones, Martin Prinz andCone1 Alexander, for their reviews of thismanuscript, and to Ed Scott and Rhian Jones fordiscussions of chondrule classification and forloaning us their thin sections. We also wish tothank Steve Symes for help with mansucriptpreparation, Walter Manger for loan of thecathodoluminescence apparatus, and Gary Lof-gren, Martin Prinz, Robert Hutchison, Lu Jie,John DeHart, Andrew Morse, Cone1 Alexander,Mike Lipschutz and Jeff Grossman for severalyears of rewarding collaborative research onchondrules. This work was funded by NASA grantNAGW 3519. [UC]

References 1161[II

[21131

[41

151

[61

[71

J.F. Kerridge and MS. Matthews, Meteorites and theEarly Solar System, 1269 pp., University of Arizona Press,Tucson, 1988.E.A. King, Chondrules and their Origins, 377 pp., LunarPlanet. Sci. Inst., Houston, 1983.J.A. Wood, Meteoritic constraints on processes in thesolar nebula, in: Protostars and Planets II, D.C. Blackand M.S. Matthews, eds., pp. 687-702, University ofArizona Press, Tucson, 1985.C.M. Wai and J.T. Wasson, Nebular condensation ofmoderately volatile elements and their abundances inordinary chondrites, Earth Planet. Sci. Lett. 36, 1-13,1977.E. Anders, Critique of Nebular condensation of moder-ately volatile elements and their abundances in ordinarychondrites by CM. Wai and J.T. Wasson, Earth Planet.Sci. Lett. 36, 14-20, 1977.J.M. DeHart, G.E. Lofgren, J. Lu, P.H. Benoit andD.W.G. Sears, Chemical and physical studies of chon-drites X: Cathodoluminescence studies of metamorphismand nebular processes in type 3 ordinary chondrites,Geochim. Cosmochim. Acta 56, 3791-3807, 1992.D.W.G. Sears, J. Lu, P.H. Benoit, J.M. DeHart and G.E.Lofgren, A compositional classification scheme for mete-oritic chondrules, Nature 357, 207-211, 1992.

@I[91

11111121

[131

[141

[151

1171

WI

[191

DO1

I211

[221

[231

J.A. Wood, Chondritic meteorites and the early solarnebula, Annu. Rev. Earth Planet. Sci. 16, 53-72, 1988.E.R.D. Scott, R.H. Jones and A.E. Rubin, Classification,metamorphic history, and pre-metamorphic compositionof chondrules, Geochim. Cosmochim. Acta 58, 1203-1209, 1994.J.A. Wood, A. Hashimoto and B.B. Holmberg, Chon-drules as near-equilibrium assemblages that formed infractionated systems, Lunar Planet. Sci. XX, 1215-1216,1989.A. Kracher, Why do some chondrules have high primaryFe0 contents?, Lunar Planet. Sci. XVI, 467-468, 1985.J. Lu, D.W.G. Sears, B. Keck, M. Prinz, J.N. Grossmanand R.N. Clayton, Semarkona type I chondrules com-pared with similar chondrules in other classes, LunarPlanet. Sci. XIII, 720-721, 1990.H. Nagahara, I. Kushiro and B.O. Mysen, Evaporation ofolivine: Low pressure relations of the olivine system andits implication for the origin of chondritic components inthe solar nebula, Geochim. Cosmochim. Acta 58, 1951-1964, 1994.A. Tsuchiyama, H. Nagahara and I. Kushiro, Volatiliza-tion of sodium from silicate melts and its application tothe formation of chondrules, Geochim. Cosmochim. Acta45, 1345-1367, 1981.A. Hashimoto, Evaporation metamorphism in the earlysolar nebula-evaporation experiments on the melt FeO-MgO-SiO,-CaO-AI,O, and chemical fractionations ofprimitive materials, Geochem. J. 17, 111-145, 1983.H.Y. McSween, Jr., Chemical and petrographic con-straints on the origin of chondrules and inclusions incarbonaceous chondrites, Geochim. Cosmochim. Acta 41,1777-1790, 1977.E.R.D. Scott and G.J. Taylor, Chondrules and othercomponents in C, 0, and E chondrites: similarities intheir properties and origins, Proc. Lunar Planet. Sci.Conf. 14th, B275-B286, 1983.R.H. Jones, Classification of porphyritic, pyroxene-richchondrules in the Semarkona ordinary chondrite, LunarPlanet. Sci. XXIII, 629-630, 1992.R.H. Jones, Petrology of FeO-poor, porphyritic pyroxenechondrules in the Semarkona chondrite, Geochim. Cos-mochim. Acta, in press, 1994.M. Kimura and K. Yagi, Crystallization of chondrules inordinary chondrites, Geochim. Cosmochim. Acta 44,589-602, 1980.M. Blander, H.N. Planner, K. Keil and L.S. Nelson, Theorigin of chondrules: Experimental investigation ofmetastable liquids in the system Mg,SiO,-SiO,,Geochim. Cosmochim. Acta 40,889-896, 1976G.E. Lofgren and W.J. Russell, Dynamic crystallizationof chondrule melts of porphyritic and radial pyroxenecomposition, Geochim. Cosmochim. Acta 50, 1715-1726,1986.R.H. Jones and E.R.D. Scott, Petrology and thermalhistory of type IA chondrules in the Semarkona (LL3.0)

7/30/2019 Condrulos

13/13

D. W.G. Sears et al . Eart h and Pl anetar y Science Lett ers 131 (1995) 27-39 39

chondrite, Proc. Lunar Planet. Sci. Conf. 19th, 523-536,1989.

[24] R.T. Dodd, W.R. Van Schmus and D.M. Koffman, Asurvey of the unequilibrated ordinary chondrites,Geochim. Cosmochim. Acta 31, 921-951, 1967.

[25] T.J. McCoy, E.R.D. Scott, R.H. Jones, K. Keil and G.J.Taylor. Composition of chondrule silicates in LL3-5chondrites and implications for their nebular history andparent body metamorphism, Geochim. Cosmochim. Acta55, 601-619. 1991.

[26] J.M. DeHart, Cathodoluminescence and microprobestudies of the unequilibrated ordinary chondrites, Ph.D.Thesis, Univ. Arkansas, Fayetteville, 1989.

[27] J. Lu, The physical and chemical studies of chondrulesfrom type 3 ordinary chondrites, Ph.D Thesis, Univ.Arkansas, Fayetteville, 1992.

[28] D.W.G. Sears, A.D. Morse, R. Hutchison, R.K. Guimon,J. Lu, C.M.OD. Alexander, P.H. Benoit, I. Wright, C.T.Pillinger, T. Xian and M.E. Lipschutz, Metamorphismand aqueous alteration in low petrologic type ordinarychondrites, Meteoritics 30. in press, 1995.

[29] S. Huang, J. Lu, M. Prinz, M.K. Weisberg, P.H. Benoitand D.W.G. Sears, Chondrules: Their diversity and therole of open-system processes during their formation,Icarus, in prep., 1995.

[30] R.H. Jones, Petrology and mineralogy of type II chon-drules in Semarkona (LL3.0): Origin of closed-systemfractional crystallization, with evidence for supercooling,Geochim. Cosmochim. Acta 54, 1785-1802, 1990.

[31] J. Lu, D.W.G. Sears, P.H. Benoit, M. Prinz and M.K.Weisberg, Related compositional and cathodolumines-cence trends in chondrules form Semarkona, Meteoritics26, 317, 1991.

[32] T.J. McCoy, K. Keil, R.D. Ash, A.D. Morse, CT.Pillinger, R. Wieler, T.K. Mayeda, R.N. Clayton, P.H.Benoit, D.W.G. Sears, I. Casanova, D.W. Muenow, C.B.Moore, C.F. Lewis and I.E. Wilson, Roosevelt County075: A petrologic, chemical and isotopic study of themost unequilibrated known H chondrite, Meteoritics 28.681-691, 1993.

[33] I. Steele, Mineralogy of meteorites revealed by cathodo-luminescence, in: Spectroscopic Characterization of Min-erals and their Surfaces, L.M. Coyne, S.W.S. McKeeverand D.F. Blake, eds.. Am. Chem. Sot. Symp. Ser. 415,150- 164.

[34] D.W.G. Sears, F.A. Hasan, J.D. Batchelor and J. Lu,

Chemical and physical studies of type 3 chondrites XI:metamorphism, pairing, and brecciation of ordinarychondrites, Proc. Lunar Sci. Conf. 21th, 493-512, 1991.

[35] E.R.D. Scott, T.J. McCoy and K. Keil, Post-metamorphicbrecciation in type 3 ordinary chondrites, Lunar Planet.Sci. XXIV, 1267-1268, 1993.

[36] S. Huang, P.H. Benoit and D.W.G. Sears, Group A5chondrules in ordinary chondrites: their formation andmetamorphism, Lunar Planet. Sci. XXV, 573-574, 1994.

[37] R.H. Hewins, Experimental studies of chondrules, in:Meteorites and the Early Solar System, J.F. Kerridge andM.S. Matthews, eds., pp. 660-679, University of ArizonaPress, 1988.

[38] G.E. Lofgren, Dynamic crystallization of chondrule meltsof porphyritic olivine composition: Texture experimentaland natural, Geochim. Cosmochim. Acta 53. 461-470.1989.

[39] S. Huang, P.H. Benoit and D.W.G. Sears, Metal andsulfide in Semarkona chondrules and rims: Evidence forreduction, evaporation and recondensation during chon-drule formation, Meteoritics 28, 367-368, 1993.

[40] S. Matsunami, K. Ninagawa, S. Nishimura, N. Kubono. I.Yamamoto, M. Kohata, T. Wada, Y. Yamashita, J. Lu,D.W.G. Sears and H. Nishimura, Thermoluminescenceand compositional zoning in the mesostasis of a Se-markona group Al chondrule and new insights into thechondrule-forming process, Geochim. Cosmochim. Acta57, 2102-2110, 1992.

[41] S.W. Kieffer, Droplet chondrules, Science 189. 333-340,1975.

[42] J.L. Gooding and K. Keil, Relative abundances of chon-drule primary textural types in ordinary chondrites andtheir bearing on conditions of chondrule formation, Me-teoritics 16, 17-43, 1981.

[43] R.T. Dodd, Compositions of droplet chondrules in theManych (L31 chondrite and the origin of chondrules,Earth Planet. Sci. Lett. 40, 71-82. 1978.

[44] R.T. Dodd, Meteorites: A Petrologic Chemical Synthesis,Cambridge University Press, 1981.

[45] J.L. Gooding, Petrogenetic properties of chondrules inunequilibrated H-, L-, and LL-group chondritic mete-orites. Ph.D Thesis, Univ. New Mexico, Albuqerque.1979.

[46] J.V. Heyse, The metamorphic history of LL-group ordi-nary chondrites, Earth Planet. Sci. Lett. 40. 365-381,1978.