petrology and origin of ferromagnesian silicate … et al.: petrology and origin of ferromagnesian...

Lauretta et al.: Petrology and Origin of Ferromagnesian Silicate Chondrules 431

431

Petrology and Origin of FerromagnesianSilicate Chondrules

Dante S. LaurettaUniversity of Arizona

Hiroko NagaharaUniversity of Tokyo

Conel M. O’D. AlexanderCarnegie Institution of Washington

Ferromagnesian silicate chondrules are major components of most primitive meteorites. Theshapes, textures, and mineral compositions of these chondrules are consistent with crystalliza-tion of a molten droplet that was floating freely in space in the presence of a gas. The textureand mineralogy of a chondrule reflects the nature and composition of its precursor material aswell as its thermal history. There is an enduring debate about the degree to which chondrulesinteracted with the ambient gas during formation. In particular, it is uncertain whether or notchondrules experienced evaporation during heating and recondensation during cooling. Theextent to which these processes took place in chondrule melts varied as a function of the du-ration of heating as well as the environmental conditions such as pressure, temperature, com-position, and size. Thus, locked in chondrule bulk compositions, mineralogy, and textures areclues to the ambient conditions of the solar nebula and the nature of material processing in theinner solar system at the early stages of planet formation. Here we survey what is known aboutthe properties of ferromagnesian chondrules in primitive meteorites and use this informationto place constraints on these important parameters. The topics covered in this chapter are listedin Table 1.

1. INTRODUCTION

The most primitive meteorites, chondrites, have texturesthat suggest formation by accumulation of material that wasfreely floating in the solar nebula (Fig. 1). In a sense, theycan be viewed as sedimentary rocks that swept up materialin specific regions of the early solar system. While part oftheir parent asteroids, chondrites experienced varying de-grees of aqueous alteration, thermal metamorphism, andimpact shock. Mineral compositions within and among theleast-altered chondrites are highly variable. This wide chem-ical variation indicates that the different components havenot been in chemical communication long enough to homog-

TABLE 1. Contents of this chapter.

Section Topic

1 Introduction2 Properties of Ferromagnesian Chondrules3 Precursor Chemistry and Mineralogy4 Thermal Histories and Crystallization of Chondrules5 Metallogenesis and Redox Reactions in the Melt6 Elemental Evaporation and Recondensation7 Volatile Recondensation and Formation of

Chondrule Rims8 Constraints on Models of Chondrule Formation

Fig. 1. Plane-polarized transmitted light image of the Semarkona(LL3.0) primitive unequilibrated ordinary chondrite (from Lau-retta and Killgore, 2005). Chondrules make up ~70 vol% of thismeteorite. The remainder is composed of dark, opaque, fine-grained matrix. FOV = 1.5 cm.

432 Meteorites and the Early Solar System II

enize. These meteorites, referred to as unequilibrated chon-drites, preserve a number of chemical, mineralogical, andstructural signatures inherited from the protoplanetary disk.

The major components (50–80 vol%) of most chondritesare ~1-mm-diameter igneous spheres, called chondrules.Representative bulk compositions of natural chondrulesare summarized in Table 2, and illustrate the range in theirbulk chemistry. In this chapter, we focus on the most com-mon types of chondrules, which are primarily composedof the ferromagnesian minerals olivine [(Mg,Fe)2SiO4],low-Ca pyroxene [(Mg,Fe)SiO3], and high-Ca pyroxene[Ca(Mg,Fe)Si2O6]. These objects also contain a silicate-glass mesostasis rich in Na, Al, Si, Ca, and K, relative tochondrite bulk composition. Iron-based alloys, sulfides, andoxides are present in varying abundances.

In unequilibrated chondrites, chondrules are often sur-rounded by rims of fine-grained material. Chondrules andrims are themselves cemented together by a fine-grainedmaterial mineral matrix. Like the chondrules, the rims andmatrix are dominated by silicates, metals, and sulfides. How-ever, they are compositionally and mineralogically distinctfrom chondrules.

2. PROPERTIES OF FERROMAGNESIANCHONDRULES

2.1. Chondrule Classification

Gooding and Keil (1981) performed a comprehensivepetrographic survey of chondrules in unequilibrated ordi-nary chondrites. They classified chondrules based on theirtextures (see Fig. 2 and Table 3). Porphyritic chondrulescontain abundant relatively large (up to one-half the chon-

drule diameter), fairly uniformly sized crystals set in a fine-grained or glassy mesostasis. Porphyritic chondrules that aredominated by olivine (>10/1 volume ratio) are referred toas porphyritic olivine (PO) chondrules, while those that con-tain abundant pyroxene (>10/1 volume ratio) are porphy-ritic pyroxene chondrules (PP). Chondrules with silicatemineral abundances between these two extremes are calledporphyritic olivine-pyroxene (POP) chondrules. Granular(G) chondrules contain many small grains of uniform size(<10 µm) with poorly defined crystal outlines. Granularchondrules are also subdivided into GO, GOP, and GPgroups. Barred-olivine chondrules (BO) contain olivine crys-tals that have one dimension that is markedly longer thanthe other two when viewed in thin section. Olivine grains inbarred chondrules either all have the same crystallographicorientation or contain several sets of such oriented grains.Radial-pyroxene chondrules (RP) consist of fan-like arraysof low-Ca pyroxene, which radiate from one or more pointsnear their surface. Cryptocrystalline chondrules are objectsthat do not exhibit any recognizable crystal structure un-der an optical microscope. They have a high abundance ofglassy material and may contain submicrometer dendrites.Metallic chondrules consist of Fe,Ni metal, some troilite(FeS), and minor amounts of accessory phases such asschreibersite and metallic Cu (La Blue and Lauretta, 2004).Neither chondrule size nor shape is strongly correlated withtextural type. Since the study of Gooding and Keil (1981),two other chondrule classification schemes have been de-veloped. The most commonly used scheme divides chon-drules into two broad categories: type-I and type-II, basedon bulk FeO contents (McSween, 1977; Scott and Taylor,1983). Type-I chondrules are characterized by the presenceof FeO-poor olivine and pyroxene (Fo and En > 90). Type-IIchondrules contain FeO-rich olivine and pyroxene (Fo andEn < 90). The textural characteristics of both the type-I andtype-II series are gradational. These two main categories arefurther subdivided into subtypes A and B based on theirabundances of olivine and pyroxene. Type-IA and IIA chon-drules contain abundant olivine (>80 vol%), type-IB and IIBchondrules contain abundant pyroxene (>80 vol%), and chon-drules with intermediate abundances of olivine and pyrox-ene are classified as type-IAB or IIAB. This approach alongwith the Gooding and Keil (1981) classification scheme areoften used in conjunction with one another.

In addition to the textural and chemical classificationschemes presented above, another, independent chondruleclassification system has been proposed by Sears et al.(1992) and DeHart et al. (1992). This system does not relyon textural properties or modal mineralogy. Instead, chon-drules are classified based on their cathodoluminescenceproperties and mineral compositions. Cathodoluminescencerefers to the emission of visible light from a solid that isexcited by interaction with an electron beam. Cathodolu-minescence is highly dependent on trace-element abun-dances and crystal lattice defects. Chondrules that luminescebrightly are classified as group-A chondrules, while thosethat show little or no luminescence are group-B chondrules.

TABLE 2. Representative bulk compositionsfor major chondrule types (in wt%).

IA* IAB† IB† IIA‡ IIAB§ IIB§

SiO2 44.8 56.4 57.9 45.1 53.8 58.7TiO2 0.19 0.11 0.11 0.1 0.16 0.18Al2O3 3.9 2.4 2.4 2.68 4.8 4.6Cr2O3 0.44 0.58 0.59 0.51 0.5 0.59FeO 1.18 3.4 2.8 14.9 8.8 7.8MnO 0.12 0.36 0.28 0.39 0.6 0.54MgO 40.7 33.9 33.4 31.3 25.1 21.4CaO 3.5 1.7 1.7 1.89 2.7 3.3Na2O 0.52 0.56 0.46 1.65 2.1 1.9K2O 0.07 0.06 0.05 0.17 0.28 0.29P2O5 — — — 0.34 — —

FeS 0.2 0.3 0.09 0.94 0.28 0.13FeNi 3.76 0.32 0.28 0.51 — 0.08

*Jones and Scott (1989). † Jones (1994). ‡ Jones (1990). § Jones (1996).


These two groups are further subdivided based on the meanFeO and CaO concentration in olivine and the compositionof the chondrule mesostasis. Sears et al. (1992) identifiedfour groups of chondrules as primitive and suggested that

the other groups represent alteration of primitive chondruleson meteorite parent asteroids. Cathodoluminescence is a lesswidely available technique, and as a result the Sears et al.classification is not commonly used.

Fig. 2. Optical microscopy images illustrating the variety of chondrule textures. FOV = 1.35 mm in all cases except (h) (FOV =2.7 mm). (a) PO chondrule from QUE 97008 (L). (b) Reflected image of the same chondrule shown in (a). Note the rounded dropletsof metal concentrated near the chondrule boundary. (c) Another PO chondrule from Clovis (H) illustrating a larger grain size distribu-tion compared to the chondrule in (a). (d) A PO-RP chondrule pair from EET 90066 (L). (e) A PP chondrule from ALH 78119 (largechondrule). Note also the small cryptocrystalline chondrule in the upper left. (f) A POP chondrule from Bishunpur (LL3.1). Note theolivine crystals concentrated near the center and the pyroxene toward the outer edge.

(a) (b)

(c) (d)

(e) (f)


2.2. Chondrule Petrology

The most detailed studies of chondrule petrology havebeen conducted on Semarkona (LL3.0), one of the least-metamorphosed ordinary chondrites. The results of thiswork are summarized below. However, it is important to

remember that the sizes and relative abundances of chon-drules vary significantly among chondrite classes.

Type-IA porphyritic chondrules from the Semarkona(LL3.0) chondrite have been described in detail (Jones andScott, 1989). These objects contain metallic Fe,Ni and abun-dant small olivine crystals in a glassy to microcrystalline

Fig. 2. (continued). (g) A coarse-grained porphyritic pyroxene chondrule from EET 90066 (L). (h) A barred-olivine chondrule fromBishunpur (LL3.1). (i) A barred-olivine chondrule from Saratov (L/LL4) showing multiple barred units. Note the presence of some por-phyritic olivine crystals. (j) A radiating pyroxene chondrule from ALH 78119 (L). (k) A granular-olivine chondrule from QUE 97008 (L).(l) A pair of crypto-crystalline chondrules from ALH 78119 (L).

(g) (h)

(i) (j)

(k) ( l )


mesostasis. Olivine grain sizes are typically 15–40 µm. Thissmall grain size is a diagnostic property of type-IA porphy-ritic chondrules. The olivine grains are typically euhedralwith well-formed crystal faces in contact with the meso-stasis. Some olivines contain inclusions of mesostasis. It iscommon for the low-Ca pyroxene in these chondrules to beconcentrated near the chondrule margins and poikiliticallyenclose olivine grains. Rims of Ca-rich pyroxene occur onthe low-Ca pyroxene grains. Rounded grains of Fe,Ni metaland associated minerals are evenly distributed throughoutthe chondrules or concentrated near their margins.

In type-IAB porphyritic chondrules in Semarkona, euhe-dral equant olivine crystals are up to 150 µm in diameter(Jones, 1994). Olivine (averaging 25 µm across) also occursin large poikilitic low-Ca pyroxene crystals. Crystals of low-Ca pyroxene are twinned, generally tabular, and containlamellar, compositional zoning. Calcium-rich pyroxene oc-curs as thin (<20 µm), discontinuous overgrowth rims onlow-Ca pyroxene. Contacts between low-Ca and high-Capyroxene are sharp. Chemical concentration gradients arecommon in these Ca-rich pyroxene rims from the inner edgein contact with the low-Ca pyroxene to the outer edge nextto the mesostasis. Grains of Fe,Ni metal are also common.Mesostasis is glassy and, in some cases, contains needle-like crystallites of Ca-rich pyroxene.

Type-IB porphyritic chondrules in Semarkona are FeO-poor. They are distinguished from type-IAB chondrulesby their high proportion of pyroxene and low abundance(<20%) of olivine (Jones, 1994). However, olivine is presentin even the most pyroxene-rich type-IB chondrules andcommonly occurs as relict grains (see section 3.2.2 below).

Type-IIA porphyritic chondrules in Semarkona have anoverall rounded morphology but are not, in general, per-

fectly spherical (Jones, 1990). They contain euhedral oli-vine (10–250 µm in diameter averaging 50 µm). Olivinegrains have average compositions of Fo85. They are chemi-cally zoned with FeO increasing toward grain edges. Troi-lite occurs as small (<10 µm) rounded blebs distributedthroughout the phenocrysts and mesostasis. Small (<1 µm)euhedral grains of chromite (FeCr2O4) also occur in themesostasis. Low-calcium pyroxene and Fe,Ni metal are rare.

In Semarkona, type-IIB chondrules pyroxene morpholo-gies vary from euhedral crystals to sets of subparallel bars(Jones, 1990). The pyroxene cores consist of enstatite orclinoenstatite and are overgrown with narrow rims of Ca-rich pyroxene, typically successive overgrowth of augite onpigeonite. Augite occurs as prisms on the ends of clino-enstatite laths. Mesostasis consists of glass with abundantmicrocrystallites of augite. Small rounded grains of metaland sulfide are present.

3. PRECURSOR CHEMISTRYAND MINERALOGY

In order to understand the chemical evolution of chon-drule melts, it is necessary to place some constraints on thecomposition and mineralogy of their precursor material.Based on our current models of the early solar system, thereare several possible sources of this precursor material, in-cluding presolar material, nebular condensates, fragments ofearlier generations of chondrules, and fragments from plan-etesimals.

Presolar material is primordial dust inherited from theprotosolar molecular cloud, some of which still survives inthe chondrule rims and matrices of chondrites, includingcircumstellar grains and interstellar organic matter (Huss et

TABLE 3. Chondrule classification systems.

Textural Classification

Abbreviation Abundance in OCs (%) Texture

POP 47–52 Porphyritic, both olivine and pyroxenePO 15–27 Porphyritic, dominated by olivine (>10/1 by volume)PP 9–11 Porphyritic, abundant pyroxene (>10/1 by volume)RP 7–9 Radial pyroxeneBO 3–4 Barred olivineCC 3–5 Cryptocrystalline

GOP 2–5 Granular olivine-pyroxeneM <1 Metallic

Compositional Classification

SilicateType Composition Subtype Silicate Abundances

I FeO-poor A Abundant olivine (>80 vol%)(Fo, En > 90)

AB Intermediate abundances of olivine and pyroxeneII FeO-rich

(Fo, En < 90) B Abundant pyroxene (>80 vol%)


al., 1981). A significant proportion of material in the innersolar system may have been vaporized in the high-tempera-ture environment near the proto-Sun. As the nebula cooled,the most refractory material condensed as solid grains.These nebular condensates may have then been reheatedduring the event (or events) responsible for chondrule for-mation. Some models suggest that the vapor pressure of themajor rock-forming elements was high enough in the innersolar system for chondrule melts to condense directly fromthe gas phase. If there were multiple heating events, thenchondrules could have been fragmented and remelted manytimes, becoming their own precursors. Finally, it is possiblethat chondrules formed after larger planetesimals underwentchemical differentiation. In this case, chondrule precursorscould be material that was ejected by collisions betweenplanetesimals or during volcanic eruptions on asteroid sur-faces. In order to constrain the various models for chondruleprecursors, we look to their mineralogy, bulk chemistry, andisotopic compositions, as well as to the information pre-served in the relict grains that occur in many chondrules.

3.1. Constraints on Chondrule Precursorsfrom Bulk Compositions

Studies of chondrule bulk compositions have been used toinfer the nature of chondrule precursors. In a series of de-tailed investigations, the bulk compositions, accompanied bypetrographic control, were determined for ordinary (Gross-man and Wasson, 1982, 1983a; Swindle et al., 1991a,b),enstatite (Grossman et al., 1985), and carbonaceous chon-drite (Rubin and Wasson, 1987, 1988) chondrules. Detailedreviews of these studies are given in Grossman and Wasson(1983b), Grossman et al. (1988), and Grossman (1996).

Elemental correlations are present in chondrules thatappear to be related to their volatility and chemical affin-ity, although there are some variations between chondritegroups. The three most universal correlations occur withinthe refractory lithophiles (Al, Ca, Ti, Sc, and REEs), com-mon siderophiles and some chalcophiles (Fe, Co, Ni, Se),and the alkalis (Na and K). The behavior of elements likeMg, Mn, and Cr are more complex, and this is reflected inhow the compositions of olivine and pyroxene correlate withthe different elemental groups. For instance, the refractorylithophiles correlate with the presence of FeO-poor olivinein the ordinary and enstatite chondrites, with FeO-poor py-roxene in the CM and CO chondrites, and with neither in theCV chondrites. As would be expected, the siderophile andchalcophile elements are generally associated with metaland troilite, although the behavior of Au, As, and Ga variesbetween the chondrite groups and the refractory siderophilesIr, Os, and Ru appear to have affinities with the commonsiderophiles and refractory lithophiles.

These elemental correlations are thought to reflect vari-able abundances of different nebular condensates in theirprecursors. For example, the refractory component in ordi-nary chondrite chondrules has been interpreted as being

olivine-rich condensates formed above the condensationtemperature of enstatite, while the nonrefractory component(FeO-rich pyroxene, moderately volatile lithophiles, etc.)may have formed from fine-grained materials that were ableto equilibrate at lower nebular temperatures (Grossman andWasson, 1983b). Type-IIA chondrules from Semarkona(LL3.0) have (Na + K)/Al atomic ratios that approach butdo not exceed unity, which has been taken as evidence thatthe principle alkali-bearing precursor was albite (Hewins,1991). The fact that chondrules from different chondritegroups exhibit different element-mineral correlations wouldrequire that the condensate precursors for each chondritegroup formed under somewhat different conditions.

There are two potential problems with this interpretation.By the time olivine condenses there are several predictedcondensate minerals that would be the hosts for the refrac-tory elements: corundum (Al2O3), hibonite (CaAl12O19), spi-nel (MgAl2O4), perovskite (CaTiO3), melilite [Ca2(Mg,Al)(Si,Al)2O7], anorthite (CaAl2Si2O8), and diopside (CaMgSi2O6).The correlation between the refractory lithophiles would re-quire little fractionation between these minerals, despite thefact that some fractionation between condensates is requiredto produce the less-refractory lithophile component(s). Per-haps more importantly, none of these minerals have beenfound as relict grains in ordinary chondrite chondrules (seesection 3.2.2). Nor have they been found in significantamounts in the rims or matrix of the least-altered chondrites.Yet rims and matrix do preserve primitive materials, includ-ing presolar materials.

An alternative explanation for the interelement correla-tions is that they were produced by some combination ofevaporation, metal/sulfide loss, FeO reduction, and/or ran-dom sampling of recycled chondrule material (Sears et al.,1996; Alexander, 1994, 1996). Starting with a homogenouspopulation of precursors, variable degrees of evaporation,FeO reduction, and metal/sulfide loss would automaticallyinduce a correlation between refractory lithophile elementsby enriching them in the residual chondrules. As discussedin section 3.2.2, recycled chondrule material was an impor-tant component of chondrule precursors. All the refractorylithophiles behave incompatibly in chondrule melts, so theyare concentrated in the glass. Thus, a correlation between therefractory lithophiles could be produced by variable amountsof glass from an earlier generation of chondrules in the pre-cursors. Metal/silicate fractionation during chondrule for-mation would create siderophile-element correlations. Theupper limit of 1 for (Na + K)/Al ratios in type-II Semarkonachondrules may simply reflect the fact that for the alkalisto be easily accommodated in a silicate melt or glass re-quires the coupled substitution (Na,K)AlO2 — SiO2 (Alex-ander, 1994).

Ultimately, the relative importance of precursors andchondrule formation conditions depends on the degree towhich individual chondrules were open or closed chemicalsystems during formation (see section 6). However, bulk-chondrule refractory-element fractionations are unlikely to


be produced during evaporation and recondensation asso-ciated with ferromagnesian chondrule formation. Therefore,if they are present they must reflect variations in precursorcompositions.

The rare earth elements (REEs) and other refractorylithophiles are very useful probes of geochemical and cos-mochemical processes. The REEs can be strongly fraction-ated from one another during planetary differentiation,during high-temperature condensation, and during extremeevaporation. Most chondrules from ordinary (Grossman andWasson, 1983a; Kurat et al., 1984; Alexander, 1994), ensta-tite (Grossman and Wasson, 1985), and CH (Krot et al.,2001a) chondrites have nearly flat CI-normalized REE pat-terns, although some have small anomalies in Eu and pos-sibly Yb. There is no evidence in the REE patterns of thesechondrules that their precursors had experienced planetarydifferentiation, evaporation, or condensation.

Some chondrules from CV and CO carbonaceous chon-drites exhibit REE patterns that resemble patterns in someCAIs (Misawa and Nakamura, 1988a,b, 1996). The abun-dance of CAIs in CVs and COs makes the inclusion of CAImaterial in some chondrule precursors very likely, unlessCAIs and chondrules were mixed together after chondruleformation. The preservation of a few CAIs inside chon-drules (Itoh and Yurimoto, 2003) suggests that they weremixed together before chondrule formation. The reason thatmore CAIs are not found inside chondrules is probablybecause their liquidus temperatures are lower than those ofmost chondrules.

There are very few CAIs in ordinary or enstatite chon-drites (Guan et al., 2000; Fagan et al., 2001). Thus, it isperhaps not surprising that most chondrules in these mete-orites have unfractionated REE patterns. However, Pack etal. (2004) have found two chondrules in two ordinary chon-drites with large negative anomalies in Sm, Eu, and Yb. Thepattern of anomalies are consistent with predictions of con-densation under very reducing conditions, but have not beenobserved even in CAIs from the highly reduced enstatitechondrites. The two chondrules themselves are not veryreduced, but reduced condensates could have been minorcomponents of their precursors. If this is the case, it is sur-prising that similar patterns have not been found in enstatitechondrite chondrules.

3.2. Mineralogical Constraints onChondrule Precursors

3.2.1. Primitive chondrules. Chondrules exhibit a widevariety of textures and grain sizes. Since many chondruleswere heated enough to melt, some chondrules must haveexperienced temperatures between ambient and melting.The finest-grained chondrules are considered to be the leastaltered and therefore best preserve precursor mineralogyand bulk compositions. Foremost among such objects areagglomeratic olivine chondrules, also referred to in the lit-erature as “dark zoned chondrules” (Dodd, 1971; Hewins,

1997) and “fine-grained lumps” (Rubin, 1984). A compre-hensive description of these objects is given by Weisbergand Prinz (1996). They are comparable in size to true drop-let chondrules. They are typically composed of a coarser-grained core surrounded by material primarily comprised offine-grained (<5 µm) olivine, which constitutes 57–94 vol%.Larger grains of olivine, up to 400 µm, can also be present.They also contain lesser amounts of pyroxene and felds-pathic glass (<20 vol%). Minor components include chrom-ite, metal, and sulfide (<5 vol%). Some agglomeratic olivinechondrules contain fragments of prior generations of chon-drules and refractory inclusions.

Olivine compositions in a single agglomeratic chondrulespan a wide range of FeO contents, indicating formationin diverse environments with little subsequent equilibration.Many of these objects are layered and exhibit distinct tex-tural, mineralogical, compositional, and O-isotopic differ-ences between core and rim. Their textures are consistentwith a thermal history in which they were briefly (minutesto hours) heated up to temperatures high enough to resultin sintering, but not enough to produce significant melting.Based on the characteristics of agglomeratic chondrules, wecan infer that chondrule precursor materials were dominatedby olivine crystals 2–5 µm in size with varying amounts ofpyroxene, metal, troilite, chromite, and alkali feldspar orfeldspathic glass.

Metal is a common constituent of type-I and agglomer-atic chondrules and was likely a component of chondruleprecursors. Additional metal may have formed as a resultof in situ reduction either by carbonaceous material in thechondrule precursors or reaction with the nebula gas. Car-bonaceous material occurs in chondrule interiors as graphiteinclusions in metal (Mostefaoui and Perron, 1994), as fine-grained carbonaceous material associated with metal andsulfide (Lauretta and Buseck, 2003), and as poorly graphi-tized C included in chondrule silicates (Fries and Steele,2005).

In many chondrules sulfide occurs as droplets and aseutectic mixes when associated with metal. These texturesare consistent with crystallization of a sulfide melt. In thefinest-grained, least-heated chondrules, a troilite-pentland-ite assemblage occurs associated with minor amounts ofkamacite. The co-occurrence of these two sulfide phases isconsistent with experimental studies of kamacite sulfuriza-tion under solar nebula conditions (Lauretta et al., 1997,1998) and suggests that sulfides were an important compo-nent of some chondrule precursors (Yu et al., 1996).

Some researchers have suggested that chondrules couldhave been produced by volcanism on planetesimals (Hutchi-son and Graham, 1975; Kennedy et al., 1992) or by im-pacts between partially molten planetesimals (Hutchison etal., 1988). Chondrules formed in this way would containdistinct geochemical or isotopic signatures associated withthe partial melting and differentiation of their planetary bod-ies. Igneous fragments occur in primitive ordinary chon-drites that contain just such signatures (Graham et al., 1976;


Hutcheon and Hutchison, 1989; Hutchison and Graham,1975; Hutchison et al., 1988; Kennedy et al., 1992). How-ever, such igneous fragments are relatively rare and do notappear to have been the primary source of chondrule pre-cursors, nor do chondrules show the major- or trace-elementfractionations that are characteristic of partial melting anddifferentiation.

3.2.2. Relict grains in chondrules. Some olivine andorthopyroxene grains have compositions, particularly FeOcontents, that are distinctly different from those of the othergrains in their host chondrules (Nagahara, 1981; Rambaldi,1981). These so-called relict grains often display composi-tional zoning that is inconsistent with in situ crystallizationfrom their host chondrule melt (Rambaldi, 1981). They ap-pear to be precursor material that survived chondrule forma-tion without melting or dissolving. Therefore, they providea record of the compositions and mineralogies of chondruleprecursors, although this record is biased toward those grainswith higher melting temperatures (and therefore lower dis-solution rates) than the peak temperatures experienced bythe chondrules. Thus, it is likely that there were many morephases in the precursors that did not survive.

Rambaldi (1981) estimated that approximately half of allporphyritic chondrules in unequilibrated ordinary chondritescontain relict grains, while Jones (1996) estimated that 15%of all chondrules contain relict grains. Jones (1996) foundforsteritic relict grains in 3 out of 11 type-II Semarkona(LL3.0) chondrules and in 5 out of 6 such chondrules inALHA 77307 (CO3.0). However, it is important to keep inmind that there is no a priori reason to assume that all relictgrains will have anomalous compositions. Such grains aresimply the easiest to identify. It is likely that there are manymore relict grains that are not readily identified becausetheir compositions are not disparate.

Dusty relict olivines, which contain numerous submicro-meter to 10-µm inclusions of Fe-metal, are probably alsorelict and are observed in ~10% of all chondrules in ordinarychondrites (Nagahara, 1983; Grossman et al., 1988). Iron-metal grains in dusty olivines are severely depleted in Ni,relative to CI abundances and likely formed when the chon-drule melted in the presence of a reducing agent, such asH2 in the surrounding gas or solid carbonaceous material(Rambaldi, 1981). Dusty olivine grains are less commonin CM and CV chondrites (Jones, 1996). Olivine crystalsthat are completely enclosed in pyroxene grains in porphy-ritic chondrules may also be relict grains (Nagahara, 1983).However, Lofgren and Russell (1986) showed that similartextures can be produced during chondrule crystallization.Nearly pure forsterite exhibiting blue cathodoluminescenceoccurs inside chondrules of carbonaceous and ordinarychondrites and is thought to represent another populationof relict grains (Steele, 1986).

Misawa and Fujita (1994) report the occurrence of anintact CAI fragment in a ferromagnesian chondrule fromAllende, which confirms that some CAIs contributed tochondrule precursor material. Furthermore, Itoh and Yuri-moto (2003) discovered a ferromagnesian chondrule frag-

ment embedded in a CAI from the Yamato 81020 (CO3.0)chondrite, providing evidence that either chondrules con-tributed some material to CAI precursors, or that remelt-ing of CAIs took place during or after chondrule formation.Initial 26Al/27Al ratios in Allende chondrules and CAIs alsosupport the idea that chondrule formation began contem-poraneously with the formation of CAIs (Bizzaro et al.,2004).

Based on major- and minor-element compositions, mostrelict grains formed in earlier generations of chondrules(Jones, 1996; Nagahara, 1983); relict forsterite and enstatitegrains found in FeO-rich chondrules were derived from FeO-poor chondrules, and dusty olivine grains found in FeO-poor chondrules formed in FeO-rich chondrules (Jones,1996). Oxygen-isotopic data also support the view that thelow-FeO relict grains formed in a previous generation oflow-FeO porphyritic chondrules that were subsequentlyfragmented (Kunihiro et al., 2004). Relict grains, chondrulefragments, and isolated olivine and pyroxene grains in ma-trix all show that fragmentation and recycling of chondrulematerial was an important process in the chondrule-formingregion (Alexander et al., 1989a; Jones, 1992). Furthermore,based on the abundances of nonspherical chondrules (arbi-trarily defined as having aspect ratios >1.20) in CO3.0 chon-drites, Rubin and Wasson (2005) suggest that chondrulerecycling was the rule in the CO chondrule-formation re-gion and that most melting events produced only low de-grees of melting. They also suggest that the rarity of sig-nificantly nonspherical, multilobate chondrules in ordinarychondrites may reflect more-intense heating of chondruleprecursors in the region of the solar nebula in which theyformed.

Weinbruch et al. (2000) suggest that some refractory,relict forsterite grains cannot have formed by crystallizationfrom chondrule melts. They favor an origin by condensa-tion from an oxidized nebular gas. Rambaldi et al. (1983)noted that igneous olivine within Qingzhen (EH3) chon-drules always contain detectable amounts of CaO, whilerelict olivines are essentially CaO-free. They also suggestthat the relict olivines did not have an igneous origin andmight represent condensates from the solar nebula. Regard-less of their origin, the sporadic occurrence of large re-lict grains in such chondrules suggests a mix of occasionallarge grains among finer material in chondrule precursors(Hewins, 1997).

3.3. Experimental Constraints onChondrule Precursors

There have only been a few experiments that constrainthe nature of chondrule precursors. In a series of experi-ments, it was demonstrated that the grain size of chondruleprecursors has a significant effect on the resulting chondruletextures (Radomsky and Hewins, 1990; Connolly and Hew-ins, 1995, 1996; Connolly et al., 1998; Hewins and Fox,2004). These experiments suggest that granular chondruletextures (grain size <10 µm) result from the incomplete


melting of fine-grained starting materials, less than 20 µm indiameter. Furthermore, Hewins and Fox (2004) show that itis very difficult to generate porphyritic chondrules by melt-ing such fine-grained precursors. These textures require pre-cursor grain sizes >40 µm.

Experimental simulations of chondrule-like heatingevents using hydrous silicates as precursors always yieldhigh abundances of vesicles. The fact that vesicles are ex-ceedingly rare in chondrules is given as evidence that chon-drules were generated from anhydrous precursor minerals(Maharaj and Hewins, 1994, 1998). However, if chondrulemelts were heated to within 50°–100° of their liquidus tem-perature, vesicles could have migrated out of the spheres.Deloule et al. (1998) report high concentrations of water inchondrule silicates from the ordinary chondrites Semarkonaand Bishunpur, which they interpret as being due to thepresence of hydrous silicates in their precursors. On theother hand, both chondrites experienced parent-body aque-ous alteration (Alexander et al., 1989a), and Grossman et al.(2002) concluded that the zoning of water and halogens inthe Semarkona chondrules was due to this alteration.

3.4. Constraints on ChondrulePrecursors from Isotopic Studies

Chondrules in some carbonaceous chondrites show evi-dence of 16O-rich components. This signature is often in-terpreted as evidence for the survival or limited processingof a presolar component in the chondrule precursors (Wood,1981; Clayton et al., 1991; Yu et al., 1995). However, theabsence of correlated anomalies in any other element, ex-cept perhaps Cu (Luck et al., 2003), and the rarity of 16O-rich presolar grains (Zinner, 1998) strongly argue againstthe presolar precursor explanation. Heterogeneity of O-iso-topic ratios within these chondrules may be the result ofincorporation of relict grains from objects such as amoe-boid olivine aggregates, followed by solid-state chemicaldiffusion without concomitant O equilibration (Jones et al.,2004). More recently, it was argued that the mass-indepen-dent fractionations observed in chondrule O isotopes resultfrom chemical processes in the gas phase that accompaniedtheir formation (Thiemens, 1996; Clayton, 2004; Lyons andYoung, 2003; Jones et al, 2004). Even if the diversity ofO-isotopic compositions are due to non-mass-dependentisotopic fractionation, the relationships between the O-iso-topic compositions and different chondrite components arestill unknown.

Isotopic anomalies of Ti are present in chondrules fromAllende (CV3) (Niemeyer, 1988). There is no known wayof generating such anomalies during or after chondruleformation, requiring that the Ti-isotopic diversity in chon-drules is a feature that they inherited from their precursors.The chondrule with the largest measured Ti-isotopic anom-aly is also Al-rich, suggesting that these anomalies may re-flect the presence of CAIs in their precursors.

The isotopic composition of Sr was obtained for eightolivine chondrules from Allende (CV3) (Patchett, 1980).

The Allende chondrules are mass fractionated by up to1.4‰/amu in the lighter Sr isotopes. Patchett (1980) sug-gests that these chondrules derived their fractionated Srfrom a region of the nebula made isotopically light by par-tial kinetic mass separation of elements in the vapor phase.Later, the solid objects may have moved to a more isotopi-cally normal region, where the Allende matrix accreted.However, these signatures could also be due to condensa-tion or metamorphism. Allende chondrule mesostasis isalmost always recrystallized, chemically zoned (as are theolivines), and can contain halogens, sodalite, etc.

3.5. Direct Condensationfrom a Vapor Phase

Several researchers have suggested that chondrule meltscondensed directly from the vapor phase (e.g., Blander andKatz, 1967; Blander et al., 1976, 2001, 2004). Nelson etal. (1972) suggested that chondrule textures can result fromrapid crystallization of relatively slowly cooling moltendroplets and suggested that this history is consistent withcondensation from a nebula of solar composition and solidi-fication in an ambient medium at high temperature.

Ebel and Grossman (2000) performed detailed equilib-rium calculations of the sequence of condensation of theelements from gases with various enrichments in a dust ofCI-chondrite composition. They determined that the FeOcontents typical of type-IIA chondrules can only form athigh total pressures (>10–3 bar) and high dust enrichments(100–1000× solar) at temperatures high enough for dust-gas equilibration to occur (~900°C). Even under these con-ditions, the condensates contain much more CaO and Al2O3(relative to MgO and SiO2) than most ordinary-chondritechondrules. Furthermore, most chondrule compositions aretoo Na-rich compared to predictions of condensate bulkcompositions. However, the higher chondrule Na contentscould reflect fractionation prior to chondrule formation oralkali reintroduction at lower temperatures.

A condensation origin has been proposed for chondrulesin the Hammadah al Hamra 237 and QUE 94411 bencub-binite chondrites, as well as the recently recognized CHgroup of chondrites (Krot et al., 2001a,b). Chondrules inthese meteorites generally have barred-olivine and cryptoc-rystalline textures and are essentially metal-free. Zonedmetallic chondrules are also abundant in these meteorites.The presence of inclusions with chondrule-like textures inthe zoned metal grains suggests that the two componentsformed in the same nebular region, with silicate chondrulesforming at temperatures above those of metal condensation(1773°C).

Two formation scenarios for these chondrules have beenconsidered. Either these chondrules formed by direct gas-liquid condensation or by prolonged heating of chondruleprecursors above their liquidus temperatures, which de-stroyed all solid nuclei. The complementary behavior ofrefractory lithophile elements (Ca, Al, Ti, REEs) in barred-olivine and cryptocrystalline chondrules suggests that these


chondrules formed in a chemically closed system (Krot etal., 2001a). Furthermore, high Cr2O3 (0.4–0.7 wt%) andFeO (1–4 wt%) contents in the chondrules and complemen-tary depletions in the FeNi-metal grains suggests formationin a closed system under relatively oxidizing conditions.These compositional trends are consistent with formationin a dust-enhanced (10–50× solar) region of the solar nebula(Krot et al., 2001b; Petaev et al., 2001). The inferred en-hanced initial dust/gas ratios supports the theory that thechondrules in these meteorites formed by direct gas-liquidcondensation (Ebel and Grossman, 2000).

One of the primary arguments against direct condensa-tion is the extreme conditions required to produce the FeOcontents of type-II chondrules. However, Blander et al.(2001, 2004) suggest that the condensation of metallic Fewas suppressed because of nucleation constraints. The in-hibition of Fe condensation results in supersaturation of Fein the gas phase. As a result, higher levels of FeO enteredsilicate melts by condensing directly from the vapor. Thismodel also explains the general absence of metal in type-IIchondrules.

The strongest evidence that at least some chondrulesformed through melting of preexisting materials and did notcondense from a gas is the presence of large, composition-ally distinct relict grains (see section 3.2.2 above). Thesegrains clearly survived the chondrule formation processwithout melting. In addition, porphyritic chondrules, themost abundant type of chondrule, require the preservationof numerous crystal nuclei (see section 4.1), implying thattheir precursors were also at least partially crystalline.

4. THERMAL HISTORIES ANDCRYSTALLIZATION OF CHONDRULES

It has long been established that many of the featuresseen in chondrules are the result of crystallization from amelt. In the previous section, we describe the constraintsthat can be placed on the precursor material. Here we iden-tify the constraints that can be placed on the peak tempera-tures and duration of heating from experimental simulationsof chondrule formation, and considerations of melting ki-netics of relict grains. We also describe how chondrule tex-tures and disequilibrium features, such as compositionalzoning in minerals, can be used to constrain the cooling his-tories of chondrules.

4.1. Peak Temperatures andPreservation of Nucleation Sites

A prerequisite for the crystallization of a melt is thepresence of crystal nuclei, and nuclei abundance is a majorfactor in determining chondrule textures (Lofgren, 1996).On the timescales of chondrule formation, spontaneousformation of nuclei in the melt is a relatively slow process.As a result, nuclei-free chondrule melts can cool well belowtheir liquidus temperatures and become supercooled. Whena nucleus forms there is a burst of rapid crystal growth,producing textures similar to those of barred-olivine and

radial-pyroxene chondrules (Lofgren and Lanier, 1990;Radomsky and Hewins, 1990).

Porphyritic and granular textures form when abundantnuclei are present during cooling. The simplest interpreta-tion of this result is that chondrule peak temperatures wereslightly below their liquidus temperatures (Lofgren andRussell, 1986), which enabled the survival of nuclei. Theseunmelted relicts provided abundant nucleation sites thatgenerated the characteristic porphyritic textures on cooling(Lofgren, 1989; Hewins and Radomsky, 1990). Chondruleliquidus temperatures therefore provide important con-straints on the peak temperatures experienced by chondrulemelt droplets. Droplets that were completely molten (e.g.,those that crystallized as barred, excentroradial, and cryp-tocrystalline chondrules) have liquidus temperatures thatrange from 1200°C to over 1900°C and are highly depen-dent on bulk composition (Herzberg, 1979; Hewins andRadomsky, 1990). Given this wide range of temperatures, itis clear that bulk composition is at least as important as ini-tial temperature in controlling chondrule textures (Connollyand Hewins, 1996). Based on the abundances, compositions,and textures of incompletely melted vs. completely meltedchondrules in carbonaceous and ordinary chondrites, it hasbeen estimated that few chondrules with liquidus tempera-tures over 1750°C were completely melted, and few withliquidus temperatures under 1400°C were incompletelymelted. Therefore, it was concluded that most chondrulesreached peak temperatures of 1500°–1550°C (Radomskyand Hewins, 1990).

Further constraints of chondrule thermal histories havebeen obtained by experimental simulations that accuratelyreproduce chondrule grain sizes and textures (Hewins andRadomsky, 1990; Lofgren, 1996; Connolly et al., 1998; Tsu-chiyama et al., 2004). These studies show that it is impor-tant to understand fully not only the peak temperatures thatchondrules experienced, but also the duration of the heat-ing event. This dual consideration is important because, forkinetically driven processes, the effect of heating to a highpeak temperature for a short time can be equivalent to thatof heating to a lower temperature for longer times (Joneset al., 2000). The total amount of energy, as opposed to thepeak temperatures, is therefore constrained to be below thatrequired both to raise the precursor material to the liquidusand to supply the necessary latent heat of fusion to melt allsolid precursor material. Connolly et al. (1998) showed thata short heat pulse, with a peak temperature well above theliquidus, can also produce porphyritic chondrule textures.Nuclei survive because the timescales are shorter than themelting and dissolution kinetics. Thus, the maximum tem-perature achieved during chondrule formation could havebeen approximately 2100°C, 200°C higher than the liquidustemperatures of even the most refractory chondrules. Thedistinctive olivine rims on barred-olivine chondrules alsoprovide information about peak temperatures. The forma-tion of compact olivine rims on these chondrules requires aheating temperature about 100°C or more above the liquidusto promote heterogeneous nucleation of olivine on the sur-face of the melt droplet (Tsuchiyama et al., 2004).


It is also possible that chondrule formation occurred ina dusty environment, which provided abundant nucleationsites. Experimental simulations that seeded molten dropletswith small grains of dust have reproduced a wide varietyof chondrule textures (Connolly and Hewins, 1995). Tex-tures vary as a function of the temperature at which dust isintroduced to the melt, the number density of dust grains,and their mineralogy.

4.2. Kinetics of Melting andSurvival of Relict Grains

Another independent constraint on the duration of chon-drule melting is provided by the presence of relict grainsthat were not in contact with melt for a long enough timeto dissolve. Early studies used thermal diffusion calculationsto determine the relationship between the duration of thechondrule heating event and the maximum attainable tem-perature inside chondrules (Fujii and Miyamoto, 1983).These researchers determined that the duration of eachheating event had to be less than 0.01 s for chondrules withradii of 0.3 mm for the survival of relict grains.

More recently, Greenwood and Hess (1996) used a modelfor the kinetics of congruent melting to place constraintson the duration and maximum temperature experienced bythe interiors of relict-bearing chondrules. They argue thatrelict forsteritic olivine grains in chondrules were not heatedabove 1900°C, approximately the melting point of forsterite.Their temperature limit for chondrules with Mg-rich py-roxene relicts is 1557°C, the incongruent melting point ofenstatite. In both cases, the input of energy could not havelasted for more than a few seconds. At temperatures belowtheir melting points, relict grains will dissolve if they arenot in equilibrium with the melt. Dissolution is slower andquantitatively less well understood than melting, but ulti-mately relict grains may be used to better constrain the ther-mal histories of a wide range of chondrules.

4.3. Chondrule Cooling Rates

4.3.1. Chondrule textures. Chondrule textures are con-trolled by nucleation density, degree of undercooling, andcooling rate. The density of nucleation sites is a functionof the duration and peak temperature of the heating eventas well as the abundance of dust particles in the chondruleformation region (section 4.2). The abundance of nucleationsites in the melt also controls the degree of undercoolingthat can occur. These factors together with the subsequentcooling rate have direct effects on the development of bothcrystal morphology and the elemental distributions withinthe resulting chondrule grains.

Dynamic crystallization experiments reveal that forsteritecrystallizes with one of four distinct morphologies: euhe-dral (well-defined crystal faces), tabular (preferential devel-opment in two directions), hopper (hourglass shape withhexagonal outline), and dendritic (intricate branching struc-tures) (Faure et al., 2003). The morphology of forsterite crys-tals varies as a function of the cooling rate, degree of un-

dercooling, and abundance of crystallization nuclei. As thedegree of undercooling increases, forsterite morphologyevolves from tabular to hopper. At the most rapid coolingrates, dendritic crystals form.

Euhedral olivine crystals develop in silicate melts withabundant nuclei. No supercooling occurs and crystallizationbegins as soon as the temperature drops below the liquidus.The resulting textures are uniformly sized olivine phenoc-rysts in a glassy mesostasis (Lofgren, 1989; Radomsky andHewins, 1990). The cooling rates required to produce por-phyritic textures in experimental chondule melts are lessthan 1000°C/h with those on the order of 100°C/h best re-producing the textures in porphyritic chondrules (Lofgren,1989; Radomsky and Hewins, 1990).

Barred-olivine chondrules contain olivine crystals withsingle-plate dendritic texture. These textures are the resultof crystallization of nuclei-free melt droplets. A significantamount of undercooling may have occurred in these meltdroplets. Comparison of the olivine morphology in barredchondrules with analog textures produced in experimentsindicates that these chondrules cooled at rates of 500°–2300°C/h (Lofgren and Lanier, 1990; Tsuchiyama et al.,2004).

Pyroxene crystallization textures also vary as a functionof melt composition and cooling history. Porphyritic pyrox-ene textures are produced only in experiments that do notexceed the liquidus temperature or exceed it only for a briefperiod of time (Lofgren and Russell, 1986). These texturesform when cooling rates are well in excess of 100°C/h.However, at cooling rates greater than 2000°C/h, the crys-tal sizes are significantly below those seen in most chon-drules of this type.

Crystals in radial pyroxene chondrules are generallyhighly acicular, radial, and often in the form of spherulites.These morphologies are indicative of growth from a meltthat experienced very high degrees of undercooling. Nucle-ation sites, which are initially absent in the melt, form fromcrystal embryos that exceed some critical radius. Once nu-cleated, growth of the pyroxene crystals is very rapid. Thelarge degree of supercooling stabilizes the highly non-equi-librium crystal shapes. The radial or spherulitic textures candevelop even at quite slow cooling rates (5°C/h) in any meltwhere olivine is not a liquidus or near-liquidus phase (Lof-gren and Russell, 1986).

4.3.2. Zoning profiles. Experimental studies suggestthat chondrule textures largely reflect the abundance of nu-cleation sites in the melt, which directly influences the de-gree of undercooling that a droplet can experience. Evenchondrules that clearly crystallized rapidly, e.g., radial py-roxene chondrules, can form at relatively slow cooling rates.Thus, chondrule texture appears to be controlled predomi-nantly by the intensity and duration of the heating event.However, olivine crystals in many chondrules are hetero-geneous and are chemically zoned from core to rim. Thesezoning profiles indicate that the chondrules cooled tooquickly for the olivines to maintain equilibrium with theremaining melts and for diffusion to homogenize the oli-vines. Thus, detailed characterization of the chemical varia-


tion in chondrule olivines can provide additional informa-tion about the cooling rates that these objects experienced.

Olivine zonation is particularly prominent in type-IIAporphyritic olivine chondrules in the least-equilibrated or-dinary chondrites, such as Semarkona (Jones, 1990). Zona-tion occurs in FeO (9–27 wt%), Cr2O3 (0.2–0.6 wt%), MnO(0.2–0.7 wt%), CaO (0.1–0.4 wt%), and P2O5 (0–0.5 wt%),with abundances increasing from cores to rims. Both theolivine textures and the normal zoning profiles are consis-tent with an igneous origin. The zonation likely results fromfractional crystallization of the chondrule melts.

The zoning profiles in porphyritic olivine chondruleshave been compared to those in experimentally producedanalogs (Jones and Lofgren, 1993; Weinbruch et al., 1998).In Semarkona chondrules, the shapes of the zoning profiles,as well as the absolute values of FeO and minor-elementcontents, of the natural and synthetic olivine crystals aresimilar at cooling rates as low as 5°C/h. Olivine zoningis much more restricted at low cooling rates in bulk com-positions, which permit pyroxene nucleation and growth(Radomsky and Hewins, 1990).

4.3.3. Exsolution microstructures. Another constrainton cooling rate is exsolution microstructure (Carpenter,1979; Kitamura et al., 1983; Muller et al., 1995; Watanabeet al., 1985; Weinbruch and Muller, 1995; Weinbruch et al.,2001). Lamellar exsolution has been observed in some chon-drule pyroxenes. In particular, pigeonite/diopside exsolutionlamellae occurs in granular-olivine-pyroxene chondrulesfrom the Allende meteorite (Weinbruch and Muller, 1995).It is suggested that these textures are produced when coolingrates are on the order of 5°–25°C/h in the temperature rangeof 1200°–1350°C. These exsolution textures have also beenproduced in synthetic analogs of chondrules (Weinbruch etal., 2001) under cooling rates of 10°–50°C/h between 1000°and 1455°C. In these experiments, higher cooling rates (50°–450°C/h) yield only modulated structures that are unlikethose seen in chondrule pyroxenes.

4.4. Mineral Crystallization Sequence

The sequence in which minerals crystallize in chondrulemelts is highly dependent on the bulk composition as wellas the thermal history. In many chondrules, olivine is theearliest phase nucleating from the melt, forming between1600°C and 1400°C (Ferraris et al., 2002). This temperaturerange is consistent with temperatures derived from chon-drule chromite-olivine pairs, which have Fe-Mg partitioningcharacteristic of phases that crystallized from silicate meltsat temperatures >1400°C (Johnson and Prinz, 1991).

Liquid Fe-based alloys containing less than 5.2 atom%Ni will preciptitate metal crystals with a body-centeredcubic structure (δ-Fe) that have slightly lower Ni contents(<4.1 atom%) between 1536° and 1513°C. Nickel-rich met-als (>5.2 atom% Ni) will crystallize with the face-centeredcubic structure. Crystallization temperatures decrease withincreasing Ni content, reaching a minimum of 1425°C be-tween 60 and 70 atom% Ni.

In FeO-rich porphyritic pyroxene chondrules in Semar-kona it appears that the order of crystallization of olivineand pyroxene is variable (Jones, 1996). Intergrown pyrox-ene and olivine phenocrysts indicate that both minerals werecrystallizing simultaneously as the chondrule cooled. Inseveral chondrules olivine is dominant and appears to havepreceded growth of pyroxene. In the coarse-barred chon-drules, olivine is sometimes present as small rounded in-clusions, indicating that it crystallized before pyroxene.

Textural relationships between the various pyroxenephases in Semarkona chondrules indicate that the crystalli-zation sequence was one of progressive enrichment in wol-lastonite (Wo) content. In these objects, low-Ca pyroxenesare overgrown with pigeonite followed by augite, whichonly occurs as rims on low-Ca pyroxene phenocrysts (Jones,1996). Pigeonitic rims form at temperatures above 1000°C.By subsolidus exsolution, high-pigeonite inverts to twinnedlow-pigeonite between 935° and 845°C. Augite crystallizesas euhedral grains hosted within the glassy matrix, or asskeletal crystals, most probably just after the pigeonitic rimgrowth between 1150° and 1000°C. Pigeonitic and augiticrims form augitic and enstatitic sigmoids, respectively, byexsolution between 1000° and 640°C. Based upon glasscomposition, the overall solidification temperature of themesostasis occurs close to 990°C, the minimum melt tem-perature in the quartz-albite-orthoclase system (Hamiltonand Mackenzie, 1965).

5. METALLOGENESIS AND REDOXREACTIONS IN THE MELT

5.1. Metal in Chondrule Interiors

Metal-bearing type-I chondrules have highly reducedmineral assemblages, such as silicates with low FeO con-tents (Jones and Scott, 1989) and low-Ni metal grains thatcontain variable amounts of P, Cr, and Si (Lauretta andBuseck, 2003; Rambaldi and Wasson, 1981; Zanda et al.,1994). Despite much interest, the origin of the metal in thesechondrules is still the subject of debate. Some researcherspropose that metal grains are derived from metallic nebu-lar condensates that avoided oxidation even during chon-drule formation (Grossman and Wasson, 1985; Rambaldiand Wasson, 1981). Other studies suggest that the metalformed by in situ reduction of preexisting sulfides, oxides,and FeO-rich silicates in the melt (Connolly et al., 1994;Lauretta et al., 2001; Lauretta and Buseck, 2003). In thissection, we examine the evidence for the origin of metal insilicate chondrules.

5.2. Formation of Metal by Condensation

Several researchers have suggested that metal grains inchondrule interiors formed by melting of metal and sulfidenebular condensates. It is suggested that these grains main-tained their chemical identity during the melting process,presumably by surviving as immiscible droplets that did not


exchange with the surrounding silicate melt. Evidence forthis idea comes from the abundances of siderophile andchalcophile elements in Semarkona (LL3.0) chondrules(Grossman and Wasson, 1985), which are presumably hostedby the metal and sulfide. The abundances of these elementsin chondrules are highly variable. The nonrefractory sidero-phile abundance patterns in Ni-rich chondrules are smoothfunctions of volatility, which have been interpreted as nebu-lar signatures. However, the volatile siderophile fractiona-tions may also reflect evaporation during chondrule melting.

Metal from chondrule interiors in Renazzo, Al Rais, andthe related chondrite, MacAlpine Hills 87320 (Lee et al.,1992), shows generally high Ni and Co, flat Ni and Coprofiles, and moderately large grain-to-grain compositionalvariations (even within chondrules). Metal from chondrulemargins adjacent to matrix has convex (inverted U-shaped)Ni and Co zoning profiles with the highest Ni and Co con-centrations at grain centers. These profiles are similar tothose observed in the CH chondrites, which have been in-terpreted as resulting from condensation in the solar nebula(Meibom et al., 1999) or a postimpact vapor plume (Camp-bell et al., 2002). However, Lee et al. (1992) interpretedthese patterns to reflect dilution by Fe produced by FeOreduction, which occurred during a brief period of parent-body metamorphism.

5.3. Metallogenesis by Reduction inChondrule Melts

Fine (~2 µm), Ni-poor (~1 wt%) metal grains are com-mon inclusions in the olivine in porphyritic chondrules inunequilibrated ordinary chondrites (Rambaldi and Wasson,1981, 1982). The most common occurrence of this “dusty”metal is in the core of olivine grains having clear Fe-poorrims. The original compositions of dusty grains, which havebeen determined by assuming that all Fe metal was origi-nally present as FeO (Jones and Danielson, 1997), closelymatch those of olivines in chondrules from unequilibratedchondrites. These assemblages likely reflect high tempera-ture in situ reduction of FeO from olivine (Rambaldi andWasson, 1981; Jones and Danielson, 1997).

The evidence that dusty olivine grains formed by in situreduction suggests that similar processes may have affectedthe compositions of other chondrule metal phases. Chon-drule kamacite in Krymka and Chainpur has Ni concentra-tions (3–4 wt%) that are significantly below that expectedfor nebular condensates (~5 wt%). This kamacite may havebeen diluted by Fe reduced from the silicates during chon-drule formation by either H2 gas or carbonaceous precursormatter (Rambaldi and Wasson, 1984).

The elements Cr, Si, and P are common in metal grainsin the least-metamorphosed ordinary and carbonaceouschondrites and provide additional information on redox re-actions in type-I chondrules (Zanda et al., 1994). The con-centrations of these elements are inversely correlated withthe FeO concentrations of their host chondrule olivines. Thisrelation suggests that Cr, Si, and P concentrations could not

have been established by nebular condensation, but are theresult of metal-silicate equilibration within chondrules.

A link between chondrule-derived metal and opaqueassemblages in the matrix of primitive chondrites has alsobeen established, based on the chemistry of Cr, Si, and P(Lauretta et al., 2001; Lauretta and Buseck, 2003). It islikely that the metal melted in and equilibrated with a sili-cate liquid under high-temperature, reducing conditions. Inthis environment the molten alloys incorporated variedamounts of Si, Ni, P, Cr, and Co, depending on the oxygenfugacity and temperature of the melt. As the system cooled,some metal oxidation occurred in the chondrule interior,producing metal-associated phosphate, chromite, and silica.Metal that migrated to chondrule boundaries experiencedextensive corrosion as a result of exposure to an external,oxidizing atmosphere.

Trace-element data also support a chondrule-melt originfor metal grains in ordinary chondrites (Kong and Ebihara,1996, 1997). Characteristics retained in the bulk metals ofunequilibrated H, L, and LL chondrites (abundant C; highcontents of Cr, V, Mn and low contents of W, Mo, and Gacompared to metals of equilibrated chondrites; less enrich-ment of W than Mo; and fractionation of Co from Ni) dem-onstrate that chondrite metals are not nebular condensates.All these characteristics are consistent with melting. Fur-thermore, these data suggest that chondrite metals are notmelting remnants of previously condensed metals. Rather,they have been produced by reduction of CI- or CM-likematerial during the melting process.

Volatile-siderophile-element abundance patterns of me-tallic and nonmagnetic fractions of CR chondrites are com-plementary and volatility-dependent (Kong et al., 1999). InRenazzo, fine-grained metal is located inside chondrules orin chondrule rims. Matrix contains mostly coarse metalgrains. Both types of metal, fine and coarse, exhibit simi-lar chemical signatures, comparatively high in Cr and lowin Ni, suggesting a genetic relationship (Kong and Palme,1999) in which Renazzo chondrules and metal were formedby reduction of oxidized precursors compositionally simi-lar to CI chondrites. The differences in siderophile-elementpattern between chondrules, chondrule metal, and matrixindicate evaporation and recondensation of volatile elementsduring chondrule formation.

Connolly et al. (2001) performed a detailed study ofmetal in CR chondrites from chondrule interiors, chondrulerims, and in the matrix. They found that the refractory plati-num group elements Os, Ir, and Pt behave coherently inmetal. They are all uniformly enriched, depleted, or unfrac-tionated with respect to Fe. Metal with solar ratios of Os/Fe, Ir/Fe, and Pt/Fe occur primarily in chondrule interiors.All metal grains have essentially CI ratios of Ni/Fe and Co/Fe. Almost all metal grains have lower-than-CI ratios of thevolatile siderophile elements, Au and P. Furthermore, thebulk compositions of chondrules are generally unfraction-ated relative to CI chondrites for elements more refractorythan Cr, but are depleted in the more volatile elements. Theabundances of siderophile elements correlate strongly with


the metal abundance in the chondrules, which implies thatthe siderophile depletions are due to expulsion of metalfrom the chondrule melts. These data are interpreted to sug-gest that most metal formed during melting via reductionof oxides by C that was part of the chondrule precursor.

Nickel has a lower volatility than Fe, and Humayun etal. (2002) found that Ni/Fe ratios in metal in CR chondruleinteriors tend to be higher than those in the rims and ma-trix. In addition, Pd has nearly the same volatility as Fe,and Pd/Fe ratios are nearly the same in metal in chondrules,rims and matrix. Copper is more volatile than Fe, and Cu/Fe ratios are much lower in chondrule metal relative to thatin rims and matrix. These observations can be explained byevaporation of volatile siderophile elements, including Fe,and subsequent recondensation mostly onto chondrule rims.In very FeO-poor ordinary chondrite chondrules, there is acorrelated decrease in Fe, Mn, and Cr (Alexander, 1994).Both Fe and Cr are siderophile under reducing conditionsbut Mn is not, suggesting that these correlated depletionsare due to evaporation. Rims with large correlated enrich-ments of FeO, Si, Mn, and Cu have also been reported inordinary chondrites, and may result from recondensation ofevaporated material (Alexander, 1995).

5.4. Experimental Studies of ChondruleMelt Redox Reactions

Experimental studies have added to our understandingof the role of in situ reduction in producing type-I chon-drule textures and mineral compositions. Experiments inwhich C, in the form of graphite or diamond, is flash-heatedtogether with magnesian silicates reproduce many minera-logical features characteristic of type-I chondrules, includingMg-rich and dusty olivines, Cr and Si dissolved in metal,and silica inclusions inside metal (Connolly et al., 1994).These results provide strong evidence that C in chondrulemelts is an effective reducing agent and that chondrule meltswere internally buffered. Furthermore, they suggest that Cplayed a key role in chondrule formation by briefly creatinga reducing environment within chondrule melts that was de-coupled from the nebula gas.

Further experiments have provided more detail on thenature of the reduction reaction. Cohen and Hewins (2004)performed an experimental study of metallic Fe formationin chondrule-like melts. They showed that troilite and Fe-metal are extremely volatile under canonical chondrule-formation conditions. Conversely, at higher total pressures(P H2 = 1 atm) Fe metal is less volatile and remains in themelt for longer times. These experimental results suggestthat metal-bearing chondrules formed in environments withelevated pressures relative to the canonical nebula. Alter-natively, unlike in the experiments, the chondrule melt andthe vapor phase may have reached equilibrium in the solarnebula, in which case the metal may have stabilized.

Additional studies have investigated the transformationsof olivine single crystals in contact with graphite powder.At 1150°C precipitates of Fe,Ni and a thin amorphous layerof silica formed on the surface of the olivine crystals (Lem-

elle et al., 2001). A reaction layer composed of Fe-Ni pre-cipitates and FeO-depleted olivine develops between thesurface and an inner reaction front at 1350°C (Lemelle etal., 2000). The reaction layer contains two reaction fronts(inner and outer) corresponding to the beginning of metalprecipitation and to the total depletion of Fe2+ from oliv-ine. The propagation of the two reaction fronts follows aparabolic law and is limited by the interdiffusion of Fe/Niand Mg in the olivine lattice. In both natural and experi-mental samples, micrometer-sized metal blebs observed inthe dusty region often show preferential alignments alongcrystallographic directions of the olivine grains, have lowNi contents (typically <2 wt%), and are frequently sur-rounded by a silica-rich glass layer (Leroux et al., 2003).These features suggest that dusty olivines are formed by asubsolidus reduction of initially fayalitic olivines. Compari-son with experimentally produced dusty olivines suggeststhat timescales on the order of minutes usually inferred forchondrule formation are also adequate for the formation ofdusty olivines. These observations are consistent with thehypothesis that at least part of the metal phase in chondritesoriginated from reduction during chondrule formation.

There have also been recent experiments on the natureof chemical reduction in silicate melts in a reducing atmos-phere (Everman and Cooper, 2003). Melt samples wereheated to 1300°–1400°C under a flowing gas mixture of1-atm CO/CO2 ( fO2

= QIF – 2). The melt compositions areenriched in Al2O3 and depleted in MgO and CaO relativeto ferromagnesian chondrules. The melt experienced an in-ternal reaction in which a dispersion of nanometer-scaleFe-metal precipitates formed at an internal interface. The dis-tance between this interface and the melt surface increasedthroughout the experiment. Oxygen was lost from the meltat the free surface via conversion of CO to CO2 resulting inan increase in the cation-to-oxygen ratio and net reductionof the melt.

Thus, while it seems likely that nebular condensate metalwas incorporated into chondrule precursors, it is also likelythat this metal experienced significant interaction with thesilicate melt. In chondrule melts that incorporated a signifi-cant amount of carbonaceous material or that interactedwith a highly reduced vapor phase, additional metal wascreated by reduction of Fe and associated elements fromsilicate and other oxidized phases.

Subsequent petrologic work supports the model thatsome chondrules incorporated C-rich material in their pre-cursors (e.g., Lauretta et al., 2001; Lauretta and Buseck,2003; Fries and Steele, 2005). Small (<1–20 µm) C inclu-sions are common in the metal of type-3 ordinary chondrites(Lauretta et al., 2001; Mostefaoui et al., 2000) and fine-grained carbonaceous inclusions occur in some chondruleinteriors (Lauretta and Buseck, 2003; Scott et al., 1984). Itis likely that these inclusions are the remains of carbona-ceous matter that escaped oxidation during chondrule for-mation. There are claims of high C contents (several weightpercent) in chondrule olivines from the primitive chondritesAllende (CV3), Semarkona (LL3.0), and Bishunpur (LL3.1)(Hanon et al., 1998). These researchers report that mean


C concentrations are 3× higher in type-I chondrules thanin type-II chondrules. However, nuclear reaction analysesshow that chondrule olivines have C contents of less than120 ppm (Varela and Metrich, 2000).

6. ELEMENTAL EVAPORATIONAND RECONDENSATION

6.1. Chemical and Isotopic FractionationsResulting from Evaporation

The origin of the wide range of compositions of ferro-magnesian chondrules remains controversial. Figure 3 sum-marizes the variation and the average compositions of Se-markona (Jones and Scott, 1989; Jones, 1990, 1994, 1996;Huang et al., 1996; Tachibana et al., 2003) and Renazzo(Kong and Palme, 1999) chondrules. Type-I chondrules inparticular exhibit significant chemical fractionations, withdepletions of elements like Cr, Mn, Na, K, and Fe, and en-richments of elements such as Al, Ca, and Ti, relative to Mg.

These compositional variations suggest that elementalvolatility is the primary factor controlling chondrule bulkcompositions. The relative volatility of elements can be esti-mated in two ways: (1) from the calculated equilibrium tem-perature at which 50% of an element has condensed froma gas of solar composition typically at Ptot = 10–4 bar (e.g.,Lodders, 2003), and (2) from evaporation experiments ofchondritic melts (e.g., Hashimoto, 1983; Floss et al., 1996;Wang et al., 2001). Both approaches suggest that the orderof decreasing volatility is roughly S, Na-K, Fe-Mn-Cr, Si,Mg, and Ca-Ti-Al. This order is qualitatively consistent withthe variations in chondrule compositions. However, the de-gree to which these variations were inherited from chon-drule precursors or generated during chondrule formation isunknown (e.g., Gooding et al., 1980; Grossman and Was-son, 1983b; Jones, 1990; Connolly et al., 2001; Huang etal., 1996; Sears et al., 1996; Alexander, 1994, 1996, Hewinset al., 1997). The bulk chemistry of chondrules alone is in-sufficient to resolve this debate. Two other approaches havebeen used: petrologic evidence for volatile loss and recon-densation in chondrule rims and matrix (see section 7) andisotopic signatures of evaporation and recondensation.

Both evaporation and condensation are generally accom-panied by mass-dependent isotopic fractionation. Evapora-tion can result in enrichment in the heavy isotopes in theresidue. Condensation enhances the abundances of lightisotopes in the condensates, relative to the gas. The largestisotopic fractionations occur during Rayleigh fractionationand the isotopic fractionations correlate with the extent ofvolatile loss or gain. For Rayleigh fractionation to occur,the evaporating residue or condensing gas must remain iso-topically uniform and the evaporation or condensation mustbe unidirectional (i.e., there is no back reaction). At the otherextreme, if diffusion or convection occurs in the evaporat-ing material on timescales that are much slower than evapo-ration, mass loss and elemental fractionation can occur withlittle or no isotopic effects. Similarly, high ambient gas pres-sures suppress isotopic fractionation and absolute evapo-

ration rates (e.g., Richter et al., 2002), as will any back reac-tion with the evaporated gas.

Whether or not volatile-element fractionation is accom-panied by isotopic fractionation depends on the conditionsunder which the volatile loss took place. Under Rayleighconditions, large isotopic fractionations will correlate withelemental fractionation. Such correlated elemental and iso-topic fractionations are seen in low-pressure evaporationexperiments (section 6.3) and in cosmic spherules that areseverely heated during atmospheric entry heating (Alex-ander et al., 2002). If volatile loss occurs far from the Ray-leigh condition, there is little associated isotopic fractiona-tion. In this case, it is not possible to tell whether volatileloss was inherited from chondrule precursors or producedby chondrule formation.

Since there is this potential ambiguity in the elementaland isotopic data, considerable effort has been put intodeveloping a theoretical understanding of the kinetics ofevaporation and condensation associated with chondruleformation. The ultimate aim of these efforts is to be ableto predict the volatile loss, and accompanied isotopic frac-

Fig. 3. Major-element abundances, normalized to CI chondritesfor (a) type-I and (b) type-II chondrules.


tionations, that would be expected under a given set ofchondrule formation conditions. By comparing these pre-dictions to observations of chondrule compositions, it ishoped the conditions of chondrule formation can be fur-ther constrained. Here we review the elemental, isotopic,and petrologic studies of chondrules and the experimentaland theoretical progress that has been made in understand-ing evaporation and condensation, in an attempt to placelimits on the conditions of chondrule formation.

6.2. Evidence for Evaporation andRecondensation

6.2.1. Sulfur. Sulfur is the most volatile major elementin chondrites. Consequently, it is expected to rapidly evapo-rate from chondrules and recondense on metal in rims andmatrix. Despite it being a potentially good indicator forevaporation, our knowledge of the petrology and cosmo-chemistry of S in chondrules is fairly limited. Troilite isuncommon in FeO-poor chondrules in Semarkona. Hewinset al. (1997) showed that the abundance of S in SemarkonaFeO-poor (type-I) chondrules decreases with increasing sizeof olivine phenocrysts (Fig. 4). They argue that increasinggrain size indicates more severe heating and, therefore,fewer crystal nuclei surviving to develop into phenocrysts oncooling. Tachibana and Huss (2005) measured S isotopesof small troilite grains included in Fe-Ni metal spherulesin six Bishunpur FeO-poor chondrules and four large troi-lite grains in the matrix. All the troilite grains have the sameS-isotopic compositions within analytical errors. Since evap-oration of Fe-S melt under Rayleigh conditions does resultin isotopic mass fractionation (McEwing et al., 1980), thelarge elemental fractionation without accompanying isoto-pic mass fractionation in chondrules indicates either vola-tilization of S under non-Rayleigh conditions or subsequentisotopic reequilibration.

6.2.2. Sodium and potassium. As with S, Na and Kabundances in type-I chondrules decrease with increasinggrain size of olivine phenocrysts (Fig. 4), suggesting pro-gressive alkali loss with the increasing severity of heating.

Kong and Palme (1999) showed that chondrules and ma-trix in Renazzo are complementary in the elemental abun-dances; chondrules are depleted in refractory and volatileelements. There are also observations that suggest that Naand K may have recondensed into chondrules after their for-mation. Ikeda (1982) was the first to report an increase inNa, K, and Fe near the edge of a Ca-Al-rich chondrule inALH 77003. Matsunami et al. (1993) described a Semar-kona chondrule, in which Na, Mn, and Si increase outward,that they suggest resulted from recondensation after crys-tallization of forsterite. Libourel et al. (2003) reported apositive correlation between SiO2 and Na2O in mesostasisand enrichment of Si in the outer portion of some FeO-poorchondrules in CR2 chondrites and Semarkona. They inter-pret these observations as resulting from recondensation ofalkalis and SiO2. It is, however, not clear whether the zon-ing pattern is an indication of condensation or later meta-somatic exchange of Na and Ca between chondrules andmatrix (Grossman et al., 2002).

Sodium is monoisotopic but K has two stable isotopes.The isotopic composition of K has been measured in wholeAllende (CV3) chondrules (Humayun and Clayton, 1995)and chondrule mesostasis and melt inclusions in Semarkona(LL3.0) and Bishunpur (LL3.1) (Alexander et al., 2000;Alexander and Grossman, 2005). Most K-isotopic analy-ses are normal within error with no evidence of systematicisotopic fractionation with K abundance. These results ruleout volatile loss of K under Rayleigh conditions.

6.2.3. Iron, nickel, manganese, and chromium. The re-duced state of Fe is the defining criterion for type-I chon-drules, and its abundance variation among chondrules is themost significant of all major elements (Fig. 3). The complexbehavior of Fe as metal, oxide, or sulfide has made identi-fying the origin(s) of the depletions difficult. Nevertheless,evidence exists for evaporative loss of some siderophile ele-ments in CR2 chondrules (see section 5.3).

Alexander and Wang (2001) found that olivines (Fo99.7to Fo78) in nine Chainpur (LL3.4) chondrules are Fe-isoto-pically unfractionated within analytical errors (2σ < 1–2‰/amu). Bulk Tieschitz (H3.4) chondrules show very smallfractionations of <–0.25‰/amu (Kehm et al., 2003). Thesmall light-isotope enrichments may be due either to con-densation or parent-body processes. Four Allende chon-drules show mass-dependent isotopic fractionation of upto 1.1‰/amu (Mullane et al., 2003). However, Allendechondrules are chemically and isotopically zoned, includingin Fe, and this zonation is likely a metamorphic overprint(Zhu et al., 2001). The isotopic mass fractionation of othersiderophile elements and Cr in chondrules have not beenmeasured.

The textures of barred-olivine chondrules may also bean indication of evaporation during chondrule formation(Tsuchiyama et al., 2004). Partial evaporation (mainly FeO)enhances the selective nucleation by cooling the surface,relative to the interior. This process produces the compact,coarse-grained rims that are characterisitic of these chon-drules. The occurrence of such rims on barred olivine chon-drules suggests chondrule formation in an open system.

Fig. 4. Graph illustrating the variation in volatile-element abun-dances as a function of nominal grain size in chondrules.


6.2.4. Silicon and magnesium. If type-I chondrulesformed from type-II- or CI-like precursors, there must beloss of both FeO and SiO2 during chondrule formation(Jones, 1990; Sears et al., 1996, Cohen et al., 2004). Thisscenario is consistent with the FeO- and SiO2-rich rims thatare often present on type-IA chondrules in ordinary chon-drites. There is petrologic evidence for recondensation ofSiO2 back into chondrules (Matsunami et al., 1993; Libourelet al., 2003). In the most highly reducing melts, it is pos-sible that a significant amount of Si was incorporated intothe Fe-based metallic melt component. Expulsion of someof this material is another mechanism for simultaneous Feand Si loss from bulk chondrule compositions (Lauretta andBuseck, 2003).

The Si-isotopic compositions of 34 individual chondrulesfrom type-3 ordinary chondrites analyzed by Clayton et al.(1991) exhibit a range of ±0.25‰/amu about the bulk chon-drite value. If type-I chondrules were part of the analyzedsuite of chondrules, the observed range of Si-isotopic com-positions is less than would be expected if Si was lost viaRayleigh evaporation.

Evaporative loss should be faster for smaller objects be-cause of the large surface area/volume ratio. Clayton et al.(1991) analyzed chondrules of varying size from type-3ordinary chondrites. For Dhajala (H3.8), smaller chondrulesare isotopically heavier. However, the total change in Si-isotopic fractionation is only ~0.5‰/amu. For Chainpur(LL3.4) there was no systematic isotopic variation with chon-drule size.

Magnesium should not experience much evaporative lossfrom ferromagnesian chondrules, although significant Mgloss may have occured during formation of Al-rich chon-drules. Galy et al. (2000) measured Mg-isotopic fraction-ations of up to ~1.5‰/amu in eight Allende chondrules.They also found that the degree of fractionation is inverselycorrelated with the bulk Mg/Al ratios of the chondrules.However, the isotopic fractionations are much smaller thanexpected for Mg evaporation under Rayleigh conditions.

Young et al. (2002) measured Mg-isotopic compositionsin four Allende chondrules. One barred olivine chondruleis isotopically zoned; the center has δ25Mg = 1.9–1.7‰(relative to a standard SRM980), whereas that of the mar-gin is 0.3‰. An interesting result is that the Mg-isotopicvariation is correlated with δ17O. The authors suggest thatmixing of prechondrule materials with distinct Mg- and O-isotopic compositions and later alteration may be respon-sible for this pattern.

6.3. Experimental Evaporation ofChondrule Melts

Kinetic parameters for gas-condensed phase reactionsalong with thermodynamic data can be applied to modelsof reaction processes relevant to chondrule formation. Todate, the range of isotopic fractionations in S, K, Fe, Si,and Mg measured in chondrules are much smaller than ex-pected for elemental fractionation resulting from evapora-tion under Rayleigh conditions. A number of suggestions

have been made about how evaporation from chondrulescould be achieved without corresponding isotopic fraction-ations. Subsolidus evaporation of S- and K-bearing chon-drule precursors (Young, 2000; Tachibana and Huss, 2005)is one possibility. High partial pressures of H could lead tonon-Rayleigh evaporation of melts (e.g., Nagahara andOzawa, 1996c; Galy et al., 2000). Diffusion of evaporat-ing species through the ambient gas can also suppress iso-topic fractionation (Richter et al., 2002). Alternatively,chondrules may have exchanged with the ambient gas dur-ing chondrule formation (Alexander et al., 2000; Alexander,2004).

6.3.1. Types of evaporation experiments. Experimentalstudies of silicate-melt evaporation are summarized in Table 4.In earlier experiments, continuous or stepwise heating ex-periments were employed to see how chondritic materialssuffered from “evaporation metamorphism” (Hashimoto etal., 1979) or to see how alkali elements were fractionatedduring evaporation from lunar materials (Gibson and Hub-bard, 1972; Gibson et al., 1973). These experiments focusedon single condensed phases. Hashimoto (1983) performedsystematic isothermal experiments on multicomponent sili-cate melts, which provide an excellent dataset for model-ing chemical fractionation (e.g., Alexander, 2001, 2002).Since then, a large number of experiments have been per-formed to understand the kinetics of gas-silicate melt reac-tions aimed at understanding the chemical and isotopicfractionations during chondrule formation.

Free evaporation experiments have been performed atpressures of less than ~10–4 bar for various starting com-positions, such as chondrites (Murchison and Allende, etc.)(Floss et al., 1996; Hashimoto et al., 1979), a synthetic CIcomposition (Hashimoto, 1983; Wang et al., 1994, 2001),the CMAS system with a type-B CAI composition (Rich-ter et al., 2002; Richter et al., 1999) or near-chondriticcomposition (Nagahara and Ozawa, 2003), and a simpleMg2SiO4-SiO2 system (Nagahara, 1996; Nagahara andOzawa, 1996a,b). Experiments that provide unequivocalinformation on evaporation must be isothermal and in theabsence of crystalline phases that complicate evaporationprocesses by inhibiting direct evaporation from the silicatemelt (Nagahara, 1996; Ozawa and Nagahara, 1997). Freeevaporation experiments provide evaporation coefficients,the extent of inhibition of evaporation reactions, and un-equivocal proof that chemical and isotopic fractionationsare generally coupled.

Experiments performed in the presence of ambient gasare grouped into four categories depending on the type ofthe ambient gas: (1) air (or residual gas) at pressures lessthan 1 atm (Notsu et al., 1978; Shimaoka and Nakamura,1991), (2) 1-atm experiments with specified oxygen fugac-ity controlled by flowing either H2/CO2 or CO/CO2 mixture(Richter and Davis, 1997; Richter et al., 1999, 2002; Shiraiet al., 2000; Tsuchiyama et al., 1981), (3) H2 or He gas rep-resenting solar-nebula conditions that may enhance evapo-ration (Richter et al., 2002; Richter et al., 1999; Wang et al.,1999; Yu et al., 2003), and (4) gas with species consistingof evaporating condensed phases that may inhibit evapora-


tion (Nagahara, 2000; Nagahara and Ozawa, 2000, 2003).Evaporation is suppressed at high partial pressures of melt-component species or high total pressures due to signifi-cant back reaction. Evaporation is enhanced at low oxygenfugacity and high H pressures.

6.3.2. Major-element fractionations during evaporation.Compositional changes during free evaporation of CI- and

type-B CAI-like compositions have also been studied(Hashimoto, 1983; Richter et al., 2002; Wang et al., 2001).In both systems, Fe is the most volatile, followed by Si, Mg,Ti, Ca, and Al. Silicon and Mg have fairly similar evapora-tion behavior; however, Si evaporates more rapidly at earlierstages and Mg at later stages of evaporation. Aluminum, Ca,and Ti do not evaporate until the final stages of mass loss.

TABLE 4. Experimental studies of gas-liquid reactions related to ferromagnesian chondrule formation.

Starting Composition Temperature (°C) Pressure (Pa) References

Evaporation in vacuumLunar basalts, breccias, and soil 950–1400 ~2.5 × 10–4 Gibson and Hubbard (1972),

Gibson et al. (1973)Murchison (CM) meteorite 1300, 1900 <10–3 Hashimoto et al. (1979)Synthetic chondritic Si-Al-Fe-Mg-Ca 1700–2000 <10–5 Hashimoto (1983)Etter (L5) chondrite 1150–1450 ~10–3 Shimaoka and Nakamura (1989)Allende (CV) meterorite peak: 1500–2100 ~10–4 Floss et al. (1996)Fe0.947O 1450–1750 ~10–4 Wang et al. (1994)Oxide mixture with CI composition peak: 1500–2000 ~10–4 Wang et al. (1994, 2001) Fe-Mg-Si-Ca-Al-Ti-REEMg2SiO4–SiO2 600, 1700 <10–4 Nagahara and Ozawa (1996a,b),

Nagahara (1996)Synthetic Type B CAI-like Ca-Al-Si-Mg 1700, 1800 <10–4 Richter et al. (1999, 2002)Synthetic chondrule-like CMAS 1500, 1600 <10–4 Nagahara and Ozawa (2003)

Evaporation in gasAllende (CV), JB-1 basalt ~2000 or 1300–2000 105 or ~1 Notsu et al. (1978)Etter (L5) chondrite 1300 10–10–3 Shimaoka and Nakamura (1991)Synthetic Type IIAB chondrule 1450, 1485 1–140 Yu et al. (2003), Wang et al. (1999)Mixture of silicates from H5 chondrite 1450–1600 105 Tsuchiyama et al. (1981)Knippa basalt 1300, 1325, 1350 105 Lewis et al. (1993)Synthetic Type IIAB and Type I peak: 1470–1750 105 Yu et al. (1996) and IA chondrules + sulfideMineral mix with chondrule-like peak: 1470–1750 105 Yu and Hewins (1998) compositionSynthetic Type B CAI-like Ca-Al-Si-Mg-Ti 1400 105 Richter and Davis (1997),

Richter et al. (1999, 2002)Synthetic Type IIAB chondrule 1450 105 Yu et al. (2003), Wang et al. (1999)SiO2-Na2O system 1300–1400 105 Shirai et al. (2000)Synthetic Type B CAI-like Ca-Al-Si-Mg-Ti 1500 ~20 Richter et al. (1999, 2002)Synthetic Type IIAB chondrule 1450, 1485 7–103 Yu et al. (2003), Wang et al. (1999)Mineral mixture with CI composition 1580 13 Cohen et al. (2000)Jilin (H5) meteorite 1300, 1400 PNa = 0 ~ 1.0, Nagahara and Ozawa (2000),

PK = 0 ~ 0.07 Nagahara (2000)Synthetic chondrule-like CMAS 1500, 1600 NA Nagahara and Ozawa (2003)

Transpiration (condensation?)Knippa basalt 1330 105 Lewis et al. (1993)Allende (CV), Ornans (CO) (±quartz), 1400–1500 105 Yu et al. (1995) Eagle Station pallasiteAllende (CV), Ornans (CO) chondrites + quartz 1500–1000 105 Yu et al. (1995)SiO2 + C source, CMAS, or target: 1330–1480 8–30 Tissandier et al. (2000; 2002) Murchison (CM) melt targetK2CO3 + C source, CMAS ± K target source: 1000, 105 Georges et al. (1999)

target: 1400

Dust melt collisionSynthetic Type IIA, Type IIAB chondrules melt: 1564–1600, 105 Connolly and Hewins (1995)

dust: 1100–1560


6.3.3. Trace-element fractionations during evaporation.Rare earth elements experience little evaporation from sili-cate melts in vacuum. An exception is Ce, which evapo-rates at a rate similar to Fe (Floss et al., 1996; Wang et al.,2001). This behavior is thought to result from oxidation ofCe2O3 to the more volatile form CeO2 in the residual melt.The REE abundance, relative to CI, along with the degreeof Ce depletion can thus be used as an indicator for evapo-ration of CI-composition precursors under oxidizing con-ditions.

6.3.4. Isotopic fractionation due to evaporation. Experi-mental evaporation of chondritic materials causes mass-de-pendent isotopic fractionation, which results in heavy iso-tope enrichment in Mg, Si, Ca, Ti, K, and O. The heavyisotope enrichments follow the Rayleigh law provided thatdiffusion within the experimental melts is not a rate-limit-ing process. The degree of mass fractionation is smaller thantheoretical expectations based on thermodynamic consid-erations (Floss et al., 1996; Wang et al., 2001; Richter etal., 2002). However, the relative degree of isotopic fraction-ation decreases with increasing atomic number (Wang et al.,2001). Kinetic processes at the surface of an evaporatingmelt may suppress isotopic fractionation.

6.4. Theoretical Modeling ofChondrule Melt Evaporation

6.4.1. Stability of silicate melt in equilibrium conditions.Although it is well known that chondrule formation in-volved kinetic processes, it is important to know the envi-ronmental conditions required to stabilize silicate melts.Wood and Hashimoto (1993) suggested that dust enrichmentgreater than 1000× solar is required to stabilize silicate meltin the solar nebula, assuming ideal mixing of the multicom-ponent silicate melt. Nagahara and Ozawa (1995) andOzawa and Nagahara (1997) examined melt stability overa wide range of dust-enrichment factors, and delineated thestability field as a function of total pressure vs. dust enrich-ment at a given temperature, and determined that the totalpressure of 10–6 bar is the lower limit of melt stability at1627°C. Yoneda and Grossman (1995) incorporated non-ideality in CMAS melts using the model of Berman (1983)and examined the stability of this melt over a wide range oftemperatures, pressures, and dust enrichments. They foundthat CMAS melts are stable at a much lower dust enrich-ments than suggested by Wood and Hashimoto (1993). Ebeland Grossman (2000) examined the stability of ferromag-nesian melts in the solar nebula and suggested that dustenrichments of 20× solar stabilizes ferromagnesian meltsat 10–3 bar and 1127°–1527°C.

6.4.2. Kinetic models. Equilibrium may have beenachieved between gas and condensed phases at high tem-peratures in the early solar nebula. However, petrographicobservations show that chondrule formation involved ki-netic processes, such as rapid heating and cooling. Kineticaspects that must be considered in modeling chondrule gas-melt reactions are (1) surface reaction kinetics, which is themajor barrier for evaporation and condensation; (2) elemen-

tal transport in the melt through diffusion or convection;(3) the incongruent nature of evaporation and condensation;and (4) transportation of gas species in the ambient gas.

Surface reaction kinetics have been treated by applyingthe Hertz-Knudsen equation

2πmkT

Peq − P Jmax = (1)

to multicomponent systems, which is, in the strictest sense,good only for metallic phases in the presence of monatomicgas molecules. The maximum possible evaporation rate(Jmax) is proportional to the difference between the equi-librium partial pressure (Peq) and the actual pressure (P) andthe average velocity of gas molecules. Kinetic barriers areincorporated into the above equation by multiplying P andPeq by evaporation and condensation coefficients (αevap andαcond), respectively, to give the actual evaporation rate J as

2πmkTJ =

αevapPeq ⋅ αcondP (2)

The reason why this simple equation cannot be extendedto more complex oxide systems without ambiguity is thatthe presence of other components involved in the reaction,such as O or H, cannot be properly treated by the equation.In spite of this fact, the ease of use of this approximationand the absence of well-established alternatives means thatmost researchers use it in their modeling. One approachcombines thermodynamic calculations for a given gas com-position to obtain the driving force and the evaporationcoefficient, αevap, is estimated from the results of free evapo-ration experiments (Alexander, 2002; Ebel, 2002; Grossmanet al., 2002; Richter et al., 1999). Another method expressesthe term for a specific evaporation (or condensation) reac-tion as a function of the partial pressure of related gas spe-cies in the ambient gas, the activity of the component inthe melt multiplied by evaporation coefficient, and the equi-librium constant for the reaction (Alexander, 2001, 2002;Shirai et al., 2000; Tsuchiyama and Tachibana, 1999). Theequilibrium constant is assumed to be extendable to kineticdomains according to the principle of detailed balancing(Lasaga, 1998). A third approach uses equation (2) andcombines kinetic effects and nonideality of silicate meltsinto the evaporation coefficient (Ozawa and Nagahara,2001). This approach is applicable to the situation whereO is much more abundant than metallic gas species.

6.5. Isotopic Fractionation or Lack Thereof

The fundamental discrepancy between chemical and iso-topic fractionations in chondrules is that they are depletedin volatile elements but not isotopically fractionated. Thisproblem was first addressed by Humayun and Clayton(1995) and discussed intensively by many later workers.This decoupling between chemical and isotopic fraction-ations is a major constraint on the origin of chondrules.There are several possible explanations. Highly volatile


elements can be fractionated with negligible fractionationif the evaporation Peclet number (Pe) is very large (Ozawaand Nagahara, 1998; Galy et al., 2000). The Peclet num-ber reflects the relative rates of diffusion and evaporationand is defined as

DJ ⋅ r

Pe = (3)

where J is the elemental evaporation rate, r is the radius ofthe condensed grain, and D is the elemental diffusion rate.Alternatively, volatile elements can be fractionated withnegligible isotopic fractionation if there is extensive backreaction during evaporation (Ozawa and Nagahara, 1998).Finally, the presence of a gaseous atmosphere inhibits theescape of evaporated gas species and decreases the isotopicfractionation factor to a value close to unity, thus suppress-ing isotopic fractionation while preserving notable chemi-cal fractionation (Richter et al., 2002).

Diffusion is not expected to play a limiting role in chon-drule melt droplets less than a few millimeters in size. Backreaction with the gas is essential if equilibrium between gasand silicate melt is approached. However, it is questionableif total equilibrium was achieved in the solar nebula. Evenif chondrule melts could not attain equilibrium with the sur-rounding vapors, isotopic fractionation of volatile to mod-erately volatile elements would have been suppressed whilestill achieving elemental fractionations if the chondrulemelts experienced a limited range of cooling rates (Ozawaand Nagahara, 1998). The slow escape rate of evaporatinggas species due to the presence of H is only effective if Hpressure is higher than 10–3 bar (Richter et al., 2002).

Combining the evidence for evaporation and recondensa-tion recorded in natural chondrules, experiments, and mod-eling, the most important conclusions are summarized asfollows: (1) Chondrule precursors were heated to lose mod-erately volatile components by evaporation. The degree ofevaporative loss varied depending on the melt composition,grain size, rate of heating, total heating time, and the physi-cal conditions such as pressure and temperature. (2) Theevaporated gas was not lost from the system. Rather, signi-ficant recondensation took place during cooling. Lithophileelements reentered crystallizing chondrules to various de-grees, but siderophile and chalcophile elements condensedas new phases, which are now in chondrule rims and/or ma-trix. (3) Isotopic fractionation should have been significantat the early stages of evaporation; however, isotopic reequili-bration ocurred during subsequent cooling, which resultedin little isotopic fractionation preserved in chondrules.

7. VOLATILE RECONDENSATION ANDFORMATION OF CHONDRULE RIMS

Rims around chondrules are an additional source of in-formation about chondrule formation conditions, as well astheir precursors. In addition, rims and interchondrule matrixare important complementary components to chondrulesthat are required to make up the various chondrite group

bulk compositions, and all three components may have beenderived from a common reservoir (Klerner and Palme,1999). It should be noted that in many studies no distinc-tion is made between chondrule rims and interchondrulematrix, and they are collectively referred to as matrix. How-ever, rims partially or completely surrounding chondrulesare texturally and often chemically distinct from the inter-chondrule matrix (Ashworth, 1977; Allen et al., 1980; Scottet al., 1988).

Chondrule rims can be broadly divided into either ac-cretionary or igneous rims. As the name implies, accretion-ary rims formed by the accretion of fine-grained dust ontothe chondrules. Accretionary rims exhibit a range of com-positions and textures from fine-grained, FeO-rich rims withfew discernable grains even in secondary electron images,to clastic rims that chemically, texturally, and mineralogi-cally resemble the interchondrule matrix (Ashworth, 1977;Alexander et al., 1989b). Igneous rims, on the other hand,have relatively coarse-grained igneous textures indicative ofpartial melting (Rubin, 1984; Rubin and Wasson, 1987; Krotand Wasson, 1995).

Perhaps the best-developed accretionary rims occur inCM chondrites (Metzler et al., 1992; Lauretta et al., 2000;Zega and Buseck, 2003). In the unbrecciated areas of CMchondrites (so-called primary rock), rims surround almostall objects, including chondrules, chondrule fragments, andmineral fragments. The most abundant phase in rims aroundchondrules in CM chondrites is Mg-rich serpentine. Otherminerals embedded within the rims include cronstedtite,tochilinite, chlorite, pentlandite, gypsum, olivine, kamacite,taenite, and chromite. In these rims, regions containing hy-drated and anhydrous minerals are in direct contact witheach other, which has been interpreted as rim material ac-creting directly from multiple nebular reservoirs. It has alsobeen suggested that aqueous alteration and metamorphismon the CM chondrite parent body has significantly modifiedthe primary mineralogy of the accretionary rims (Brearleyand Geiger, 1993; Chizmadia and Brearley, 2003). Thereis a clear relationship between the radius of the core ob-ject and the thickness of its rim. A similar relationship hasbeen reported for the CV chondrites (Paque and Cuzzi,1997) and may or may not be present in the ordinary chon-drites (Matsunami, 1984; Huang et al., 1993).

In a few meteorites, such as Bishunpur (LL3.1) andALH 77307 (CO3.0), there is an abundance of amorphousmaterial in the rims and interchondrule matrix (Alexanderet al., 1989a,b; Brearley, 1993). Amorphous material is verysusceptible to alteration and recrystallization, and its pres-ervation suggests that some of the primary material hassurvived. Brearley and Geiger (1993) and Chizmadia andBrearley (2003) argue from the lack of phyllosilicate pseu-domorphs that CM rims were also largely amorphous priorto alteration.

There are likely to have been several sources of amor-phous material. In Bishunpur, chondrule mineral fragmentsare abundant in the interchondrule matrix and rims. Thesemust be accompanied by amorphous, Al-, Ca-, and Si-richchondrule mesostasis. Indeed, the finest-grained rim and


matrix material tends to be enriched in these elements.However, in ALH 77307 clastic mineral fragments are muchless abundant, leading Brearley (1993) to suggest that muchof the amorphous material is a condensate. Numerous au-thors have suggested that rim and interchondrule matrix ma-terial is composed of nebular condensates (e.g., Larimer andAnders, 1967; Huss et al., 1981; Nagahara, 1984).

An alternative source of condensates is material evap-orated during chondrule formation. The shorter timescaleof chondrule formation would also be more conducive toformation of amorphous material. There is some direct evi-dence of recondensation during chondrule formation. Iron,Si, and Mn are among the most volatile of the major andminor elements, and they are depleted in type-I chondrules.Alexander (1995) described compositionally and texturallyzoned rims around some type-I Bishunpur chondrules, withlarge FeO, Si, and Mn enrichments at the rim-chondruleinterface that often grade into matrix-like compositions andtextures at the rims’ outer edges. Transmission electronmicroscopy of rims with similar compositions shows thisFe-,Si-rich material to be amorphous. There is also evidenceof partial to complete reequilibration of clastic olivine andpyroxene grains with the FeO-rich material, as well as re-action with protuberances from the chondrules.

Petrologic evidence shows that accretion of rims beganbefore expulsion of metal and sulfide from chondrules wascomplete, and that the metal/sulfide that was trapped in therims reacted with the nebular gases as the system cooled(Allen et al., 1980; Alexander et al., 1989b; Lauretta andBuseck, 2003). In the CR chondrites, at least, there is evi-dence of both evaporation of metal from chondrules andits recondensation onto rims (Kong and Palme, 1999; Con-nolly et al., 2001; Humayun et al., 2002; Zanda et al.,2002).Finally, rims and matrix-like material are sometimes trappedbetween plastically deformed or compound chondrules,indicating that the chondrules were still hot when they col-lided (e.g., Hutchison and Bevan, 1983; Scott et al., 1984;Holmen and Wood, 1986; Lauretta and Buseck, 2003). Allthese features suggest that the rims began accreting whilethe chondrules were still hot.

While Fe and Si are relatively volatile, the alkalis aremuch more so. If there was some evaporation of Fe and Siduring chondrule formation, there should have been con-siderable loss of the alkalis from chondrules, and enrich-ment of alkalis in rims and interchondrule matrix. Evidencefor recondensation of alkalis in rims is hampered by theirmobility during even very mild aqueous alteration or meta-morphism. Rims and interchondrule matrix are generallyenriched in alkalis relative to chondrules in the least altered/metamorphosed chondrites. Of the OCs, alkali enrichmentsare greatest in Semarkona (LL3.0), but even in this mete-orite there is ample evidence of reentry of alkalis into chon-drules (Grossman et al., 2002).

While there is evidence that some rims accreted whiletheir host chondrules were still hot, this cannot be true forsome or, perhaps, even most rims. The rims and matrix ofchondrites are the host of presolar material: circumstellargrains and interstellar organic matter. Indeed, the abundances

of presolar material in rim/matrix in the least-metamor-phosed members of the chondrite groups are very similar tothat of CI chondrites (Huss and Lewis, 1995; Alexander etal., 1998). The CI chondrites are thought to have containedfew if any chondrules or CAIs. They most likely escapedhigh-temperature processes prior to accretion and their pre-solar abundances should most closely resemble those of theprimordial dust. The implication of the presolar abundances,particularly since rims make up the bulk of the fine-grainedmaterial in many chondrites, is that the major componentof chondrule rims and matrix resembled CI chondrites incomposition and did not experience the high temperaturesassociated with nebula-wide condensation or with chondruleformation. The CI chondrites are almost completely aque-ously altered, but perhaps the most likely analogs for themprior to alteration are the anhydrous interplanetary dust par-ticles. They are very fine-grained and can contain abundantamorphous material.

Intuitively, one might expect accretionary rims thatformed in a dusty nebula to resemble loosely packed dustballs. Yet, accretionary rims have surprisingly low porosity(<20%). They were also robust enough to survive accre-tion onto the chondrite parent bodies, as well as any regolithprocessing that occurred prior to final burial and compac-tion. Recondensation of vaporized material from chondruleswould help cement a rim, as well as fill in porosity. Mod-els and experiments (e.g., Dominik and Tielens, 1997; Blumand Wurm, 2000) find that sticking of micrometer-sized dustis most efficient when collisions occur at less than 10–20 m/s.At speeds of up to several meters per second, the collisionsactually tend to compact fluffy aggregates. Electrostaticattraction between grains may enhance sticking efficiencies,as well as increase the range of relative speeds over whichsticking and compaction occurs (Marshall and Cuzzi, 2001).The most detailed modeling of rim formation has been donein the context of the turbulent concentration mechanism(Cuzzi, 2004). In a weakly turbulent nebula, the timescalesfor making rims by this mechanism are likely to be 10–1000 yr. Such timescales are possible for rims that accretedonto cold chondrules, but are clearly inconsistent with rimsthat show evidence of accretion while the chondrules werestill hot.

Igneous rims appear to have formed by partial meltingof preexisting fine-grained rims (Rubin, 1984; Rubin andWasson, 1987; Krot and Wasson, 1995). They probably forma continuum with dark-zoned or microporphyritic chon-drules, and with so-called enveloping compound chondrules(Wasson, 1993). Their fine grain size and their limited in-teraction with their host chondrules suggest that igneousrims formed at lower peak temperatures and from materialwith lower solidus temperatures than their host chondrules.More intense heating would presumably lead to the mutualassimilation of rim and host, producing a new chondrule.Thus, it is likely that fine-grained rim and matrix materialwas one component of the chondrule precursors.

In summary, the presence of fine-grained rims aroundchondrules suggests that they formed in relatively dusty en-vironments, consistent with chondrule cooling rates. In at


least some instances, rims began accreting while the chon-drules were still hot and acted as sites for recondensationof material evaporated from the chondrules. The dust mayalso have provided nuclei for completely melted chondrules(see section 4.1). Whether this early accreted material sur-vived chondrule formation or was mixed into the chondrule-forming region soon after it began to cool is unclear. Whilerecondensation may have aided rim accretion at this stage,it is likely that by the time rims began to accrete the rela-tive velocities of chondrules and micrometer-sized dustwere <10–20 m/s, placing some constraints on models thatinvolve gas drag for heating and/or sorting of chondrules.Later, mixing of cold chondrules and unheated material musthave occurred. The CI-like abundances of presolar materialin rim/matrix requires that most rim formation occurred afterchondrules had cooled. Finally, the presence of igneous rimson chondrules point to multiple episodes of chondrule andrim formation, requiring a relatively dynamic environment,and suggests that fine-grained rim/matrix material was onecomponent of the chondrule precursors.

8. CONSTRAINTS ON MODELS OFCHONDRULE FORMATION

The primary purpose of this chapter is to summarize thepetrologic history of those chondrules dominated by ferro-magnesian silicates. It is not the intent of this chapter todiscuss various models for chondrule formation, whichplace this history in the larger context of environments andprocesses in the early solar system (see Connolly et al.,2006). However, the information presented here is essentialfor constraining models of chondrule formation. Any suchmodel must account for all the features present in chon-drules, summarized below.

Chondrule precursors were composed of anhydrous min-erals, predominantly fine-grained olivine (<5 µm). However,larger grains of olivine, up to 400 µm, were also incorpo-rated. Lesser amounts of pyroxene and feldspathic glass,chromite, metal, and sulfide were likely also present. Thepresence of sulfide places some constraints on the minimumtemperatures of chondrule precursors. In a canonical solar-composition system, this temperature is ~427°C, where sul-fide condensation begins (Lauretta et al., 1996). However,in systems with enhanced S/H ratios, relative to solar, thistemperature could be as high as 1000°C for total gas pres-sures <10–4 bar (Lauretta et al., 2001). Carbonaceous mate-rial was included in some chondrule melts and likely playedan important role in controlling the oxidation state of theresulting mineralogy.

At least some chondrules formed through melting of pre-existing materials, indicated by the presence of large, com-positionally distinct grains. These relict grains were part ofthe precursor material that survived chondrule formationwithout melting or dissolving. Many of these grains appearto be fragments of both chondrules and refractory inclu-sions. These features imply that the process or processesof chondrule formation were repeated many times.

Chondrules in some carbonaceous chondrites show evi-dence of 16O-rich components. These signatures may beevidence of the survival or limited processing of a presolarcomponent in the chondrule precursors. Alternatively, theymay result from chemical processes in the gas phase thatoccurred prior to or during chondrule formation

Chondrule textures result from a complex interplay be-tween the peak temperature, duration of heating, and pres-ervation of nucleation sites, either due to incomplete melt-ing of chondrule precursors or the presence of fine-graineddust in the chondrule-formation region. Experimental stud-ies place a firm limit of 2100°C on the maximum tempera-ture of chondrule formation. Few chondrules with liquidustemperatures over 1750°C were completely melted, and fewwith liquidus temperatures under 1400°C were incompletelymelted. Most chondrules reached peaked temperatures of1500°–1550°C.

Simulation experiments and chondrule textures have alsoplaced constraints on chondrule cooling rates. The coolingrates required to produce porphyritic olivine textures are lessthan 1000°C/h with rates on the order of 100°C/h providingthe best experimental analogs. Porphyritic pyroxene texturesdevelop at slow cooling rates, on the order of 2°–50°C/h,regardless of the degree of undercooling. Cooling rates aslow as 5°C/h are also consistent with the zoning observedin Semarkona chondrules. Barred-olivine chondrules cooledat rates of 500° to 2300°C/h. Radial pyroxene chondrulesgrew from a melt that experienced very high degrees of un-dercooling at quite slow cooling rates (5°C/hr). Pigeonite/diopside exsolution lamellae suggest that these textures areproduced when cooling rates are on the order of 5°–25°C/h in the temperature range of 1200°–1350°C.

An important constraint on models of chondrule forma-tion is the variation in the oxidation states of their compo-nents. Type-I chondrules experienced in situ reduction eitherby incorporating reducing material in their precursors orinteracting with a reducing gas phase. Type-I chondrulesalso exhibit significant chemical fractionations, with deple-tions of elements like Cr, Mn, Na, K, and Fe, and enrich-ments of elements, such as Al, Ca, and Ti, relative to Mg.The order of these depletions and enrichments suggest vola-tility has played a role.

Chondrule precursors were heated and lost moderatelyvolatile components by evaporation. However, significantrecondensation took place during cooling. Lithophile ele-ments recondensed into crystallizing chondrules to variousdegrees during cooling, but siderophile and chalcophileelements condensed as new phases or recondensed intometal expelled from chondrules that did not completelyevaporate, which are now in chondrule rims and matrix.Isotopic fractionation should have been significant at theearly stage of evaporation; however, isotopic reequilibrationduring subsequent cooling must have been almost complete,resulting in little or no isotopic fractionation in chondrules.

Petrologic evidence shows that accretion of rims beganwhile chondrules were still molten, before expulsion ofmetal and sulfide from chondrules was complete. Metal and


sulfide was trapped in the rims and reacted with nebulargases as the system cooled. Finally, rims and matrix-likematerial are sometimes trapped between plastically de-formed or compound chondrules, indicating that the chon-drules were still hot when they collided. However, since therims and matrix of chondrites are the host of presolar ma-terial, much of this material must have accreted after chon-drules had cooled back down to ambient nebular conditions.Igneous rims appear to have formed by partial melting ofpreexisting fine-grained rims, suggesting that reheatingevents occurred well after the volatile-rich accretionary rimswere in place.

REFERENCES

Alexander C. M. O’D. (1994) Trace-element distributions withinordinary chondrite chondrules — Implications for chondruleformation conditions and precursors. Geochim. Cosmochim.Acta, 58, 3451–3467.

Alexander C. M. O’D. (1995) Trace element contents of chondrulerims and interchondrule matrix in ordinary chondrites. Geo-chim. Cosmochim. Acta, 59, 3247–3266.

Alexander C. M. O’D. (1996) Recycling and volatile loss in chon-drule formation. In Chondrules and the Protoplanetary Disk(R. H. Hewins et al., eds.), pp. 233–241. Cambridge Univ.,New York.

Alexander C. M. O’D. (2001) Exploration of quantitative kineticmodels for the evaporation of silicate melts in vacuum and inhydrogen. Meteoritics & Planet. Sci., 36, 255–283.

Alexander C. M. O’D. (2002) Application of MELTS to kineticevaporation models of FeO-bearing silicate melts. Meteoritics& Planet. Sci., 37, 245–256.

Alexander C. M. O’D. (2004) Chemical equilibrium and kineticconstraints for chondrule and CAI formation conditions. Geo-chim. Cosmochim. Acta, 68, 3943–3969.

Alexander C. M. O’D. and Grossman J. N. (2005) Alkali elementaland potassium isotopic compositions of Semarkona chondrules.Meteoritics & Planet. Sci., 40, 541–556.

Alexander C. M. O’D. and Wang J. (2001) Iron isotopes in chon-drules: Implications for the role of evaporation during chon-drule formation. Meteoritics & Planet. Sci., 36, 419–428.

Alexander C. M., O’D., Barber D. J., and Hutchison R. (1989a)The microstructure of Semarkona and Bishunpur. Geochim.Cosmochim. Acta, 53, 3045–3057.

Alexander C. M. O’D., Hutchison R., and Barber D. J. (1989b)Origin of chondrule rims and interchondrule matrices in un-equilibrated ordinary chondrites. Earth Planet. Sci. Lett., 95,187–207.

Alexander C. M. O’D., Russell S. S., Arden J. W., Ash R. D.,Grady M. M., and Pillinger C. T. (1998) The origin of chon-dritic macromolecular organic matter: A carbon and nitrogenisotope study. Meteoritics & Planet. Sci., 33, 603–622.

Alexander C. M. O’D., Grossman J. N., Wang J., Zanda B.,Bourot-Denise M., and Hewins R. H. (2000) The lack of potas-sium-isotopic fractionation in Bishunpur chondrules. Meteor-itics & Planet. Sci., 35, 859–868.

Alexander C. M. O., Taylor S., Delaney J. S., Ma P. X., andHerzog G. F. (2002) Mass-dependent fractionation of Mg, Si,and Fe isotopes in five stony cosmic spherules. Geochim. Cos-mochim. Acta, 66, 173–183.

Allen J. S., Nozette S., and Wilkening L. L. (1980) A study of

chondrule rims and chondrule irradiation records in unequili-brated ordinary chondrites. Geochim. Cosmochim. Acta, 44,1161–1175.

Ashworth J. R. (1977) Matrix textures in unequilibrated ordinarychondrites. Earth Planet. Sci. Lett., 35, 25–34.

Berman R. G. (1983) A thermodynamic model for silicate meltswith application to the system CaO-MgO-Al2O3-SiO2. Ph.D.thesis, University of British Columbia. 178 pp.

Bizzarro M., Baker J. A., and Haack H. (2004) Mg isotope evi-dence for contemporaneous formation of chondrules and re-fractory inclusions. Nature, 431, 275–278.

Blander M. and Katz J. L. (1967) Condensation of primordial dust.Geochim. Cosmochim. Acta, 31, 1025–1034.

Blander M., Planner H. N., Keil K., Nelson L. S., and RichardsonN. L. (1976) Origin of chondrules — Experimental investiga-tion of metastable liquids in system Mg2SiO4-SiO2. Geochim.Cosmochim. Acta, 40, 889–896.

Blander M., Unger L., Pelton A., and Ericksson G. (2001) Nuclea-tion constraints lead to molten chondrule precursors in the earlysolar system. J. Phys. Chem., B105, 11823–11827.

Blander M., Pelton A. D., Jung I. H., and Weber R. (2004) Non-equilibrium concepts lead to a unified explanation of the for-mation of chondrules and chondrites. Meteoritics & Planet.Sci., 39, 1897–1910.

Blum J. and Wurm G. (2000) Experiments on sticking, restructur-ing, and fragmentation of preplanetary dust aggregates. Icarus,143, 138–146.

Brearley A. J. (1993) Matrix and fine-grained rims in the unequili-brated CO3 chondrite, ALHA77307: Origins and evidence fordiverse, primitive nebular components. Geochim. Cosmochim.Acta, 57, 1521–1550.

Brearley J. and Geiger T. (1993) Fine-grained chondrule rims inthe Murchison CM2 chondrite: Compositional and mineral-ogical systematics (abstract). Meteoritics, 28, 328.

Campbell A. J., Humayun M., and Zanda B. (2002) Partial con-densation of volatile elements in Renazzo chondrules (abstract).Geochim. Cosmochim. Acta, 66, A117.

Carpenter M. A. (1979) Experimental coarsening of antiphasedomains in a silicate mineral. Science, 206, 681–683.

Chizmadia L. J. and Brearley A. J. (2003) Mineralogy and tex-tural characteristics of fine-grained rims in the Yamato 791198CM2 carbonaceous chondrite: Constraints on the location ofaqueous alteration (abstract). In Lunar and Planetary ScienceXXXIV, Abstract #1419. Lunar and Planetary Institute, Houston(CD-ROM).

Clayton R. N. (2004) The origin of oxygen isotope variations inthe early solar system (abstract). In Lunar and Planetary Sci-ence XXXV, Abstract #1682. Lunar and Planetary Institute,Houston (CD-ROM).

Clayton R. N., Mayeda T. K., Goswami J. N., and Olsen E. J.(1991) Oxygen isotope studies of ordinary chondrites. Geo-chim. Cosmochim. Acta, 55, 2317–2337.

Cohen B. A. and Hewins R. H. (2004) An experimental study ofthe formation of metallic iron in chondrules. Geochim. Cosmo-chim. Acta, 68, 1677–1689.

Cohen B., Hewins R. H. and Alexander C. M. O’D. (2004) Theformation of chondrules by open-system melting of nebularcondensates. Geochim. Cosmochim. Acta, 68, 1661–1675.

Connolly H. C. and Hewins R. H. (1995) Chondrules as productsof dust collisions with totally molten droplets within a dust-rich nebular environment — An experimental investigation.Geochim. Cosmochim. Acta, 59, 3231–3246.


Connolly H. C. and Hewins R. H. (1996) Constraints on chon-drule precursors from experimental data. In Chondrules andthe Protoplanetary Disk (R. H. Hewins et al., eds.), pp. 129–136. Cambridge Univ., New York.

Connolly H. C., Hewins R. H., Ash R. D., Zanda B., LofgrenG. E., and Bourot-Denise M. (1994) Carbon and the formationof reduced chondrules. Nature, 371, 136–139.

Connolly H. C., Jones B. D., and Hewins R. H. (1998) The flashmelting of chondrules: An experimental investigation into themelting history and physical nature of chondrule precursors.Geochim. Cosmochim. Acta, 62, 2725–2735.

Connolly H. C., Huss G. R., and Wasserburg G. J. (2001) On theformation of Fe-Ni metal in Renazzo-like carbonaceous chon-drites. Geochim. Cosmochim. Acta, 65, 4567–4588.

Connolly H. C. Jr., Desch S. J., Ash R. D., and Jones R. H. (2006)Transient heating events in the protoplanetary nebula. In Me-teorites and the Early Solar System II (D. S. Lauretta and H. Y.McSween Jr., eds.), this volume. Univ. of Arizona, Tucson.

Cuzzi J. N. (2004) Blowing in the wind: III. Accretion of dustrims by chondrule-sized particles in a turbulent protoplanetarynebula. Icarus, 168, 484–497.

Dehart J. M., Lofgren G. E., Jie L., Benoit P. H., and SearsD. W. G. (1992) Chemical and physical studies of chondrites10: Cathodoluminescence and phase-composition studies ofmetamorphism and nebular processes in chondrules of type-3ordinary chondrites. Geochim. Cosmochim. Acta, 56, 3791–3807.

Deloule E., Robert F., and Doukhan J. C. (1998) Interstellar hy-droxyl in meteoritic chondrules: Implications for the origin ofwater in the inner solar system. Geochim. Cosmochim. Acta, 62,3367–3378.

Dodd R. T. (1971) Petrology of chondrules in Sharps meteorite.Contrib. Mineral. Petrol., 31, 201.

Dominik C. and Tielens A. G. G. M. (1997) The physics of dustcoagulation and the structure of dust aggregates in space. Astro-phys. J., 480, 647–673.

Ebel D. S. (2002) Model evaporation of FeO-bearing liquids. Me-teoritics & Planet. Sci., 37, A43.

Ebel D. S. and Grossman L. (2000) Condensation in dust-enrichedsystems. Geochim. Cosmochim. Acta, 64, 339–366.

Everman R. L. A. and Cooper R. F. (2003) Internal reduction ofan iron-doped magnesium aluminosilicate melt. J. Am. Ceram.Soc., 86, 487–494.

Fagan T., McKeegan K. D., Krot A. N., and Keil K. (2001) Calci-um-aluminum-rich inclusions in enstatite chondrites (II): Oxy-gen isotopes. Meteoritics & Planet. Sci., 36, 223–230.

Faure F., Trolliard G., Nicollet C., and Montel J. M. (2003) Adevelopmental model of olivine morphology as a function ofthe cooling rate and the degree of undercooling. Contrib. Min-eral. Petrol., 145, 251–263.

Ferraris C., Folco L., and Mellini M. (2002) Chondrule thermalhistory from unequilibrated H chondrites: A transmission andanalytical electron microscopy study. Meteoritics & Planet.Sci., 37, 1299–1321.

Floss C. F., El Goresy A. E., Zinner E., Kransel G., RammenseeW., and Palme H. (1996) Elemental and isotopic fractionationsproduced through evaporation of the Allende CV chondrite:Implications for the origin of HAL-type hibonite inclusions.Geochim. Cosmochim. Acta, 60, 1975–1997.

Fries M. and Steele A. (2005) Carbonaceous inclusions in meteor-ite chondrules as characterized using confocal Raman imaging.Astrobiology, 5, 221.

Fujii N. and Miyamoto M. (1983) Constraints on the heating andcooling processes of chondrule formation. In Chondrules and

Their Origins (E. A. King, ed.), pp. 53–60. Lunar and Plan-etary Institute, Houston.

Galy A., Young E. D., Ash R. D., and O’Nions R. K. (2000) Theformation of chondrules at high gas pressures in the solar neb-ula. Science, 290, 1751–1753.

Georges P., Libourel G., and Dloule E. (1999) Potassium conden-sation experiments and their bearing on chondrule formation(abstract). In Lunar and Planetary Science XXX, Abstract#1744. Lunar and Planetary Institute, Houston (CD-ROM).

Gibson E. K. Jr. and Hubbard N. J. (1972) Thermal volatilizationstudies on lunar samples. Proc. Lunar Sci. Conf. 3rd, pp. 2003–2014.

Gibson E. K. Jr., Hubbard N. J., Wiesmann H., Bansal B. M., andMoore G. W. (1973) How to lose Rb, K, and change the K/Rb ratio: An experimental study. Proc. Lunar Sci. Conf. 4th,pp. 1263–1273.

Gooding J. L. and Keil K. (1981) Relative abundances of chon-drule primary textural types in ordinary chondrites and theirbearing on conditions of chondrule formation. Meteoritics, 16,17–43.

Gooding J. L., Keil K., Fukuoka T., and Schmitt R. A. (1980)Elemental abundances in chondrules from unequilibrated chon-drites — Evidence for chondrule origin by melting of pre-existing materials. Earth Planet. Sci. Lett., 50, 171–180.

Graham A. L., Easton A. J., Hutchison R., and Jerome D. Y.(1976) Bovedy meteorite — Mineral chemistry and origin of itsCa-rich glass inclusions. Geochim. Cosmochim. Acta, 40, 529.

Greenwood J. P. and Hess P. C. (1996) Congruent melting kinetics:Constraints on chondrule formation. In Chondrules and theProtoplanetary Disk (R. H. Hewins et al., eds.), pp. 205–212.Cambridge Univ., New York.

Grossman J. N. (1996) Chemical fractionations of chondrites:Signatures of events before chondrule formation. In Chon-drules and the Protoplanetary Disk (R. H. Hewins et al., eds.),pp. 243–256. Cambridge Univ., New York.

Grossman J. N. and Wasson J. T. (1982) Evidence for primitivenebular components in chondrules from the Chainpur chon-drite. Geochim. Cosmochim. Acta, 46, 1081–1099.

Grossman J. N. and Wasson J. T. (1983a) Refractory precursorcomponents of Semarkona chondrules and the fractionation ofrefractory elements among chondrites. Geochim. Cosmochim.Acta, 47, 759–771.

Grossman J. N. and Wasson J. T. (1983b) The composition ofchondrules in unequilibrated chondrites: An evaluation ofmodels for the formation of chondrules and their precursormaterials. In Chondrules and Their Origins (E. A. King, ed.),pp. 88–121. Lunar and Planetary Institute, Houston.

Grossman J. N. and Wasson J. T. (1985) The origin and historyof the metal and sulfide components of chondrules. Geochim.Cosmochim. Acta, 49, 925–939.

Grossman J. N., Rubin A. E., Rambaldi E. R., Rajan R. S., andWasson J. T. (1985) Chondrules in the Qingzhen type-3 ensta-tite chondrite — Possible precursor components and compari-son to ordinary chondrite chondrules. Geochim. Cosmochim.Acta, 49, 1781–1795.

Grossman J. N., Rubin A. E., Nagahara H., and King E. A. (1988)Properties of chondrules. In Meteorites and the Early SolarSystem (J. F. Kerridge and M. S. Matthews), pp. 619–659.Univ. of Arizona, Tucson.

Grossman J. N., Alexander C. M. O’D., Wang J., and BrearleyA. J. (2002) Zoned chondrules in Semarkona: Evidence forhigh- and low-temperature processing. Meteoritics & Planet.Sci., 37, 49–74.

Guan Y., McKeegan, K. D., MacPherson G. J. (2000) Oxygen


isotopes in calcium-aluminum-rich inclusions from enstatitechondrites: New evidence for a single CAI source in the solarnebula. Earth Planet. Sci. Lett., 181, 271–277.

Hamilton D. I. and MacKenzie W. S. (1965) Phase equilibriumstudies in the system NaAlSiO2-KAlSiO4-SiO2-H2O. Mineral.Mag., 34, 214–231.

Hanon P., Robert F., and Chaussidon M. (1998) High carbon con-centrations in meteoritic chondrules: A record of metal-sili-cate differentiation. Geochim. Cosmochim. Acta, 62, 903–913.

Hashimoto A. (1983) Evaporation metamorphism in the early solarnebula — evaporation experiments on the melt FeO-MgO-SiO2-CaO-Al2O3. Geochem. J., 17, 111–145.

Hashimoto A., Kumazawa M., and Onuma N. (1979) Evapora-tion metamorphism of primitive dust material in the early solarnebula. Earth Planet. Sci. Lett., 43, 13–21.

Herzberg C. T. (1979) Solubility of olivine in basaltic liquids —Ionic model. Geochim. Cosmochim. Acta, 43, 1241–1251.

Hewins R. H. (1991) Retention of sodium during chondrule melt-ing. Geochim. Cosmochim. Acta, 55, 935–942.

Hewins R. H. (1997) Chondrules. Annu. Rev. Earth Planet. Sci.,25, 61–83.

Hewins R. H. and Fox G. E. (2004) Chondrule textures and pre-cursor grain size: An experimental study. Geochim. Cosmo-chim. Acta, 68, 917–926.

Hewins R. H. and Radomsky P. M. (1990) Temperature condi-tions for chondrule formation. Meteoritics, 25, 309–318.

Hewins R. H., Yu Y., Zanda B., and Bourot-Denis M. (1997) Donebular fractionations, evaporative losses, or both, influencechondrule compositions? Antarct. Meteorite Res., 10, 275–298.

Holmen B. A. and Wood J. A. (1986) Chondrules that indent oneanother: Evidence for hot accretion? (abstract). Meteoritics,21, 399.

Huang S., Benoit P. H., and Sears D. W. G. (1993) Metal andsulfide in Semarkona chondrules and rims: Evidence for re-duction, evaporation, and recondensation during chondruleformation. Meteoritics, 28, 367–368.

Huang S. X., Lu J., Prinz M., Weisberg M. K., Benoit P. H., andSears D. W. G. (1996) Chondrules: Their diversity and the roleof open-system processes during their formation. Icarus, 122,316–346.

Humayun M. and Clayton R. N. (1995) Potassium isotope cos-mochemistry — Genetic-implications of volatile element deple-tion. Geochim. Cosmochim. Acta, 59, 2131–2148.

Humayun M., Campbell A. J., Zanda B., and Bourot-Denise M.(2002) Formation of Renazzo chondrule metal inferred fromsiderophile elements (abstract). In Lunar and Planetary Sci-ence XXXIII, Abstract #1965. Lunar and Planetary Institute,Houston (CD-ROM).

Huss G. R. and Lewis R. S. (1995) Presolar diamond, SiC, andgraphite in primitive chondrites: Abundances as a function ofmeteorite class and petrologic type. Geochim. Cosmochim.Acta, 59, 115–160.

Huss G. R., Keil K., and Taylor G. J. (1981) The matrices of un-equilibrated ordinary chondrites: Implications for the originand history of chondrites. Geochim. Cosmochim. Acta, 45, 33–51.

Hutcheon I. D. and Hutchison R. (1989) Evidence from theSemarkona ordinary chondrite for Al-26 heating of small plan-ets. Nature, 337, 238–241.

Hutchison R. and Bevan A. W. R. (1983) Conditions and time ofchondrule accretion. In Chondrules and Their Origins (E. A.King), pp. 162–179. Lunar and Planetary Institute, Houston.

Hutchison R. and Graham A. L. (1975) Significance of calcium-rich differentiates in chondritic meteorites. Nature, 255, 471.

Hutchison R., Alexander C. M. O., and Barber D. J. (1988) Chon-drules — Chemical, mineralogical and isotopic constraints ontheories of their origin. Philos. Trans. R. Soc. Lond., A325,445–458.

Ikeda Y. (1982) Petrology of the ALH-77003 chondrite (C3). Proc.7th Symp. Antarct. Meteorites, 34–65.

Itoh S. and Yurimoto H. (2003) Contemporaneous formation ofchondrules and refractory inclusions in the early solar system.Nature, 423, 728–731.

Johnson C. A. and Prinz M. (1991) Chromite and olivine in type-IIchondrules in carbonaceous and ordinary chondrites — Impli-cations for thermal histories and group-differences. Geochim.Cosmochim. Acta, 55, 893–904.

Jones R. H. (1990) Petrology and mineralogy of type-II, FeO-richchondrules in Semarkona (LL3.0) — Origin by closed-systemfractional crystallization, with evidence for supercooling. Geo-chim. Cosmochim. Acta, 54, 1785–1802.

Jones R. H. (1992) On the relationship between isolated and chon-drule olivine grains in the carbonaceous chondrite ALHA77307. Geochim. Cosmochim. Acta, 56, 467–482.

Jones R. H. (1994) Petrology of FeO-poor, porphyritic pyroxenechondrules in the Semarkona chondrite. Geochim. Cosmochim.Acta, 58, 5325–5340.

Jones R. H. (1996) FeO-rich, porphyritic pyroxene chondrules inunequilibrated ordinary chondrites. Geochim. Cosmochim.Acta, 60, 3115–3138.

Jones R. H. and Danielson L. R. (1997) A chondrule origin fordusty relict olivine in unequilibrated chondrites. Meteoritics& Planet. Sci., 32, 753–760.

Jones R. H. and Lofgren G. E. (1993) A comparison of FeO-rich,porphyritic olivine chondrules in unequilibrated chondrites andexperimental analogs. Meteoritics, 28, 213–221.

Jones R. H. and Scott E. R. D. (1989) Petrology and thermal his-tory of type-IA chondrules in the Semarkona (LL3.0) chon-drite. Proc. Lunar Planet. Sci. Conf. 19th, pp. 523–536.

Jones R. H., Lee T., Connolly H. C., Love S. G., and Shang H.(2000) Formation of chondrules and CAIs: Theory vs. obser-vation. In Protostars and Planets IV (V. Mannings et al., eds.),pp. 927–962. Univ. of Arizona, Tucson.

Jones R. H., Leshin L. A., Guan Y. B., Sharp Z. D., DurakiewiczT., and Schilk A. J. (2004) Oxygen isotope heterogeneity inchondrules from the Mokoia CV3 carbonaceous chondrite.Geochim. Cosmochim. Acta, 68, 3423–3438.

Kehm K., Hauri E. H., Alexander C. M. O., and Carlson R. W.(2003) High precision iron isotope measurements of meteor-itic material by cold plasma ICP-MS. Geochim. Cosmochim.Acta, 67, 2879–2891.

Kennedy A. K., Hutchison R., Hutcheon I. D., and Agrell S. O.(1992) A unique high Mn/Fe microgabbro in the Parnallee(LL3) ordinary chondrite — Nebular mixture or planetary dif-ferentiate from a previously unrecognized planetary body.Earth Planet. Sci. Lett., 113, 191–205.

Kitamura M., Yasuda M., Watanabe S., and Morimoto N. (1983)Cooling history of pyroxene chondrules in the Yamato-74191chondrite (L3) — an electron-microscopic study. Earth Planet.Sci. Lett., 63, 189–201.

Klerner S. and Palme H. (1999) Origin of chondrules and matrixin carbonaceous chondrites (abstract). In Lunar and PlanetaryScience XXX, Abstract #1272. Lunar and Planetary Institute,Houston (CD-ROM).

Kong P. and Ebihara M. (1996) Metal phases of L chondrites:Their formation and evolution in the nebula and in the parentbody. Geochim. Cosmochim. Acta, 60, 2667–2680.

Kong P. and Ebihara M. (1997) The origin and nebular history of


the metal phase of ordinary chondrites. Geochim. Cosmochim.Acta, 61, 2317–2329.

Kong P. and Palme H. (1999) Compositional and genetic relation-ship between chondrules, chondrule rims, metal, and matrix inthe Renazzo chondrite. Geochim. Cosmochim. Acta, 63, 3673–3682.

Kong P., Ebihara M., and Palme H. (1999) Distribution of sidero-phile elements in CR chondrites: Evidence for evaporation andrecondensation during chondrule formation. Geochim. Cosmo-chim. Acta, 63, 2637–2652.

Krot A. N. and Wasson J. T. (1995) Igneous rims on low-FeOchondrules in ordinary chondrites. Geochim. Cosmochim. Acta,59, 4951–4966.

Krot A. N., McKeegan K. D., Russell S. S., Meibom A., WeisbergM. K., Zipfel J., Krot T. V., Fagan T. J., and Keil K. (2001a)Refractory calcium-aluminum-rich inclusions and aluminum-diopside-rich chondrules in the metal-rich chondrites Hamma-dah al Hamra 237 and Queen Alexandra Range 94411. Mete-oritics & Planet. Sci., 36, 1189–1216.

Krot A. N., Meibom A., Russell S. S., Alexander C. M. O., JeffriesT. E., and Keil K. (2001b) A new astrophysical setting forchondrule formation. Science, 291, 1776–1779.

Kunihiro T., Rubin A. E., McKeegan K. D., and Wasson J. T.(2004) Oxygen-isotopic compositions of relict and host grainsin chondrules in the Yamato 81020 CO3.0 chondrite. Geochim.Cosmochim. Acta, 68, 3599–3606.

Kurat G., Pernicka E., and Herrwerth I. (1984) Chondrules fromChainpur (LL-3): Reduced parent rocks and vapor fractiona-tion. Earth Planet. Sci. Lett., 68, 43–56.

La Blue A. R. and Lauretta D. S. (2004) Metallic chondrules inNWA 1390 (H3-6): Clues to their history from metallic Cu(abstract). In Lunar and Planetary Science XXXV, Abstract#1939. Lunar and Planetary Institute, Houston (CD-ROM).

Larimer J. W. and Anders E. (1967) Chemical fractionation in me-teorites: II. Abundance patterns and their interpretation. Geo-chim. Cosmochim. Acta, 31, 1239–1270.

Lasaga T. C. (1998) Kinetic Theory in the Earth Sciences. Prince-ton Univ., Princeton. 811 pp.

Lauretta D. S. and Buseck P. R. (2003) Opaque minerals in chon-drules and fine-grained chondrule rims in the Bishunpur (LL3.1)chondrite. Meteoritics & Planet. Sci., 38, 59–79.

Lauretta D. S. and Killgore M. (2005) A Color Atlas of Meteor-ites in Thin Section. Golden Retriever Publications/SouthwestMeteorite Press.

Lauretta D. S., Kremser D. T., and Fegley B. (1996) The rate ofiron sulfide formation in the solar nebula. Icarus, 122, 288–315.

Lauretta D. S., Lodders K., and Fegley B. (1997) Experimentalsimulations of sulfide formation in the solar nebula. Science,277, 358–360.

Lauretta D. S., Lodders K., and Fegley B. (1998) Kamacite sul-furization in the solar nebula. Meteoritics & Planet. Sci., 33,821–833.

Lauretta D. S., Hua X., and Buseck P. R. (2000) Mineralogy offine-grained rims in the ALH81002 CM chondrite. Geochim.Cosmochim. Acta, 64, 3263–3273.

Lauretta D. S., Buseck P. R., and Zega T. J. (2001) Opaque min-erals in the matrix of the Bishunpur (LL3.1) chondrite: Con-straints on the chondrule formation environment. Geochim.Cosmochim. Acta, 65, 1337–1353.

Lee M. S., Rubin A. E., and Wasson J. T. (1992) Origin of metallicFe-Ni in Renazzo and related chondrites. Geochim. Cosmo-chim. Acta, 56, 2521–2533.

Lemelle L., Guyot F., Fialin M., and Pargamin J. (2000) Experi-mental study of chemical coupling between reduction and vola-tilization in olivine single crystals. Geochim. Cosmochim. Acta,64, 3237–3249.

Lemelle L., Guyot F., Leroux H., and Libourel G. (2001) Anexperimental study of the external reduction of olivine singlecrystals. Am. Mineral., 86, 47–54.

Leroux H., Libourel G., Lemelle L., and Guyot F. (2003) Experi-mental study and TEM characterization of dusty olivines inchondrites: Evidence for formation by in situ reduction. Mete-oritics & Planet. Sci., 38, 81–94.

Lewis R. D., Lofgren G. E., Franzen H. F., and Windom K. E.(1993) The effect of Na vapor on the Na content of chondrules.Meteoritics, 28, 622–628.

Libourel G., Krot A. N., and Tissandier L. (2003) Evidence forhigh temperature condensation of moderately-volatile elementsduring chondrule formation (abstract). In Lunar and PlanetaryScience XXXIV, Abstract #1558. Lunar and Planetary Institute,Houston (CD-ROM).

Lodders K. (2003) Solar system abundances and condensationtemperatures of the elements. Astrophys. J., 591, 1220–1247.

Lofgren G. (1989) Dynamic crystallization of chondrule melts ofporphyritic olivine composition — Textures experimental andnatural. Geochim. Cosmochim. Acta, 53, 461–470.

Lofgren G. E. (1996) A dynamic crystallization model for chon-drule melts. In Chondrules and the Protoplanetary Disk (R. H.Hewins et al., eds.), pp. 187–196. Cambridge Univ., New York.

Lofgren G. and Lanier A. B. (1990) Dynamic crystallization studyof barred olivine chondrules. Geochim. Cosmochim. Acta, 54,3537–3551.

Lofgren G. and Russell W. J. (1986) Dynamic crystallization ofchondrule melts of porphyritic and radial pyroxene composi-tion. Geochim. Cosmochim. Acta, 50, 1715–1726.

Luck J. M., Ben Othman D., Barrat J. A., and Albarede F. (2003)Coupled Cu-63 and O-16 excesses in chondrites. Geochim.Cosmochim. Acta, 67, 143–151.

Lyons J. R. and Young E. D. (2003) Towards an evaluation of self-shielding at the x-point as the origin of the oxygen isotopeanomaly in CAIs (abstract). In Lunar and Planetary ScienceXXXIV, Abstract #1981. Lunar and Planetary Institute, Houston(CD-ROM).

Maharaj S. V. and Hewins R. H. (1994) Clues to chondrule pre-cursors — An investigation of vesicle formation in experimen-tal chondrules. Geochim. Cosmochim. Acta, 58, 1335–1342.

Maharaj S. V. and Hewins R. H. (1998) Chondrule precursorminerals as anhydrous phases. Meteoritics & Planet. Sci., 33,881–887.

Marshall J. and Cuzzi J. N. (2001) Electrostatic enhancement ofcoagulation in protoplanetary nebulae (abstract). In Lunar andPlanetary Science XXXII, Abstract #1262. Lunar and PlanetaryInstitute, Houston (CD-ROM).

Matsunami S. (1984) Chemical compositions and textures ofmatrices and chondrule rims in eight unequilibrated ordinarychondrites: A preliminary report. Mem. Natl. Inst. Polar Res.,Spec. Issue 24, 126–148.

Matsunami S., Ninagawa K., Nishimura S., Kubono N., YamamotoI., Kohata M., Wada T., Yamashita Y., Lu J., Sears D. W. G.,and Nishimura H. (1993) Thermoluminescence and composi-tional zoning in the mesostasis of a Semarkona group A1 chon-drule and new insights into the chondrule-forming process.Geochim. Cosmochim. Acta, 57, 2101–2110.

McEwing C. E., Thode H. G., and Rees C. E. (1980) Sulphurisotope effects in the dissociation and evaporation of troilite:


A possible mechanism for 34S enrichment in lunar soils. Geo-chim. Cosmochim. Acta, 44, 565–571.

McSween H. Y. (1977) Chemical and petrographic constraints onorigin of chondrules and inclusions in carbonaceous chondrites.Geochim. Cosmochim. Acta, 41, 1843–1860.

Meibom A., Petaev M. I., Krot A. N., Wood J. A., and Keil K.(1999) Primitive FeNi metal grains in CH carbonaceous chon-drites formed by condensation from a gas of solar composition.J. Geophys. Res.–Planets, 104, 22053–22059.

Metzler K., Bischoff A., and Stoffler D. (1992) Accretionary dustmantles in CM chondrites: Evidence for solar nebula proc-esses. Geochim. Cosmochim. Acta, 56, 1873–1898.

Misawa K. and Fujita T. (1994) A relict refractory inclusion in aferromagnesian chondrule from the Allende meteorite. Nature,368, 723–726.

Misawa K. and Nakamura N. (1988a) Demonstration of REE frac-tionation among individual chondrules from the Allende (CV3)chondrite. Geochim. Cosmochim. Acta, 52, 1699–1710.

Misawa K. and Nakamura N. (1988b) Highly fractionated rare-earth elements in ferromagnesian chondrules from the Felix(CO3) meteorite. Nature, 334, 47–50.

Misawa K. and Nakamura N. (1996) Origin of refractory precur-sor components of chondrules from carbonaceous chondrites.In Chondrules and the Protoplanetary Disk (R. H. Hewins etal., eds.), pp. 99–106. Cambridge Univ., New York.

Mostefaoui S. and Perron C. (1994) Redox processes in chon-drules recorded in metal. Meteoritics, 29, 506–506.

Mostefaoui S., Perron C., Zinner E., and Sagon G. (2000) Metal-associated carbon in primitive chondrites: Structure, isotopiccomposition, and origin. Geochim. Cosmochim. Acta, 64,1945–1964.

Mullane E., Russell S. S., Gounelle M., and Mason T. F. D. (2003)Iron isotope composition of Allende and Chainpur chondrules:Effects of equilibration and thermal history (abstract). In Lu-nar and Planetary Science XXXIV, Abstract #1027. Lunar andPlanetary Institute, Houston (CD-ROM).

Muller W. F., Weinbruch S., Walter R., and Mullerbeneke G.(1995) Transmission electron-microscopy of chondrule miner-als in the Allende meteorite — Constraints on the thermal anddeformational history of granular olivine-pyroxene chondrules.Planet. Space Sci., 43, 469–483.

Nagahara H. (1981) Evidence for secondary origin of chondrules.Nature, 292, 135–136.

Nagahara H. (1983) Chondrules formed through incomplete melt-ing of the pre-existing minerals and the origin of chondrules.In Chondrules and Their Origins (E. A. King, ed.), pp. 211–222. Lunar and Planetary Institute, Houston.

Nagahara H. (1984) Matrices of type-3 ordinary chondrites —Primitive nebular records. Geochim. Cosmochim. Acta, 48,2581–2595.

Nagahara H. (1996) Evaporation induced isothermal crystalliza-tion by evaporation. In Lunar and Planetary Science XXVIII,pp. 927–928. Lunar and Planetary Institute, Houston.

Nagahara H. (2000) Chemical fractionation in the presence of am-bient gas. Meteoritics & Planet. Sci., 35, A116.

Nagahara H. and Ozawa K. (1995) Stability of silicate melt in thesolar nebula. Symp. Antarct. Meteorites, 20, 172–173.

Nagahara H. and Ozawa K. (1996a) Evaporation kinetics of sili-cate melt (abstract). Meteoritics & Planet. Sci., 31, A95–A96.

Nagahara H. and Ozawa K. (1996b) Evaporation of silicate meltin the sytem Mg2SiO4-SiO2. Symp. Antarct. Meteorites, 21,125–27.

Nagahara H. and Ozawa K. (1996c) Evaporation of forsterite in H2

gas. Geochim. Cosmochim. Acta, 60, 1445–1459.Nagahara H. and Ozawa K. (2000) Isotopic fractionation as a

probe of heating processes in the solar nebula. Chem. Geol.,169, 45–68.

Nagahara H. and Ozawa K. (2003) Was the diversity of chondrulecompositions achieved by evaporation and condensation proc-esses? (abstract). Meteoritics & Planet. Sci., 38, A125.

Nelson L. S., Skaggs S. R., Keil K., and Blander M. (1972) Use ofa CO2-laser to prepare chondrule-like spherules from super-cooled molten oxide and silicate droplets. Earth Planet. Sci.Lett., 14, 338.

Niemeyer S. (1988) Titanium isotopic anomalies in chondrulesfrom carbonaceous chondrites. Geochim. Cosmochim. Acta, 52,309–318.

Notsu K., Onuma N., and Nishida N. (1978) High temperatureheating of the Allende meteorite. Geochim. Cosmochim. Acta,42, 903–907.

Ozawa K. and Nagahara H. (1997) Stability of chondrule melt inthe solar nebula (abstract). In Lunar and Planetary ScienceXXVIII, pp. 1055–1056. Lunar and Planetary Institute, Houston.

Ozawa K. and Nagahara H. (1998) Chemical and isotopic frac-tionation during evaporation of a multi-component system:(2) General model and application to Mg-Fe olivine. Antarct.Meteorites XXIII, 126–127.

Ozawa K. and Nagahara H. (2001) Chemical and isotopic frac-tionations by evaporation and their cosmochemical implica-tions. Geochim. Cosmochim. Acta, 65, 2171–2199.

Pack A., Shelley J. M. G., and Palme H. (2004) Chondrules withpeculiar REE patterns: Implications for solar nebular conden-sation at high C/O. Science, 303, 997–1000.

Paque J. M. and Cuzzi J. N. (1997) Physical characteristics ofchondrules and rims and aerodynamic sorting in the solarnebula (abstract). In Lunar and Planetary Science XXVIII,pp. 1071–1072. Lunar and Planetary Institute, Houston.

Patchett P. J. (1980) Sr isotopic fractionation in Allende chon-drules — A reflection of solar nebular processes. Earth Planet.Sci. Lett., 50, 181–188.

Petaev M. I., Meibom A., Krot A. N., Wood J. A., and Keil K.(2001) The condensation origin of zoned metal grains in QUE94411: Implications for the formation of the Bencubbin-likechondrites. Meteoritics & Planet. Sci., 36, 93–107.

Radomsky P. M. and Hewins R. H. (1990) Formation conditionsof pyroxene-olivine and magnesian olivine chondrules. Geo-chim. Cosmochim. Acta, 54, 3475–3490.

Rambaldi E. R. (1981) Relict grains in chondrules. Nature, 293,558–561.

Rambaldi E. R. and Wasson J. T. (1981) Metal and associatedphases in Bishunpur, a highly unequilibrated ordinary chon-drite. Geochim. Cosmochim. Acta, 45, 1001–1015.

Rambaldi E. R. and Wasson J. T. (1982) Fine, nickel-poor Fe-Nigrains in the olivine of unequilibrated ordinary chondrites.Geochim. Cosmochim. Acta, 46, 929–939.

Rambaldi E. R. and Wasson J. T. (1984) Metal and associatedphases in Krymka and Chainpur: Nebular formational proc-esses. Geochim. Cosmochim. Acta, 48, 1885–1897.

Rambaldi E. R., Rajan R. S., Wang D., and Housley R. M. (1983)Evidence for relict grains in chondrules of Qingzhen, an E3-type enstatite chondrite. Earth Planet. Sci. Lett., 66, 11–24.

Richter F. M. and Davis A. M. (1997) Mass loss kinetics and iso-tope fractionation of Type B CAI melts at 1400°C and solar fO 2

(abstract). In Lunar and Planetary Science XXVIII, pp. 1161–1162. Lunar and Planetary Institute, Houston.

Richter F. M., Parsad N., Davis A. M., and Hasegawa H. (1999)


CAI cosmobarometry (abstract). In Lunar and Planetary Sci-ence XXX, Abstract #1989. Lunar and Planetary Institute,Houston (CD-ROM).

Richter F. M., Davis A. M., Ebel D. S., and Hashimoto A. (2002)Elemental and isotopic fractionation of Type B calcium-alu-minum-rich inclusions: Experiments, theoretical considera-tions, and constraints on their thermal evolution. Geochim.Cosmochim. Acta, 66, 521–540.

Rubin A. E. (1984) Coarse-grained chondrule rims in type-3 chon-drites. Geochim. Cosmochim. Acta, 48, 1779–1789.

Rubin A. E. and Wasson J. T. (1987) Chondrules, matrix andcoarse-grained chondrule rims in the Allende meteorite —Origin, interrelationships and possible precursor components.Geochim. Cosmochim. Acta, 51, 1923–1937.

Rubin A. E. and Wasson J. T. (1988) Chondrules and matrix inthe Ornans CO3 meteorite — Possible precursor components.Geochim. Cosmochim. Acta, 52, 425–432.

Rubin A. E. and Wasson J. T. (2005) Non-spherical lobate chon-drules in CO3.0 Y-81020: General implications for the forma-tion of low-FeO porphyritic chondrules in CO chondrites. Geo-chim. Cosmochim. Acta, 69, 211–220.

Scott E. R. D. and Taylor G. J. (1983) Chondrules and other com-ponents in C-chondrite, O-chondrite, and E-chondrite — Simi-larities in their properties and origins. Proc. Lunar Planet. Sci.Conf. 14th, in J. Geophys. Res., 88, B275–B286.

Scott E. R. D., Rubin A. E., Taylor G. J., and Keil K. (1984)Matrix material in type-3 chondrites — Occurrence, heteroge-neity and relationship with chondrules. Geochim. Cosmochim.Acta, 48, 1741–1757.

Scott E. R. D., Barber D. J., Alexander C. M. O’D., HutchisonR., and Peck J. A. (1988) Primitive material surviving in chon-drites: Matrix. In Meteorites and the Early Solar System (J. F.Kerridge and M. S. Matthews, eds.), pp. 718–745. Univ. of Ari-zona, Tucson.

Sears D. W. G., Lu J., Benoit P. H., Dehart J. M., and LofgrenG. E. (1992) A compositional classification scheme for mete-oritic chondrules. Nature, 357, 207–210.

Sears D. W. G., Huang S., and Benoit P. H. (1996) Open-systembehavior during chondrule formation. In Chondrules and theProtoplanetary Disk (R. H. Hewins et al., eds.), pp. 221–232.Cambridge Univ., New York.

Shimaoka T. and Nakamura N. (1989) Vaporization of sodiumfrom a partially molten chondritic material. Proc. NIPR Symp.Antarct. Meteorites, 2, 252–267.

Shimaoka T. and Nakamura N. (1991) The effect of total pres-sure on vaporization of alkalis from partially molten chondriticmaterial. In Origin and Evolution of Interplanetary Dust (A. C.Levasseur-Regourd and H. Hasegawa, eds.), pp. 79–82. Klu-wer, Dordrecht.

Shirai T., Tachibana S., and Tsuchiyama A. (2000) Evaporationrates of Na from Na2O-SiO2 melt at 1 atm (abstract). In Lunarand Planetary Science XXXI, Abstract #1610. Lunar and Plane-tary Institute, Houston (CD-ROM).

Steele I. M. (1986) Compositions and textures of relic forsterite incarbonaceous and unequilibrated ordinary chondrites. Geo-chim. Cosmochim. Acta, 50, 1379–1395.

Swindle T. D., Caffee M. W., Hohenberg C. M., Lindstrom M. M.,and Taylor G. J. (1991a) Iodine-xenon studies of petrographi-cally and chemically characterized Chainpur chondrules. Geo-chim. Cosmochim. Acta, 55, 861–880.

Swindle T. D., Grossman J. N., Olinger C. T., and Garrison D. H.(1991b) Iodine-xenon, chemical, and petrographic studies ofSemarkona chondrules — Evidence for the timing of aqueousalteration. Geochim. Cosmochim. Acta, 55, 3723–3734.

Tachibana S. and Huss G. R. (2005) Sulfur isotope compositionof putative primary troilite in chondrules from Bishunpur andSemarkona. Geochim. Cosmochim. Acta, 69, 3075–3097.

Tachibana S., Nagahara H., Mostefaoui S., and Kita N. T. (2003)Correlation between relative ages inferred from 26Al and bulkcompositions of ferromagnesian chondrules in least equili-brated ordinary chondrites. Meteoritics & Planet. Sci., 38, 939–962.

Thiemens M. H. (1996) Mass-independent isotopic effects inchondrites: The role of chemical processes. In Chondrules andthe Protoplanetary Disk (R. H. Hewins et al., eds.), pp. 107–118. Cambridge Univ., New York.

Tissandier L., Libourel G., Robert F., and Chaussidon M. (2000)SiO2 condensation experiments and implications for protosolarmaterials (abstract). In Lunar and Planetary Science XXXI,Abstract #1553. Lunar and Planetary Institute, Houston (CD-ROM).

Tissandier L., Libourel G., and Robert F. (2002) Gas-melt inter-actions and their bearing on chondrule formation. Meteoritics& Planet. Sci., 37, 1377–1389.

Tsuchiyama A. and Tachibana S. (1999) Evaporation rates of el-ements, such as sodium, from melts under solar nebula condi-tions (abstract). In Lunar and Planetary Science XXX, Abstract#1535. Lunar and Planetary Institute, Houston (CD-ROM).

Tsuchiyama A., Nagahara H., and Kushiro I. (1981) Volatiliza-tion of sodium from silicate melt spheres and its applicationto the formation of chondrules. Geochim. Cosmochim. Acta,45, 1357–1367.

Tsuchiyama A., Osada Y., Nakano T., and Uesugi K. (2004) Ex-perimental reproduction of classic barred olivine chondrules:Open-system behavior of chondrule formation. Geochim. Cos-mochim. Acta, 68, 653–672.

Varela M. E. and Metrich N. (2000) Carbon in olivines of chon-dritic meteorites. Geochim. Cosmochim. Acta, 64, 3433–3438.

Wang J., Davis A. M., and Clayton R. N. (1994) Modeling ofisotopic fractionation during the phase transition of a diffusioncontrolled reservoir. In Eighth International Conference onGeochronology, Cosmochronology, and Isotope Geology. U.S.Geological Survey Circular 1107.

Wang J., Yu Y., Alexander C. M. O.’D., and Hewins R. H. (1999)The influence of oxygen and hydrogen on the evaporation ofK (abstract). In Lunar and Planetary Science XXX, Abstract#1778. Lunar and Planetary Institute, Houston (CD-ROM).

Wang J., Davis A. M., Clayton R. N., Mayeda T. K., and Hashi-moto A. (2001) Chemical and isotopic fractionation during theevaporation of the FeO-MgO-SiO2-CaO-Al2O3-TiO2 rare earthelement melt system. Geochim. Cosmochim. Acta, 65, 479–494.

Wasson J. T. (1993) Constraints on chondrule formation. Mete-oritics, 28, 13–28.

Watanabe S., Kitamura M., and Morimoto N. (1985) A transmis-sion electron-microscope study of pyroxene chondrules inequilibrated L-group chondrites. Earth Planet. Sci. Lett., 72,87–98.

Weinbruch S. and Muller W. F. (1995) Constraints on the coolingrates of chondrules from the microstructure of clinopyroxeneand plagioclase. Geochim. Cosmochim. Acta, 59, 3221–3230.

Weinbruch S., Buttner H., Holzheid A., Rosenhauer M., andHewins R. H. (1998) On the lower limit of chondrule coolingrates: The significance of iron loss in dynamic crystallizationexperiments. Meteoritics & Planet. Sci., 33, 65–74.

Weinbruch S., Palme H., and Spettel B. (2000) Refractoryforsterite in primitive meteorites: Condensates from the solarnebula? Meteoritics & Planet. Sci., 35, 161–171.


Weinbruch S., Muller W. F., and Hewins R. H. (2001) A trans-mission electron microscope study of exsolution and coarsen-ing in iron-bearing clinopyroxene from synthetic analogues ofchondrules. Meteoritics & Planet. Sci., 36, 1237–1248.

Weisberg M. K. and Prinz M. (1996) Agglomeratic chondrules,chondrule precursors, and incomplete melting. In Chondrulesand the Protoplanetary Disk (R. H. Hewins et al., eds.),pp. 119–128. Cambridge Univ., New York.

Wood J. A. (1981) The interstellar dust as a precursor of Ca,Al-rich inclusions in carbonaceous chondrites. Earth Planet. Sci.Lett., 56, 32–44.

Wood J. A. and Hashimoto A. (1993) Mineral equilibrium in frac-tionated nebular systems. Geochim. Cosmochim. Acta, 57,2377–2388.

Yoneda S. and Grossman L. (1995) Condensation of CaO-MgO-Al2O3-SiO2 liquids from cosmic gases. Geochim. Cosmochim.Acta, 59, 3413–3444.

Young E. D. (2000) Assessing the implications of K isotope cos-mochemistry for evaporation in the preplanetary solar nebula.Earth Planet. Sci. Lett., 183, 321–333.

Young E. D., Ash R. D., Galy A., and Belshaw N. S. (2002) Mgisotope heterogeneity in the Allende meteorite measured by UVlaser ablation-MC-ICPMS and comparisons with O isotopes.Geochim. Cosmochim. Acta, 66, 683–698.

Yu Y. and Hewins R. H. (1998) Transient heating and chondruleformation: Evidence from sodium loss in flash heating simu-lation experiments. Geochim. Cosmochim. Acta, 62, 159–172.

Yu Y., Hewins R. H., Clayton R. N., and Mayeda T. K. (1995)Experimental study of high-temperature oxygen-isotope ex-change during chondrule formation. Geochim. Cosmochim.Acta, 59, 2095–2104.

Yu Y., Hewins R. H., and Zanda B. (1996) Sodium and sulfur inchondrules: Heating time and cooling curves. In Chondrulesand the Protoplanetary Disk (R. H. Hewins et al., eds.),pp. 213–220. Cambridge Univ., New York.

Yu Y., Hewins R. H., Alexander C. M. O.’D., and Wang J. (2003)Experimental study of evaporation and isotopic mass fraction-ation of potassium in silicate melts. Geochim. Cosmochim.Acta, 67, 773–786.

Zanda B., Bourot-Denise M., Perron C., and Hewins R. H. (1994)Origin and metamorphic redistribution of silicon, chromium,and phosphorus in the metal of chondrites. Science, 265, 1846–1849.

Zanda B., Bourot-Denise M., Hewins R. H., Cohen B. A., andDelaney J. (2002) Accretion textures, iron evaporation and re-condensation in Renazzo chondrules (abstract). In Lunar andPlanetary Science XXXIII, Abstract #1852. Lunar and Plane-tary Institute, Houston (CD-ROM).

Zega T. J. and Buseck P. R. (2003) Fine-grained-rim mineralogyof the Cold Bokkeveld CM chondrite. Geochim. Cosmochim.Acta, 67, 1711–1721.

Zhu X. K., Guo Y., O’Nions R. K., Young E. D., and Ash R. D.(2001) Isotopic homogeneity of iron in the early solar nebula.Nature, 412, 311–313.

Zinner E. (1998) Stellar nucleosynthesis and the isotopic compo-sition of presolar grains from primitive meteorites. Annu. Rev.Earth Planet. Sci., 26, 147–188.

petrology and origin of ferromagnesian silicate … et al.: petrology and origin of ferromagnesian...

Documents