Author's personal copy - UCLAcosmochemists.igpp.ucla.edu/2010-1.pdf · Author's personal copy Pyroxene-selective impact smelting in ureilites ... Received 1 March 2010; accepted in

Author's personal copy

Pyroxene-selective impact smelting in ureilites

Paul H. Warren *, Alan E. Rubin

Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA 90095-1567, USA

Received 1 March 2010; accepted in revised form 18 May 2010; available online 1 June 2010

Abstract

A characteristic feature of ureilite meteorites is reduction of FeO. But the reduction is usually confined to the rims of oliv-ine. In the LAR 04315, LAP 03587 and Almahata Sitta ureilites, pyroxene was extensively reduced by impact smelting. InLAR 04315, the impact caused nearly all of the original pigeonite to melt or otherwise become sufficiently structurally com-promised to allow smelting, and yet a minor proportion of the pyroxene escaped smelting and survived with its original com-position (En74.1Wo10.2). Olivine mosaicism confirms that LAR 04315 experienced a major shock event. The smelted pyroxenesalso show a distinctive patchiness in their interference colors (although each grain’s basic optical continuity, often includingtwinning, is still discernible). They also have reduced compositions, are ubiquitously porous (�15%), and contain sprinklingsof Fe-metal and felsic glass. For the most part the olivine underwent only very slight reduction. Much of the (small) pyroxenecomponent of LAP 03587 shows the same oddly porous texture. LAR 04315 also contains large traces of silica and felsic glass(with a typical composition of, in wt%, 61 SiO2, 23 Al2O3, 11 CaO, 3.7 Na2O) glass; these two phases together form selvagesthat line the walls of many of the largest voids in the rock. Silica is a by-product of pyroxene smelting. The felsic glass prob-ably derives largely from interstitial basaltic melt that predated the impact. However, the comparatively stiff surrounding/included silica may have promoted unusually high melt retention within LAR 04315 through the smelting episode (one aspectof which was a major stream-out, through the same large voids, of COx gas). The impact-smelted pyroxene of LAP 03587 isenigmatic because this ureilite also features little-shocked euhedral graphite laths and no olivine mosaicism. The fine-grainedureilitic component of Almahata Sitta appears to have likewise formed by impact smelting, but with more extensive melting ofpyroxene (especially a Ca-rich pyroxene component), more pulverization and melting of olivine, and more displacement ofboth. However, in places the original coarse-equant ureilite texture is still discernible in relict form. Ordinarily, an impactshock melts olivine before, or at least no later than, pyroxene. But in the case of LAR 04315 and LAP 03587, the great shockevent evidently occurred when the material was already anatectic or very nearly so; and thus the difference in melting temper-ature between pyroxene and olivine, �300 degrees lower for pyroxene, was decisive. If literature inferences of extremely fastcooling rates, implying shallow burial depths, are accurate, the proportion of COx gas generated by ureilite smelting exceededby a very large factor (of order 103 but possibly much greater) the volume represented as porosity in the final ureilites. Theoutflow of so much gas may have, by near-surface explosive expansion and jetting, enhanced the thoroughness of the impact-triggered catastrophic impact disruption of the parent asteroid.� 2010 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Ureilites are extremely depleted asteroidal-mantle peri-dotites, typically containing about 2/3 olivine, 25% pyrox-

ene (mainly pigeonite), and generally no observablefeldspar. As reviewed by Goodrich et al. (2004), the numberof ureilites has reached well over 100, qualifying this dis-tinctive group as the second most numerous variety ofachondrite (only HEDs are more common). Ureilites showconsiderable diversity in terms of their modal proportionsof olivine, pigeonite, orthopyroxene and augite, and inthe compositions of these silicates (e.g., olivine ranges fromFo75 to Fo96); and yet they are remarkably consistent in

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

doi:10.1016/j.gca.2010.05.026

* Corresponding author. Tel.: +1 3108253202; fax: +13102063051.

E-mail address: [email protected] (P.H. Warren).

www.elsevier.com/locate/gca

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 74 (2010) 5109–5133


several important respects. Almost every ureilite containsan important proportion (average 3 wt%) of carbon, mostlyas either amorphous “C-matrix” or, in some of the lowest-shock ureilites, graphite. The rims of ureilite olivines showdistinctive indications of FeO reduction: tiny inclusions ofFe-metal scattered amidst olivine that is reversely zoned to-ward pure Mg-olivine. In contrast, the mafic-silicate coresin any individual monomict ureilite are remarkably uniformin composition. Nearly all ureilites exhibit a remarkablystep-wise cooling history: slow at first, but very rapid later(Miyamoto et al., 1985).

Takeda (1987) was the first to propose that ureilitesformed as asteroidal mantle restites, contrary to the once-popular view (e.g., Berkley et al., 1980; Goodrich et al.,1987) that they originated as igneous cumulates. Today, lar-gely in consideration of the great oxygen-isotopic diversityamong ureilites (Clayton and Mayeda, 1988), there is a gen-eral consensus (e.g., Warren and Kallemeyn, 1992; Scottet al., 1993; Singletary and Grove, 2003; Goodrich et al.,2004; Kita et al., 2004; Warren and Huber, 2006b) thatmost ureilites formed as mantle restites. Still, the role ofsmelting (“the chemical reduction of a metal from its oreby a process usually involving fusion” — Thrush, 1968)in ureilite petrogenesis remains controversial. Manyauthors argue that smelting affected ureilites from an earlystage of their igneous evolution, and engendered the widerange of ureilite mafic-silicate core compositions (e.g.,Singletary and Grove, 2003; Goodrich et al., 2004; Wilsonet al., 2008). A contrasting interpretation holds that smelt-ing occurred almost exclusively as a short-lived disequilib-rium process which left the silicate-grain cores essentiallyunaltered. The presumed trigger for both the abortivesmelting and the rapid cooling that limited it to the olivinerims was a large impact that disrupted the preheated, ana-tectic parent asteroid(s), with the ordinary mechanical-dis-ruptive effect of impact greatly amplified by rapidproduction, within the hot, fractured mantle, of copiousCOx gas as a by-product of smelting (Warren and Kalle-meyn, 1992; Warren and Huber, 2006b; cf. Scott et al.,1993; Goodrich et al, 2004, and more elaborately Downeset al., 2008, have argued for derivation from a single aster-oid). As noted by Warren and Huber (2006b), this processwas not in a strict sense “smelting,” because the materialswere cooling, not melting, as the reduction of the olivinerims took place.

The Larkman Nunatak (LAR) 04315 ureilite describedin this paper (and preliminarily by Warren and Rubin,2006) represents an exceptional case of extensive smelting(sensu stricto) as the result of intense shock, includingpigeonite-localized shock melting, of an otherwise normalureilite. One of this meteorite’s most distinctive features isunusual porosity within the altered pyroxenes, presumablyproduced by smelting (yielding COx gas). However, in con-trast to the limited reduction of olivine rims in other urei-lites, reduction in LAR 04315 primarily affected shock-melted pyroxene. Although impact smelting is exemplifiedby LAR 04315, we also describe similar manifestationswithin a relatively low-shock ureilite, LaPaz Icefield(LAP) 03587; and in the Almahata Sitta ureilite, whichwas tracked, as asteroid TC3, on its approach to Earth

(Jenniskens et al., 2009). Pyroxene-selective impact smeltingprobably accounts for the reported characteristics of severalother Antarctic ureilites (Saito and Takeda, 1990) and mayhave also occurred in the very recently discovered porousureilite, Jiddat al Harasis 422 (Janots et al., 2009). Im-pact-smelted ureilites afford new insight into the evolution,and especially the probable catastrophic disruption, of theureilite asteroid(s).

2. METHODS

Backscattered electron (BSE) and secondary-electronimages were acquired on a LEO 1430 SEM and a JEOLJXA-8200 electron probe. Mineral compositions weredetermined using wavelength-dispersive detectors on theJXA-8200. In general, analyses were run at an acceleratingvoltage of 15 KV, with focused beam and count durationsof 15–20 s. However, for volatilization-prone phases (alkalifeldspars and glasses) currents as low as 3 nA, beam diam-eters as wide as 3 lm, along with �0.5 times shorter counttimes, were employed. Semiquantitative analyses, a few ofwhich were never superseded by quantitative electron-probeanalyses, were obtained using an EDAX system on theSEM. In addition, bulk compositional data for three splitsderived from a single 1.27-g chip of LAR 04315 were ob-tained using the INAA technique of Warren et al. (2006).

3. EVIDENCE FOR IMPACT SMELTING OF

PYROXENE

3.1. Exemplary LAR 04315

LAR 04315 is a 1.165 kg ureilite that was recognized astexturally anomalous by McCoy (2005). We have studiedtwo thin sections, LAR 04315,20 and LAR 04315,21, each�1.1 cm2. Modal analysis (1700 points from LAR 04315,20and 2700 from LAR 04315,21) indicates that, discountingporosity, the meteorite consists of 47 vol% olivine,40 vol% pyroxene, 9.4 vol% Fe-oxides (weathering productssuch as limonite derived, judging from their consistentlyinterstitial context in relation to the silicates, from a combi-nation of Fe-metal, Fe-sulfide and possibly a carbonphase), 2.2 vol% Fe-metal, 0.95 vol% sulfides, 0.26 vol%of low-reflectance silicate phases, probably mostly felsic–si-licic glass and silica (see below), plus a trace of Cr-spinel.The coarsest Fe-metals and Fe-sulfides occur as grains(Fe-metal up to 0.25 mm long: McCoy, 2005) at intersticesand along boundaries between apparent individual silicategrains of the pre-shock peridotite.

It is difficult to reconstruct the unweathered Fe-metaland sulfide abundances, but they were surely considerablyhigher than the present abundances. Contextual relation-ships (adjacent grains) allow us to estimate that for �30%of the Fe-oxides the parent phase was probably Fe-metal,while for 13% it was probably sulfide. For the other 57%,contextual relationships provide little or no hint as to theoxide’s parent phase. But probably roughly half of the totalFe-oxide (i.e., 4.7 vol% of the rock) is weathered Fe-metal;the rest is weathered sulfide. To complete the modal recon-struction, we need to convert from vol% Fe-oxide back to

5110 P.H. Warren, A.E. Rubin / Geochimica et Cosmochimica Acta 74 (2010) 5109–5133


vol% of Fe-metal (or Fe-dominated sulfide). If we assumethat the Fe-oxides are OH-rich and somewhat porous,and thus average roughly 3000 kg m�3, then productionof these oxides would inflate the original volumes by a fac-tor of �2.6� starting from Fe-metal, or 1.6� starting fromFeS. Our reconstruction thus implies that �1.8 vol% Fe-metal and �3 vol% Fe-sulfide need to be added to themode, while 9.4 vol% Fe-oxide needs to be deleted. Afterrenormalization of all values to yield a sum of 100%, thereconstructed mode becomes 50% olivine, 42 vol% pyrox-ene, �4 vol% Fe-metal, �4 vol% sulfide, roughly0.28 vol% low-reflectance silicate phases, and a trace ofCr-spinel. Although the preterrestrial Fe-metal abundance(which figures prominently in our discussion below) isuncertain, 3 vol% seems a very safe lower limit.

The preterrestrial porosity, a distinctive feature of thisureilite (Figs. 1–5), is mainly concentrated within expansesof pyroxene (i.e., former individual pyroxene grains, in thepre-shock ureilite). Including this porosity and minor in-cluded (porosity-associated) opaques, these relict pyroxene

grains constitute �50 vol% of the sample. Within them,average porosity is variable but averages �15 vol%; i.e.,8% of the overall section area. In the meteorite’s presentstate an additional 5 vol% consists of elongate voids locatedalong boundaries between the silicates (i.e., associated withthe coarsest Fe-metal and Fe-sulfide, and Fe-oxide weather-ing products). Some of this grain-boundary porosity prob-ably formed by a combination of weathering and (lessimportantly) plucking during section preparation. How-ever, for the most part these large, interstitial-elongatepores appear to be preterrestrial dilated/flushed spaces,based on the common occurrence of delicate selvages of fel-sic glass that appear to have once lined the walls of thepores (Figs. 4 and 5 show some examples). Overall, the pre-terrestrial porosity was probably between 9% and 12%.

The original, pre-shock configuration of the silicategrains is easily discernible, and it was the typical equigran-ular texture of ureilites, with olivines and pyroxenes bothup to 3 mm in length, but more typically 1–2 mm across(see Fig. EA-1 in the Electronic Annex to this paper). Shock

Fig. 1. LAR 04315,21 in plane-polarized light (lower image, crossed polarizer). This 3.5 mm wide area consists mostly of pyroxene, includingtwo regions of “intact” pigeonite: blue in upper left of bottom image, purple in lower right. Twinning is manifest in the lower right pyroxene.The central area is dominated by a typically mosaicized olivine. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

Pyroxene-selective impact smelting in ureilites 5111


metamorphism altered the olivines and pyroxenes in verydifferent ways. The olivines now consist of mosaics of small(usually �0.5 mm, typically of order 0.1 mm, across)often angular pieces (Fig. 6; cf. the initial description ofLAR 04315: McCoy, 2005), indicating that they werestrongly shocked, shattered, and subsequently weldedtogether by an annealing process. Mosaicized olivine is typ-ical of highly shocked ureilites such as Goalpara and ALH81101 (Berkley et al., 1980; Berkley, 1986; cf. Rubin, 2006;and Mittlefehldt et al., 1998, who cite 5 additional examples).

In terms of porosity and general texture, two distincttypes of pyroxene are readily observable. A few widely scat-tered patches, constituting less than 1% of the total pyrox-ene, have sharp optical features (birefringence, also rarecleavages) and are free of all pores greater than about2 lm across (Fig. 2). By far the largest such patch is anear-conterminous zone of �0.9�0.7 mm (Fig. 2, top).Although, for simplicity, we refer to these isolated remnantsas intact, even within them the submicrometer porosity isgenerally 2–3% (Fig. 3). Patches completely free (�1%) ofsubmicrometer porosity are only a minor subvolume of

what we call “intact” pyroxene. In shape, the submicrome-ter pores of the “intact” pyroxenes tend to be nearly round.Compositionally (Fig. 7), the intact pyroxenes form a tightcluster in the usual fashion of ureilite pigeonites (e.g.,Goodrich et al., 1987) at �En74.1Wo10.2 (Table 1); it is ananalytical challenge to resolve any significant variation.The boundaries of the intact patches tend to be wavy,and vary (even from side to side of the same intact patch)between sharp and gradational.

The other, far more prevalent type of LAR 04315 pyrox-ene features the aforementioned �15 vol% porosity in a dis-tinctive vermiculate form (Fig. 2). The vermicular pores aregenerally �10 lm in diameter, and thus when sliced for athin section are typically about 1/4 as deep as the thicknessof the thin section. They often appear light in secondary-electron SEM images, but always appear dark in backscat-tered electron images and generally also in optical-lightimages. From 2-dimensional slices the total lengths of thesetubular pores are difficult to assess, but most appear curvy-elongate, with apparent maximum dimension of 30–40 lm.Compositionally, this porous variety of pyroxene is diverse

Fig. 2. The largest expanses of “intact” pyroxene in LAR 04315,21 (top pair of images) and LAR 04315,20 (bottom pair). The left two imagesare cross-polarized light photos; the right two are backscattered electron (BSE) images. Such intact pyroxenes are atypical. Most LAR04315pyroxene has been impact-smelted into the ragged condition (porous, with scattered Fe-metal and glass inclusions) seen in the upper left ofLAR 04315,21 (top) images and the lower right of the LAR 04315,20 (bottom) images.



but consistently higher-mg than the intact type (Fig. 7), pre-sumably as a result of smelting (kinetics and local chemicalgradients apparently caused the second-generation, FeO-depleted pyroxenes to nucleate with diverse Ca contents).The vermicular pores may not be completely empty. InX-ray maps, they show locally higher Ca, Mg and O thanthe larger, interstitial voids. However, this may merely bean artifact related to the vermicular pores’ frequently shal-low depth. The common, porous type of LAR 04315 pyrox-ene also contains a smattering of Fe-metals, usually <10 lmacross (e.g., Figs. 2 and 5), and similar-sized patches of Si-rich glass.

This prevalent, 15%-porous (smelted) type of LAR04315 pyroxene was not completely disordered by shockmelting. After the impact, the grains consistently recoveredinto birefringent and optically near-continuous crystals(Figs. 1b, and 2a and c). The recovery was not complete(optical continuity is far from perfect), but it is nonethelessobvious within each and every former pigeonite crystal. Inthe rare areas where isolated patches of intact pigeonite sur-vived, the surrounding smelted portion of the same grainconsistently shows interference color close to that of the in-tact remnant. Twin boundaries manifest as sharp, distinctplanes (Fig. 1b). These coarse twins probably developedduring rapid cooling, as the initial post-shock crystalstransformed from protoenstatite to pigeonite.

The most obvious effect of shock on the olivines wasmosaicism. Olivine mosaic tiles show a loose tendency to-ward being smaller near rims (of the original large olivinegrains) and larger near cores. As usual with ureilite olivines,the (original) rim areas are exceptionally magnesian as a re-sult of reduction (e.g., darkening of the olivine near thelower left corner of Fig. 6). In some areas of the approxi-

mate (original) rims, clusters of small, reduced and oftenrounded olivine remnants are set amidst interstitialpyroxene, mostly of relatively high-Ca, magnesian (subcal-cic diopside) composition, and/or glass of that approximatecomposition (Figs. 8 and 9). Our informal term for thisphenomenon is “olivine microporphyry” texture. Larger,interior mosaic tiles are commonly surrounded by a“necklace” of minute (of order 1 lm) opaques, probablymostly Fe-metal with lesser Fe-sulfide (Fig. 6). At leastsome of these Fe-metals (those distributed in long quasicir-cular arcs unlikely to represent injections) probably formedby reduction of olivine FeO; ultra-high-contrast BSE imag-ery suggests subtle compositional variation between adja-cent large interior tiles (Fig. 6).

In LAR 04315, most of the olivine appears to haveundergone only limited reduction. Ureilites in generalfeature an excellent correlation between core (unreduced)olivine mg and low-Ca pyroxene mg (Fig. 10). The tightcompositional clustering for the structurally “intact” typeof LAR 04315 pigeonite (from four separate originalgrains) suggests this composition is unaltered pigeonite.Based on the general ureilite correlation (Fig. 10) theolivine that coexisted with this pigeonite was originally veryclose to (most likely within 1 mol% of) Fo81. The observedcompositions for (original) core-portion olivines clustertightly at Fo81.9; so the extent of reduction, apart from (ori-ginal) rim materials, appears to have raised the core olivinemg by only �1 mol%, and certainly less than 4 mol%.

The pyroxene Fe/Mg vs. Fe/Mg trend (Fig. 12) is to firstapproximation consistent with simple loss (reduction) ofFeO with negligible reduction of MnO; i.e., the data distrib-ute close to a line between the intact pigeonite compositionand the origin. There is a slight tendency for high-Ca vari-eties of reduced pyroxene to have lower Fe/Mn than theirlower-Ca counterparts. Curiously, the trend of the reducedcompositions extrapolates slightly to the high-Fe/Mn sideof the intact composition. The pyroxene Fe/Mg vs.Cr/Mg trend (Fig. 13) indicates that Cr was also lost byreduction. On this diagram, pure Fe loss would drive thecompositions, as they diverged from the intact pigeonite,horizontally toward the left. The actual data instead trenddownward in Cr/Mg, although the trend is complicatedby a strong tendency for high-Ca varieties of reducedpyroxene to have higher Cr/Mg than their lower-Cacounterparts.

As shown on the pyroxene quadrilateral (Fig. 7), apartfrom the very minor intact and near-intact patches, the finalpyroxenes have mg consistently >87 mol%. In the data setshown in our figures (e.g., Fig. 7), extreme high-mg andCa-rich compositions are, along with intact patches, surelyoverrepresented, as we have sought to characterize the fullcompositional diversity engendered by the smelting process.McCoy (2005) found that the pyroxene is “dominantly”

Fs7–9Wo7–9; i.e., pigeonitic with mg �90. Assuming (a) thatthe pre-shock peridotite had negligible high-Ca pyroxene,(b) that the average composition for the final, mostlysmelted pyroxene is near to a line from the Fs apex(Fig. 7), and (c) that its mg is �89 mol%, we infer the aver-age final composition to be En79Fs10Wo11. Thus, the pyrox-ene FeO content has been reduced from an original 10.2 to

Fig. 3. BSE image of the largest “intact” pyroxene in LAR04315,21. At high magnification, most nominally intact pyroxenemanifests extremely fine porosity (along with scattered micrograinsof Fe-metal), such as in the fringe region shown here. However, afew small regions, usually deep within the “intact” patches, areessentially devoid of porosity, even at this scale. Some typicalimpact-smelted pyroxene, with its relatively coarse porosity, isvisible in the upper left corner.



�6.4 wt%. Translated from pyroxene into whole-rock com-position, this FeO reduction is �1.6 wt% of the whole rock.

At the same time, FeO was also undergoing reductionwithin the LAR 04315 olivine. Reduction was minor (lead-ing to �1 mol% increase in Fo) in regions of the original-grain cores, but extensive (locally all the way to Fo99) inthe original-grain rims. Assuming that 80–90% of the oliv-ine is core material little-modified by reduction (if mg in-creased from Fo81 to Fo82, the implied FeO reductionamounted to �0.9 wt% of this olivine, i.e., from �17.9 to17.0 wt%), and that on average the rim olivine (10–20%of the total) was reduced to �Fo91, we estimate as a roughaverage for the total LAR 04315 complement of olivine anextent of FeO reduction that amounts to �2.0 wt% of theolivine (from �17.9 to 15.8 wt%), or �1.1 wt% of the wholerock.

Probably there was also finite reduction of FeO from Cr-spinel (cf. Warren, 2004). In total, the textural and mineral-compositional observations suggest that FeO constitutingroughly 3 wt%, or as a fairly conservative lower limit2.5 wt%, of the original rock was reduced to Fe-metal. In-deed, this ureilite is uncommonly rich in Fe-metal. Themodal abundance of �4 vol% translates into 9 wt%, 1.5

times the highest known precedent among ureilites(5.9 wt% in Goalpara: Wiik, 1972).

The most common occurrence of glass in LAR 04315,apart from minute patches within smelted pyroxene, is asa thin selvage, often interstitial to a silica phase, on therim of coarse olivine or pyroxene grains. As suggested byFigs. 4 and 5, and EA-2, these selvages usually occur wherethe outward (in relation to the coarse mafic silicate) side ofthe selvage is a large elongate pore; possibly a former nar-row space between grains that has been dilated by gasexpansion during smelting. In these selvages the innermostgrains of the silica phase, where silica more or less continu-ously coats the complexly altered rim of a coarse mafic sil-icate, typically have an elongate, vermicular appearance,with apparent dimensions of order 100�5 lm, or even long-er (Figs. 4 and 5, and EA-2). Further out, where the silicaphase tends to be embedded in a comparable volume of fel-sic glass, the silica grains appear (at least in 2-dimensionalthin section views) as shorter, more equidimensional ovoids(Fig. 4). The silica phase generally contains <2 wt% Al2O3

(Table 1).The felsic glasses of LAR 04315 are remarkably diverse

in composition (Fig. 11), probably because a complex mix

Fig. 4. BSE images of glass- and silica-rich selvage areas on rims of olivine in LAR 04315,21 (top) and LAR 04315,20 (bottom). Darkest grey,ovoid/vermicular phase is silica. Medium grey phase intergrown with silica is felsic (�21 wt% Al2O3) glass. Whites are sulfides and Fe-metals.Inward toward the olivine, the selvages consist of a variety of high-mg pyroxenes and glasses, which manifest as a patchwork of differentshades of grey. For context, cf. Fig. EA-2, which has lower magnification images centered on same areas.



Fig. 5. BSE image of LAR 04315,20, showing selvages developed along the interfaces between large, tubular pores and large grains of botholivine (bottom and right edges of the view) and impact-smelted pigeonite (upper left). The field of view is �480 � 360 lm. For most of theirlength, the selvages here are dominated by silica (darkest grey). In lower left (10–80 lm to the left of the large white opaque), the silica isintergrown with felsic (�21 wt% Al2O3) glass, which is a lighter grey. Glass of similar composition (18 wt% Al2O3) is also present within theimpact-smelted pigeonite. Farther from the central void channel, the extended selvages consist of a complex variety of high-mg pyroxenes andglasses (cf. Fig. 4). Note toward lower right corner that a thin septum of selvage-silica separates the long, sausage-shaped, vertically alignedvoid from the similarly broad, near-round void to its left.

Fig. 6. Typical olivine mosaicism in LAR 04315,21. The larger, cross-polarized light view is 800 � 600 lm. The BSE image (left; likewise600 lm from top to bottom) shows that some individual mosaic tiles in the interior of the olivine have undergone perceptible reduction(manifested by the relatively darker grey of the central mosaic tile). However, the difference is so slight it cannot be detected by electron-probeanalysis (with normal count duration, beam current, etc., settings).



of disequilibrium processes led to their origin. Al2O3 is ashigh as 25 wt%, and shows a strong inverse correlation withNa2O (Fig. 11). Three of our analyses represent a round (d= 8 lm) blister-like feature (Fig. EA-3) within the rim of anolivine (the blister is surrounded to a distance of �20 lm bypure, albeit reduced, olivine, and is 40–50 lm from the typ-ically complex selvage at the rim of the olivine). Possibly asan original characteristic, or perhaps due to late reactionwith the enclosing olivine, this glass has a distinctively Al-poor and yet CaO-rich composition (Fig. 11). Glasses deepwithin grains of smelted pyroxene have moderate Al2O3 buthigh SiO2 and Na2O (Fig. EA-4).

The glass-rich region shown in Fig. 8 is unique in thatthe glass + silica texture, usually found as a selvage betweena large mafic silicate grain and a comparably large pore, ishere found spanning the space between the rims (albeithighly modified) of two olivine grains. This locale appearsto represent an exceptional clotting of the terminal-stagemelt (i.e., the melt was not blown out through the dilatedspace between grains). Compositionally, the glasses heretend to be unusually Na-poor compared to other selvageglasses (Fig. 11).

The metals of LAR 04315 are also compositionally di-verse (Table 3, and Figs. 14 and EA-5). The diversity inSi is especially interesting, because the partitioning of Siinto Fe-metal (i.e., the reduction of SiO2 into metallic Si)is a strong function of fO2 (Ziegler et al., 2009). Thus, inthe extremely reduced enstatite (E) chondritic meteorites,Si contents in metals are typically 0.2–2.0 wt% in EL and2–4 wt% in EH (Zhang et al., 1995; these authors also notedthat Si content tended to increase as the E chondritic metalswere altered by thermal and impact metamorphism). Mostureilite metals contain <0.4 wt% Si (Goodrich et al., 2009),but several metals in ALH 78019 have 2.8–2.9 wt% (Berkleyand Jones, 1982). Paradoxically, in LAR 04315 the lowestSi contents are found for the metals most obviously affectedby smelting; i.e., metals that are deeply embedded withinpyroxene (and thus probably formed primarily by smelting

of pyroxene FeO). The large, elongate metal grains sited inthe interstitial gaps between the original coarse mafic sili-cates have the highest Si, at 2.1–2.6 wt%, comparable tothe metals of extremely reduced EH chondrites (Fig. 14).A small ovoid (d � 20 lm) of metal within one of the fel-sic–silicic selvages yielded two analyses with �2.0 wt% Si.Four small (roughly 5 lm) metal grains in one of the largestregions of “olivine microporphyry” texture (i.e., whereshattered remnants from the rim of an original-coarse oliv-ine have been rounded and reduced to near-pure forsteriteby reaction with interstitial, compositionally near-pyroxenemelt) have relatively low Si levels of order 0.2–0.4 wt%,comparable to metals of the least metamorphosed EL chon-drites (Zhang et al., 1995).

In the felsic–silicic selvages Fe-sulfide is far more abun-dant than Fe-metal. High levels of Cr in some of the Fe-sul-fides (Table 3; also numerous semiquantitative analyses)provide additional evidence for smelting having occurredat extremely low fO2. The large sulfide shown in Fig. EA-6 contains 6.3 wt% Cr. The lowest Cr (1.2 wt%) was foundin a sulfide within an olivine microporphyry area.

3.2. Enigmatic LAP 03587

LAP 03587 is a 0.26 kg ureilite, when paired samplesLAP 03721, 03722 and 031109 are included (LAP 03587,per se, is 0.13 kg). LAP 03587 is an extraordinarily ferroanureilite. Our 24 analyses of core portions of olivines (6 dif-ferent grains) average Fo74.7. There are two precedents, ornear precedents, for Fo75 olivine among ureilites, LEW88774 (Goodrich and Harlow, 2001; Warren and Huber,2006a) and Y-791839 (Yanai et al., 1987), but no true urei-lite has olivine significantly more ferroan than LAP 03587(NWA 1500 is more ferroan, but Mittlefehldt and Hudon,2004, and later Goodrich et al., 2006, concluded that isnot a ureilite). LAP 03587 is enigmatic inasmuch as itspyroxene, like that of LAR 04315, is porous and reduced;and yet its overall texture, as described below, suggests amild shock history, quite unlike the heavy shock inferredfor LAR 04315.

We have studied thin section LAP 03587,7 (�1.0 cm2). Itcontains only �5 vol% pyroxene. The rest is nearly all oliv-ine, except for typical ureilitic abundances of carbon(mostly graphite, in this case) and other opaque phases.The texture (Fig. 15; see also Figs. EA-7–EA-9) is typicallyureilitic inasmuch as the mafic silicates are granular andmostly 1–2 mm across. In their initial description Corriganand McCoy (2005) noted that the pyroxenes “exhibit mosa-icism.” We note that the portions of the pyroxenes that ex-hibit mosaicism also show porosity, with the poresstrikingly similar in abundance, size and shape (Fig. 15)to the pores in a typically smelted LAR 04315 pyroxene.In this case, a larger proportion (roughly 1/3) of the pyrox-ene survived with approximately “intact” structural andcompositional characteristics. Compositionally, the intactpigeonite is tightly clustered at En69Wo10, whereas the re-duced, porous pyroxenes span a compositional range simi-lar to that observed in LAR 04315 (Fig. 7). Pyroxeneminor-element variation trends (Figs. 12 and 13) arealso similar to those observed in LAR 04315. One small

Fig. 7. Major-element compositions of pyroxenes (possibly includ-ing some glasses of near-pyroxene composition) in the LAR 04315and LAP 03587 ureilites. Original-composition, “intact” pigeonitesare denoted by dark blue symbols. (For interpretation of thereferences to color in this figure legend, the reader is referred to theweb version of this article.)



Tab

le1

Rep

rese

nta

tive

elec

tro

nm

icro

pro

be

anal

yses

(oxi

des

inw

t%)

of

ph

ases

inth

eL

AR

0431

5u

reil

ite.

Oli

vin

eco

res

in,2

0+

,21

Pyr

oxe

ne

“in

tact

”

in ,20

+,2

1

Pyr

oxe

ne

red

uce

din ,2

0+

,21

Pyr

oxe

ne

red

uce

din ,2

0+

,21

px/

gla

ino

livi

ne

mic

rop

orp

hyr

y

px/

gla

ino

livi

ne

mic

rop

orp

hyr

y

Gla

ssw

ith

inp

xin

,20

+,2

1

Gla

ssw

ith

ino

lin

,21

Set

of

sim

ilar

glas

ses

in,2

1

Set

of

sim

ilar

glas

ses

in,2

0

Set

of

sim

ilar

glas

ses

in,2

1

Set

of

sim

ilar

glas

ses

in ,20

+,2

1

Gla

sses

in ,20

Gla

sses

in ,20

Sil

ica

insm

elte

dp

xin

,20

+,2

1

Sil

ica

area

“6” in

,20

Sil

ica

area

“5” in

,20

Sil

ica

tub

ula

rin

,21

Sil

ica

glo

bu

lar

in,2

1

Cr-

spin

elin

.20

N aver

aged

3690

4319

711

22

33

2122

24

214

924

102

SiO

239

.23

55.1

157

.25

54.9

156

.36

54.1

571

.13

58.3

258

.64

58.1

461

.15

61.7

064

.04

67.0

895

.60

94.9

196

.78

98.0

996

.17

0.32

TiO

20.

010.

080.

090.

530.

030.

150.

721.

190.

430.

000.

430.

661.

110.

930.

060.

030.

040.

030.

020.

71A

l 2O

30.

030.

710.

431.

380.

230.

8517

.03

11.5

025

.14

24.7

822

.61

20.8

318

.44

17.8

61.

122.

390.

740.

532.

0124

.84

Cr 2

O3

0.57

1.08

0.74

0.51

0.94

1.13

0.07

0.70

0.09

0.10

0.05

0.08

0.00

0.10

0.07

0.03

0.03

0.03

0.05

43.7

9M

gO42

.89

26.9

732

.94

22.3

929

.57

22.1

80.

972.

924.

610.

673.

653.

504.

002.

370.

560.

100.

050.

020.

0814

.47

CaO

0.33

5.16

4.15

17.9

55.

9714

.97

4.41

20.5

910

.60

13.8

311

.10

10.7

210

.02

5.88

0.60

0.83

0.35

0.20

0.81

0.14

Mn

O0.

420.

440.

420.

330.

490.

390.

040.

260.

040.

010.

080.

090.

120.

070.

040.

010.

010.

010.

010.

64F

eO16

.91

10.2

33.

721.

336.

635.

270.

923.

300.

791.

300.

270.

330.

250.

500.

460.

190.

740.

250.

6014

.87

Na 2

O0.

010.

060.

010.

060.

020.

054.

120.

610.

530.

361.

041.

511.

434.

040.

060.

300.

050.

090.

130.

01K

2O

0.02

0.02

0.02

0.02

0.03

0.01

0.08

0.01

0.01

0.01

0.00

0.00

V2O

30.

030.

030.

030.

050.

080.

080.

24S

um

100.

4399

.86

99.7

799

.40

100.

2899

.22

99.4

999

.41

100.

8999

.20

100.

4099

.45

99.4

298

.91

98.5

998

.80

98.7

999

.24

99.8

910

0.02

Sto

ich

iom

etry

:T

ota

lca

tion

sb

ase

do

nre

leva

nt

nu

mb

ero

fo

xyg

ens

(id

eal

py

rox

ene

=4

cati

on

sp

er6

ox

yg

ens;

oli

vin

ea

nd

spin

el=

3p

er4

ox

yg

ens;

sili

ca=

2p

er4

ox

yg

ens)

3.00

4.00

4.00

3.99

4.00

4.00

2.03

2.04

2.02

2.01

2.03

3.01

En

74.1

86.7

62.1

78.7

61.8

Fs

15.8

5.5

2.1

9.9

8.2

Wo

10.2

7.8

35.8

11.4

30.0

mg

81.6

582

.46

94.0

96.8

88.8

88.2

65.1

61.2

91.2

47.9

96.0

95.0

96.6

89.4

68.4

48.5

11.5

10.6

18.6

63.4

Ab

bre

via

tio

ns:

px,

pyr

oxe

ne,

ol,

oli

vin

e.a

“p

x/gl

”d

esig

nat

esfi

ne-

grai

ned

mat

eria

lsth

atco

uld

be

eith

ercr

ysta

llin

ep

yro

xen

eo

rgl

ass

that

isve

ryn

ear

top

yro

xen

ein

com

po

siti

on

.



olivine-rim area was noted that is similar to the “olivinemicroporphyry” rim areas of LAR 04315 (Fig. EA-9).

Apart from its porous-reduced pyroxenes, the texture ofLAP 03587 does not indicate particularly intense shockmetamorphism. Roughly half of its olivines show wavyextinction, but none shows clear mosaicism (one grain doesshow a vague, coarse mosaicism). Its graphite is almostexclusively in the form of long, euhedral laths, often withpyramidal terminations (Fig. 15; see also Figs. EA-8 andEA-9). The longest such euhedron in LAP 03587,7 is2.5 mm long and nowhere wider than 0.10 mm. Olivinenear-contact with graphite is typically reduced and cloudedby abundant tiny opaques (mostly Fe-metal) to a depth ofroughly 100 lm. Euhedral ureilitic graphite is generallyinterpreted as evidence of an uncommonly mild shock his-tory, as in the cases of ALH 78019 (Berkley and Jones,1982), ALH 83014 (Bischoff et al., 1999) and Nullarbor010 (Treiman and Berkley, 1994).

However, Steele et al. (2009) have detected submicrom-eter diamond aggregates within a shock-metamorphosedgraphitic phase in LAP 03587. These diamonds must haveformed by at least moderately intense shock. Accordingto Matsuda et al. (1995) a minimum shock P of roughly25 GPa is required (cf. Bischoff et al., 1999; Nakamutaand Aoki, 2000). Another curious and possibly shock-re-lated feature of LAP 03587: some of its olivines contain asprinkling (�1 vol%) of tiny, submicrometer Fe-rich opa-ques, uniformly scattered even within the grain cores. Wehave not been able to identify the opaque phase(s), but itseems likely that they are dominated by Fe-metal thatformed when the host olivines underwent a dispersed albeitexceedingly mild degree of smelting in the immediate after-math of a fairly intense shock event. Mainly based on thestate of its pyroxenes and the presence of diamond, we inferthat despite its euhedral graphite, LAP 03587 probablyexperienced a typically ureilitic intensity of shock.

Fig. 8. This BSE image of LAR 04315,21 is centered on an unusual clot-like mass of selvage-like glass + silica matter. The surrounding coarsegrains were originally olivine (upper right), pigeonite (lower right corner; now a typical mass of impact smelting products) and olivine, whichin this area (lower left) is largely broken up into an “olivine microporphyry” zone (cf. Fig. 9). The darkest grey phase in the selvage is silica(however, an area of extremely reduced olivine near the upper left corner of the image is equally dark). The silicas are set amidst two differentphases that are difficult to distinguish here. Above and to the left of the (discontinuous) yellow line, the interstitial phase is highly felsic(�24 wt% Al2O3) glass. To the right of the yellow line, the interstitial phase is a pyroxene, or glass with almost perfect pyroxene stoichiometry,the composition of which clusters near En93Wo6 (this pyroxene manifests as a slightly darker grey than the felsic glass). In the olivinemicroporphyry zone, the dark grey is reduced (Fo90) olivine; the light grey interstitial phase is a high-CaO (17–19 wt%) glass, whosecomposition is reduced (mg = 90 mol%) and slightly Si-poor relative to pyroxene stoichiometry. The extension of the selvage into the moreintact olivine (upper right) consists of a complex variety of high-mg (and mostly high-Ca) pyroxenes and glasses (cf. Fig. 4). (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)



3.3. Embrangled Almahata Sitta

The Almahata Sitta meteorite fell as a shower of manytens of stones in 2008 (Jenniskens et al., 2009). It is a urei-lite-dominated but extraordinarily complex and porouspolymict breccia with a wide variety of ureilitic materials,including some that are unusually fine-grained, plus clastsof various kinds of chondrites (Bischoff et al., 2010). Wehave studied thin sections from five distinct Almahata Sittastones: three acquired direct from Jenniskens and twoloaned from John Kashuba, who in turn acquired themfrom Siegfried Haberer.

One of these (Jenniskens’ stone number 25) is an H5chondrite containing olivine (Fa18.1±0.4), low-Ca pyroxene(Fs16.2±0.4Wo1.8±1.0), 6–10-lm-size plagioclase grains(Ab82.9±3.6Or3.9±1.3), and minor chromite. Opaque phasesinclude polycrystalline kamacite averaging 0.49 wt% Co,martensite with 14.1 ± 5.0 wt% Ni, and thin veins of troilitewith variable Ni (<0.04 – 0.39 wt%). Shock-stage S3 is in-ferred from polysynthetically twinned low-Ca pyroxene

grains as well as undulose extinction and planar fracturesin many olivine grains. Haberer’s stone number 32 is anEL6 chondrite, containing enstatite (Fs0.21Wo1.6), relativelycoarse (10–80 lm across) plagioclase (Ab82.0Or3.5), kama-cite (0.75 wt% Si), and sulfides including oldhamite, troiliteand ferromagnesian alabandite. Almahata Sitta chondritecomponents reported by Bischoff et al. (2010) includeEH3, EH4/5, EH5, EL3/4 and EL6.

Another of the studied Almahata Sitta stones (Jennis-kens’ number 27) is an essentially unbrecciated commonureilite with Fo85.3 olivine cores and En82.1Wo4.7 pigeonite.

Fig. 9. Areas of “olivine microporphyry” material in LAR04315,20 (top) and LAR 04315,21 (bottom). Scale bars are25 lm. The dark grey rounded grains are reduced (Fo89–93)olivines. The light grey interstitial matter is either pyroxene or,more likely, a glass, typically CaO-rich (�10–15 wt%), reduced (mg

�89 mol%), and very close to pyroxene stoichiometry. The opaques(white) are mostly Fe-metal. Typically porous products of impactsmelting of pigeonite are seen in the lower left and upper right ofthe top image.

Fig. 10. When compositions of “intact” pigeonite are used, theolivine (core) and pigeonite mg ratios of LAR 04315 and LAP03587 conform to a strong correlation among ureilites in general(using data compilation of Mittlefehldt et al., 1998).

Fig. 11. Concentrations of four oxides (the balance is almostentirely SiO2) in various glasses from LAR 04315. Most of theseglasses are found in selvages at boundaries between the originalcoarse mafic silicates. However, as indicated by Fig. EA-3, the“blister” is uniquely situated within the intact interior of an olivinecrystal. For Na2O, open symbols designate glasses within the “clot”area shown in Fig. 8; these glasses are unusually Na-poor.



The original texture was probably poikilitic, with equantolivines 1–2 mm across and pigeonites in optical continuityup to 6 mm, but the optical continuity issue has been ob-scured somewhat by shock (the pigeonites show abundantand complex twinning; olivines show undulose extinction).Olivine reduction affected to a mildly unusual extent thedeep interiors of some grains, as the reducing fluid appar-ently coursed through in-grain cracks (Fig. EA-10). Theolivine cores average 0.70 wt% Cr2O3, but the reduction

trend shows depletion of Cr along with Fe, so that the mostextreme composition measured is Fo97.2 with 0.31 wt%Cr2O3. One of the largest pigeonites contains within300 lm (mostly within 100 lm) of its rim numerous blebs,mostly 20–40 lm across, that appear to have been alteredby reduction; they are seldom more reduced thanEn84.0–86.3Wo5.0–5.8, but one position, <4 lm from thegrain’s rim, is En94.8Wo1.1. In general, however, this pieceof Almahata Sitta is a rather ordinary ureilite.

The other two Almahata Sitta stones that we studied aresamples of the unusually fine-grained ureilitic lithologiesunique (as major components of a meteorite) to AlmahataSitta. One of these, Haberer’s number 109 (hereafter AS-H109), consists of a porous mix that is largely (roughlyhalf) two distinctive sublithologies (Figs. 16 and 17 andEA-11); the other half consists of similar but more chaoticarrangements of the same minerals (possibly derived bymixing-brecciation of the same ingredients that were pre-cursors to the two sublithologies). Both sublithologies oc-cur in the form of domains typically �0.5 mm across. Themore abundant (�30 vol%) of the two sublithologies we callAS-H109-OP (for olivine + pyroxene). It consists of Fo85–98

(average Fo91) olivine as the dominant (�80 vol%) phase, inthe form of equant and typically roundish grains of order10 lm across; along with, as anhedral masses interstitialto the olivine, a phase (�20 vol%) that is either subcalcicpyroxene or glass with near-pyroxene composition (averageEn63Wo30, the stoichiometry shows a slight excess of Si). InFig. 18, all of the “Haberer 109” pyroxenes with Wo >9 mol% (except for a few analyses from undocumented ran-dom positions) are from the OP sublithology. Even the leastreduced (Fo85–88) 109OP olivines have moderate “Cr2O3”,0.31–0.60 wt%.

The less abundant (�20 vol%) sublithology we call AS-H109-LPS (for low-Ca pyroxene + silica). It consists

Fig. 12. Fe/Mg and Fe/Mn variations in LAR 04315 and LAP03587 pyroxenes (including some suspected glasses with pyroxenestoichiometry). Symbols adorned with central black dots indicatehigh-Ca pyroxenes. The lines through the data are not regressionfits to these data; they reproduce fits obtained by Goodrich et al.(1987) from data for a diverse suite of ureilites.

Fig. 13. Fe/Mg and Cr/Mg variations in LAR 04315 and LAP03587 pyroxenes (including some suspected glasses with pyroxenestoichiometry). Symbols adorned with central black dots indicatehigh-Ca pyroxenes.

Fig. 14. Compositions of Fe-metals in LAR 04315 and the H109portion of Almahata Sitta (AS). Shown for comparison are fields forextremely reduced EL and EH chondrites, from Zhang et al. (1995).



mainly (�95 vol%) of low-Ca pyroxene grains of order10 lm across; along with, as scattered anhedral massesinterstitial to the pyroxene, a phase (�5 vol%) that is eithersilica or glass with near-silica composition. The largestpores and opaques tend to occur at OP/LPS boundaries.The pyroxenes that dominate 109LPS range from En81Wo4

cores to magnesian rims such as En97Wo1.3 and En93Wo5.6.They tend to be more angular than the rounded olivinesthat dominate 109OP. The 109LPS pyroxenes typicallyshow effects of reduction, in the form of FeO-depleted(BSE-dark) rims; in some areas their interiors are richlysprinkled with minute (�1 lm) opaques that are probablyFe-metals. Otherwise, however, Fe-metal tends to be coar-ser and also slightly more abundant in LPS than in OP. Intransmitted light, the LPS sublithology appears nearly opa-que; and elsewhere in the thin section it assumes long, nar-row schlieren-like forms (Fig. EA-11), suggesting that itwas once, locally, mobilized in a near-fluid state.

A third, extremely minor sublithology in AS-H109 con-sists of �10 lm olivine (Fo94–96) grains with interstitial feld-

spathic glass (wt% concentrations: 3.3 Na2O, 7.2 MgO, 18.5Al2O3, 65 SiO2, 0.4 K2O, 3.3 CaO, 0.4 TiO2, 0.8 FeO). Theolivines in this sublithology are angular and euhedral(Fig. 17) compared to typical olivines in OP domains. Prob-ably each �0.5-mm OP or LPS domain represents most ofthe volume of a former relatively coarse (normal ureilitic)grain that underwent granulation followed by pervasive im-pact smelting.

The other fine-grained sample, Jenniskens’ stone num-ber 1, is similar to the chaotic 1/2 of AS-H109, inasmuchas the dominant texture features rounded �10 lm olivinesintergrown with pyroxene (or pyroxenitic glass) amidstextraordinary porosity. However, in this case there is littleif any OP material (i.e., material in which olivine is prepon-derant and pyroxene only a minor interstitial phase) or LPSmaterial. Domains of striking modal consistency are sel-dom >100 lm across. The pyroxenes intergrown with oliv-ine are mostly low-Ca but have Wo ranging, even amongmoderate (< 92 mol%) mg compositions, from 0.7 to33 mol% (Fig. 18). The most extremely reduced pyroxeneis En99.0Wo0.15. The least reduced (Fo83–84) stone 1 olivineshave diverse “Cr2O3”, 0.32–0.81 (average 0.47) wt%.

In texture and mineralogy, the AS-H109-OP sublitholo-gy is similar to the minor “olivine microporphyry” compo-nents of LAR 04315 (Fig. 9) and LAP 03587 (Fig. EA-9).The extent of the lithology is vastly greater in Almahata Sit-ta, but as will be discussed below, a similar basic scenario ofimpact smelting probably determined the late evolutions ofall three meteorites.

4. DISCUSSION

4.1. Other possible instances of ureilite pyroxene smelting

According to a preliminary description (Janots et al.,2009), Jiddat al Harasis 422 is an impact-melt breccia with8.4 vol% vesicles. The role of impact in engendering the re-ported diversity of pyroxene composition (En84–92Wo0–5

orthopyroxene and En56–71Wo20–31 augite) is not yet clear,but amidst so much impact melting and porosity, thepyroxene may have undergone some degree of smelting.

Saito and Takeda (1990) described five ureilites (LEW85328, LEW 86216, Y-75154, Y-791839 and Y-8448) withdiverse pyroxene compositions that were interpreted as“shock effects . . . produced during the break-up of theirparent body.” In general, the pyroxene compositionaldiversity in these ureilites resembles what we have observedin LAR 04315 and LAP 0587 (Fig. 7; cf. Fig. EA-12). In theinterpretation of Saito and Takeda (1990), for most of theureilites of this group (Y-75154, Y-791839 and Y-8448)the compositional diversity and “cloudy” appearance ofpyroxene was simply attributed to some type of shock ef-fect, without explicit reference to reduction. In general,the interpretation was that “the chemical variations ofpyroxenes . . . can be explained by partial melting andvaporization mainly at grain boundaries and rims.” How-ever, in the case of the “moderately shocked” LEW85328, “portions of a pigeonite were partially melted” andSi-rich glasses found within cloudy pigeonite “may havebeen produced by reduction.” In the highly shocked

Fig. 15. BSE images of LAP 03587,7 showing two of its pyroxenes,both of which consist of roughly equal proportions of intactpigeonite and impact-smelted (porous) material. The other silicates(similarly grey) are olivine, which typically shows reduction (darkershades of grey and sprinkling of tiny white Fe-metals) near its rims.Elongate dark grains are euhedral graphites.



(mosaicized olivine) ureilite LEW 86216, its pyroxene diver-sity “may be one of characteristics of shock reduction, but. . . it is difficult to deduce its origin.”

Reviewer C. A. Goodrich reports seeing in ALH 81101,a ureilite with mosaicized olivine and compositionally di-verse pigeonite (Takeda et al., 1985), “all the features(including abundant pores)” of the LAR 04315 pyroxenetexture. The Y-790981 ureilite was described by Ogataet al. (1991) as having compositionally diverse “cloudypigeonite in which many phases [such as augite, chromite,metal and glass] were produced by shock partial melting.”Finally, the NWA 2634 ureilite has been described in theMeteoritical Bulletin (89) as “highly shocked with pyrox-enes completely converted to mosaic texture, olivine lessso.” Except for Jiddat al Harasis 422, none of these ureiliteshas been previously described (in print) as especially por-ous. However, from the perspective of LAR 04315 andLAP 03587, diverse pigeonite compositions in a monomictureilite might be viewed as prima facie evidence of pyrox-ene-focused impact smelting.

4.2. Extreme shock metamorphism, ureilite-style

In LAR 04315, the terminal-igneous smelting processcharacteristic of all ureilites took an unusual course, with

smelting more extensive, and overall more destructive ofFeO, in pyroxene than in olivine. LAP 03587 is similar, ex-cept it has much less modal pyroxene. The smelting obvi-ously occurred as a disequilibrium, uncompleted process,in a rapidly cooling environment. In an asteroidal context,the trigger for such a cascade of rapid change was presum-ably a large impact. Apart from stochastic factors includingpossible proximity to the impact and heterogeneous shock-wave propagation, it is unclear why LAR 04315 and LAP03587 happened to undergo unusually extensive impactsmelting. However, we can suggest a reason why impactsmelting in ureilites tended to be pyroxene-selective.

Heating by impact shock is a function of the targetmaterial’s volume response during shock compression andrelease, i.e., the area between the loading and unloadingpaths on a plot of volume V vs. P. We have used the for-malism of Sharp and DeCarli (2006; especially their Appen-dix A) and equation of state constants based primarily onAhrens and Johnson (1995) to obtain rough estimates forpost-shock temperature TS in orthopyroxenite and dunite(“bronzitite, Stillwater” and “dunite, low density”), asnear-approximations of the pigeonites and olivines in urei-lites. Key material-specific variables in the calculation of TS

include the initial density q0, the bulk compressional veloc-ity C, and the dimensionless slope constant S. The product

Fig. 16. BSE image showing the two main sublithologies in Almahata Sitta sample Haberer 109. Areas shown by shading in the map aresublithology LPS; the rest of this region consists of sublithology OP. Most of the opaques (white; in both OP and LPS) are Fe-metal.



q0C defines the “shock impedance” (He et al., 1996). Toconvert from the q0 listed by Ahrens and Johnson (1995)to one appropriate for an anatectic or near-anatectic urei-lite at roughly 1100 �C, we modeled thermal expansion afterFei (1995); i.e., the values of q0 shift from 3262 to 3164kg/m3 for the dunite and from 3277 to 3172 kg/m3 for theorthopyroxenite. We assume (conventionally) that for bothmaterials the specific heat is 1 kJ kg�1 K�1 and the Grunei-sen parameter �1. The thin curves in Fig. 19 show resultsusing values of C and S from Ahrens and Johnson (1995).

Fig. 19 also shows a set of results (thick curves) designedto account for the effects on C of a pre-shock T of order

1100 �C, along with relatively higher Fe in the olivine, aswell as higher Fe and Ca in the pyroxene (caveat: paralleleffects on S are admittedly unknown). For simplicity, we as-sume that for both pyroxene and olivine the overall effect(the DT effect dominates) is a reduction in C by a factorof 0.86. Compositional and thermal effects on C for maficsilicates were reviewed by Hood and Jones (1987), but webase dC/dT for olivine on the more recent data and reviewof Speziale et al. (2005). Although the modeling behindFig. 19 is crude, and neglects possible complexities suchas pre-existing porosity or unusually strong shock reverber-ations, it is probably safe to infer that the shock pressurethat engenders mosaicism in olivine (roughly 20–50 GPa:Stoffler et al., 1955) could lead to a TS of anywhere froma few tens to roughly 500 degrees above the pre-shock T;and that under such conditions the post-shock temperaturerise will not differ greatly between pigeonite and olivine.

The latter conclusion in turn implies that proximity ofthe pre-shock temperature T0 to the melting temperatureof the pigeonite was probably crucial. Raman studies indi-cate that when magnesian orthopyroxene (bronzite) isshocked at room temperature, it retains its crystalline struc-ture in almost unaltered form up to shock P of 62.5 GPa(Johnson et al., 2002) and undergoes only limited structuralchange even at 100 GPa (Estep et al., 1972). But if T0 wasclose to 1100 �C, and the shock P was a few tens of GPa (asimplied for LAR 04315 by the olivine mosaicism), TS within

Fig. 17. BSE images of Almahata Sitta sample Haberer 109. Thetop image shows an area of LPS sublithology. The dominant phase,low-Ca pyroxene, features reduced (a darker shade of mediumgrey) rims and a sprinkling of submicrometer sized opaques(almost certainly Fe-metal; but the abundance of this sprinklingphase is uncommonly high in this particular region of LPS). Largerwhite grains are mostly Fe-metal but include some FeS. The darkgrey interstitial phase (largest exposure is �5 lm to the left of thecenter of the image) is silica or very high-SiO2 glass. The bottomimage shows an area of unusual olivine + feldspathic glass subli-thology (the center and lower left portion of the view), along withthe more common OP sublithology (most of the upper 1/4 and therightmost 1/4 of the image). This particular area of OP has anuncommonly high ratio of pyroxene (light grey) to olivine(medium-dark grey).

Fig. 18. Compositions of pyroxenes (including fine-grained inter-stitial phases, possibly glass, with stoichiometry very near topyroxene) in two of the fine-grained lithologies of Almahata Sittaand, for comparison, LAR 04315.



pigeonite was likely high enough to induce scattered melt-ing, and widespread degradation of whatever crystallinestructure survived, to a point at which diffusive communica-tion between grain interiors and rims was vastly faster thanin a normally crystalline silicate. However, the survivaland/or recovery of the LAR 04315 pigeonites in a state ofnear optical continuity (Figs. 1 and 2, and EA-1) indicatesthat they were never completely melted. In general, the crys-talline structure of the pigeonite was degraded, but noterased.

We envisage transformation of most of the LAR 04315pigeonite (in the immediate lead-up to its subsequent smelt-ing) into a state comparable to that of the plagioclase-de-rived glassy phase “maskelynite.” The detailed nature andorigin of maskelynite are controversial. Although maskely-nite has been customarily regarded as a diaplectic glass,Chen and El Goresy (2000) argued from flow features inshocked meteorites that it must actually be a quencheddense melt. However, Yamaguchi and Sekine (2000) pro-duced analogous flow features experimentally, and inferredthat maskelynite “may be a mixture of plagioclase crystalsand fused glass.” A similar and probably still mainstreamview is that of Fritz et al. (2005; cf. Sazanova et al.,2007), who consider maskelynite to be a diaplectic glass“intermediate between a truly glassy state and the crystal-line state.” From the evidence of LAR 04315, we infer thatonce a ureilitic pigeonite was transformed into such a state,under P-T-fO2 conditions that favored smelting (rapidlydeclining T probably became the limiting factor), this mate-rial was vulnerable to thorough-going smelting.

Shock-degradation of crystalline structure is presumablyrelated to the melting temperature TM of the pure mineral.In LAR 04315, the localization of penetrative, thoroughsmelting within pigeonite probably reflects this mineral’slow TM in comparison to olivine (Fig. 20). For Fo70,Fo80, and Fo90 olivine, TM is �1580, 1670 and 1770 �C,respectively (Presnall, 1995). In contrast, by interpolationfrom data for Fe-free pigeonites and other pyroxenes, weestimate that TM for LAR 04315-like (En74Wo10; mg

�82 mol%) pigeonite is probably roughly 1350 �C. Thus,for the original LAR 04315 silicates, TM was about 300 Khigher for olivine than for pigeonite. From the concentra-tion of smelting into pigeonite, we can infer that the pre-shock temperature was probably such that TM,pigeonite –T0 was < 300 K; i.e., that LAR 04315 was still stewing ata T not far below 1200 �C when the impact that catastroph-ically altered its evolution, and probably that of the entireparent asteroid, struck.

The size-frequency distribution of pores formed withinLAR 04315 pigeonite is decidedly bimodal, with modes atroughly 1 lm (Fig. 3) and 20 lm (as an equivalent diameterfor pores that, in 2-dimensional exposures, are typically�40 � 10 lm). We speculate that this bimodality developedbecause at a scale of order 10 lm the bubbles reached coa-lescence, at which point access to reducing fluid was greatlyenhanced, and thus the smelting rate was enhanced. Fur-ther widening may have been limited by a combination ofincreasingly easy gas outflow as the local volume becamerich in closely spaced and interconnected pores, and dimi-nution of gas production as cooling led to slower andslower smelting.

Even a small amount of porosity, by rendering a mate-rial more compressible, leads to higher TS (Sharp and DeC-arli, 2006). Conceivably the pre-shock pyroxene was porouswhile the olivine was not, perhaps because an episode ofannealing erased olivine porosity while leaving some pyrox-ene porosity. But such a complex pre-history is difficult toreconcile with most of the textural evidence. The near opti-cal continuity of the pyroxene, even after the smelting pro-cess, suggests it was in a coherent condition before themajor smelting-trigger shock. The only textural featuresthat arguably point toward previous brecciation are a fewinstances where shock-injected opaque veins appear (butthe evidence is ambiguous) to cross “intact” zones ofpigeonite without extending into the surrounding ground-mass of smelted pyroxenitic matter. Olivine-pyroxeneboundaries are consistently smooth, as if unaltered from anormal ureilitic texture, except at scales below �10 lm,where in places granulation left the rims somewhat jagged.

In the LAR 04315 olivine, individual mosaic tiles do notshow shock effects, such as undulose extinction or planarfractures. We infer that a single major shock event was fol-lowed by an episode of reheating, or slow cooling, whichcaused sufficient annealing to erase the shock effects. Evenif the mosaicizing of the olivine involved only minor dis-placement, the interiors of the olivine domains (i.e., of theoriginal-coarse olivines) were rendered somewhat porous.Without an intervening period of annealing-compaction,reduction would have operated almost as effectively in theinteriors as on the edges. We assume that this thermal-com-

Fig. 19. Post-shock temperature TS calculated as a function ofshock pressure for orthopyroxenite and dunite (as stand-ins forolivine and pigeonite) starting from assumed TS = 1100 �C (seetext). Thinner curves assume values for bulk compressional velocityC and dimensionless slope constant S from Ahrens and Johnson(1995). Thicker curves show models with C modified to bettermodel a pre-shock temperature T0 of order 1100 �C, along withrelatively higher Fe in the olivine, as well as higher Fe and Ca in thepyroxene. For orthopyroxenite, each plotted curve represents asmoothing (by polynomial fit) of three separate curve segments(i.e., the three different sets of equation-of-state regimes cited byAhrens and Johnson, 1995).



paction process, and also the shock twinning of the pyrox-ene (Fig. 1), both could have happened soon after the initialshock, before shock heating triggered the impact smelting.Admittedly, however, these aspects of the history of LAR04315 are enigmatic.

The smelting of pigeonite more than olivine is especiallynoteworthy in the case of LAP 03587. Typically, in ureilites,olivines have prominent reduction rims but pyroxenes donot (e.g., Berkley, 1986). The lack of mosaicism in olivineand the preservation of euhedral graphite both suggest thatin LAP 03587 the shock pressure of the smelt-triggeringevent was not much greater than the minimum P for pro-duction of its diamonds (�25 GPa: Matsuda et al., 1995),and thus the increment from the pre-shock T0 to TS wasprobably not �100 K. We can only speculate that in thisinstance T0 was even nearer to the TM of this ureilite’spigeonite than was the case with LAR 04315; and/or somecircumstance caused this sample to undergo an unusualthermal evolution (e.g., an unusual pre-shock porosity; orperhaps after the smelt-triggering event the cooling ratewas unusually moderate). It may not be pure coincidencethat LAP 03587 is the most ferroan ureilite (Fig. 10), whichimplies an especially low TM,pigeonite (although probablyonly by a few degrees relative to the mg �82 mol% pigeon-ite of LAR 04315). But the impact process is highly stochas-tic, so it should not surprise that other almost equallyferroan ureilites show little evidence for pyroxene smelting(Goodrich et al., 2001; Warren and Huber, 2006a).

4.3. LAR 04315 as a post-depressurization ureilite

The impact smelting of LAR 04315 appears to representan extreme example of the complex, disequilibrium process-ing that altered all ureilites at the terminal stage of theirigneous evolution, when rapid depressurization occurredin association with impact disruption of the parent asteroid(Warren and Kallemeyn, 1992; Weber et al., 2003; Good-rich et al., 2004). Depressurization-triggered anatexis isthe dominant process that produces mantle melts on Earth.

With ureilites, the depressurization involved only a few tensof bars; and instead of triggering anatexis, the overall pro-cess modified the end and very short-term aftermath ofanatexis. In typical ureilites, with redox effects mainly evi-dent only in the rims of the olivines, the disruption causedthe thermal regime to abruptly transition from one of man-tle anatexis (or cooling shortly after anatexis) into one ofrapid, small-body cooling. But in the case of LAR 04315,the same impact that led to rapid cooling also initiallycaused pyroxene-localized impact melting, which in thechanged pressure conditions immediately took on the char-acter of impact smelting.

In all ureilites, the key consequence of hot excavation-depressurization was oxidation of solid carbon into COgas; in most cases with the oxygen derived almost exclu-sively from reduction of olivine rims (Warren and Huber,2006b). The olivine reduction reaction, the form of“smelting” observed in virtually all ureilites, isapproximately

2Cþ 2ðMg3FeSi2O8Þ ¼ 2ðMg2SiO4Þ þ 2ðMgSiO3Þþ 2Feþ 2CO ðgasÞ ð1aÞ

where 2(Mg3FeSi2O8) is an atomistic way of denoting 4 mo-les of Fo75 olivine. This assumes that the SiO2 liberated byFe-olivine reduction manages to efficiently react with sur-viving Mg-olivine to form pyroxene. Under disequilibriumconditions, the reaction might go more like

3Cþ 3ðMg3FeSi2O8Þ ¼ 4ðMg2SiO4Þ þ 3Fe

þ 3CO ðgasÞ þMgSiO3

þ SiO2 ð1bÞ

or even

2Cþ 2ðMg3FeSi2O8Þ ¼ 3ðMg2SiO4Þ þ 2Fe

þ 2CO ðgasÞ þ SiO2 ð1cÞ

which yields silica instead of Mg-pyroxene among the finalproducts.

Fig. 20. Melting temperatures of pure pyroxenes (Bowen et al., 1933; Carlson, 1988; Presnall, 1995). Temperatures for ferrosilite (Fs) andhedenbergite (Hd) are shown in italics because these minerals melt incongruently. Olivine melting temperatures (Presnall, 1995) are about300 �C higher than the interpolated TM for LAR 04315 and LAP 03587 pigeonites.



The pyroxene reduction process exemplified by LAR04315 is approximately

CþMgFeSi2O6 ¼MgSiO3 þ Feþ CO ðgasÞ þ SiO2 ð2Þ

where we use a simplified chemical formula, sans CaO, forthe pigeonite. In both Eqs. (1) and (2), the presence of COgas only on the products side implies that the reaction mustbe extremely pressure-sensitive. Note that the yield ratio ofSiO2 per mole of FeO that is reduced (or per mole of Fe-metal formed) is two times higher in the pyroxene reductionprocess (2) than in even the (1c) version of olivine-reduc-tion. In both cases, the detailed fate of this SiO2 is unclear.It may, in principle, manage to recombine with olivine toform new pyroxene; it may enter into an interstitial silicatemelt; and/or it may remain as isolated new grains of silica.Probably it partitions to some degree into all three of theseend states. But presumably the yield of silica, and/or silica-rich interstitial melt, was relatively high in the case of LAR04315. The silicate melt was probably (as discussed in detailbelow) fugitive, and may also have entrained some solidsilica. But even at �0.26 vol%, the modal abundance ofsilica + felsic–silicic glass is indeed uncommonly high fora ureilite (Figs. 4, 5 and 11, and EA-2).

Before the advent of porous impact-smelted ureilitessuch as LAR 04315, the most striking examples of depres-surization effects were found in a few extraordinarily ferro-an ureilites such as Asuka-881391 (Ikeda, 1999), NWA 766and LEW 88774 (Warren and Huber, 2006a). Having equil-ibrated (i.e., undergone anatexis) at relatively high fO2 (e.g.,Warren and Kallemeyn, 1992), ferroan ureilites are moreprone to contain Cr-spinel. They also started with moreFeO to “burn”. As, upon depressurization, the creation ofCO gas out of solid C sucked free oxygen out of the system,Cr tended to be reduced into a divalent state (valency of Cras a function of fO2 has been constrained by Hanson andJones, 1998). In ferroan ureilites NWA 766 (Warren,2004) and LEW 88774 (Goodrich et al., 2001; Warrenand Huber, 2006a), reduction of Cr2O3 from Cr-spinelled to the spinels being mantled by a thin layer of silicatesin which Cr2+ is a major, sometimes predominant, cation.These ferroan ureilites contain, in association with minorsilica (or near-pure-SiO2 glass) and in at least some casesalso with Cr-silicates (although Asuka-881391 seems to lackCr-silicates: Ikeda, 1999), felsic glasses. These glasses typi-cally contain �70 wt% SiO2, 18 wt% Al2O3 (cf. the felsicglasses associated with silica in LAR 04315: Tables 1 and2; Figs. 4, 5, 8 and 11).

In LAR 04315 and LAP 03587, as indicated by theexamples of Figs. 4, 5 and 8, and EA-2, -3 and -9, the felsicglass + silica association tends to occur as a rimming sel-vage where coarse silicates of the pre-impact lithology wereseparated by large voids. These assemblages appear to haveformed by crystallization/quenching of felsic melts thatclung as linings of the large voids. However, it would bea mistake to assume that the original melt can be modeledas a recombination of the observed subequal proportions ofglasses and silicas. Even as the glasses were crystallizing (oranyway, reacting with) the silicas, other processes wereoccurring, including smelting, which potentially, per (1c)and/or (2), added SiO2. Ureilites such as Asuka-881391,

NWA 766 and LEW 88774 show that major smelting ofpyroxene was not necessary to form a tiny proportion of sil-ica in association with relatively larger amount of felsicglass.

As suggested for Asuka-881391 by Ikeda (1999), all ofthese glasses probably derived in part from minor propor-tions of interstitial anatectic melts that were present beforethe impact disruption of the ureilitic mantle. However, thehigh relative amount of silica in LAR 04315 (in severalareas, the rimming selvages consist preponderantly of silica:e.g., Figs. 5 and EA-3) is presumably related to the extraor-dinary importance of pyroxene smelting (2) in this rock.Compared to most of its counterparts in Asuka-881391,LEW 88774 and NWA 766 (Table 2), the typical LAR04315 selvage-area felsic glass composition is lower inSiO2, and richer in CaO and especially Al2O3. We speculatethat the LAR 04315 felsic glass became (or stayed) rela-tively Si-poor because it was consistently in near-contactwith a silica phase, whereas in other ureilites the felsic glasswas generally more isolated from silica, and became (orstayed) supersaturated in SiO2 as a consequence of nucle-ation kinetics.

Although none of the LAR 04315 glasses have Na2O>4.5 wt%, an anticorrelation between Al2O3 and Na2O(Fig. 11, excepting the olivine-blister glass) may in partreflect volatilization of Na. However, Na volatilizationis probably not the sole reason for the diversity ofAl2O3 in the LAR 04315 glasses, because loss of�4 wt% Na2O (K2O is <0.1 wt% even in the most Na-rich glasses) cannot account for increase in Al2O3 from�18 to 25 wt%. Also, if another refractory element, Ti,is substituted for Al, the anticorrelation disappears intoa cloud of random scatter.

SiO2-rich clasts occur in some enstatite and ordinarychondrites (Rubin, 1983; Hezel et al., 2006). However, theseclasts do not seem very relevant. The SiO2-rich clasts in theordinary chondrites apparently represent survival of ex-treme products of solar-nebular (e.g., fractional condensa-tion) processing (Hezel et al., 2006).

High-SiO2 glasses that occur as minor components ofsome peridotites derived from the mantle of the Earth (Hir-schmann et al., 1998) may be more useful analogs. Some areeven associated with a “spongy” texture within incipientlymelted pyroxene that superficially resembles the LAR04315 smelted pyroxenes (Carpenter et al., 2002; Zhu,2008). These terrestrial glasses are believed to representquenched anatectic melts, and factors that lead to high-SiO2 are believed to include low melt fraction, low P, andhigh alkali-oxide activities (Hirschmann et al., 1998). Alkaliions tend to reduce Si–O–Si linkages in the melt, whichtranslates into negative deviations from ideality for mixingbetween alkalis and silica, which in turn requires high alka-lis to be accompanied by high silica for liquids in equilib-rium with peridotite. The importance of low P may stemfrom a tendency for alkali-carbonate complexes to format high P (Schiano et al., 1998). Experiments at P � 1 GPa(still off-scale high by asteroidal standards) indicatethat incipient peridotite melts begin to acquire unusuallyhigh-SiO2 as the melt Na2O + K2O surpasses �9 wt%(Hirschmann et al., 1998).



Tab

le2

Rep

rese

nta

tive

elec

tro

nm

icro

pro

be

anal

yses

(oxi

des

inw

t%)

of

ph

ases

inL

AP

0358

7,A

lmah

ata

Sit

taan

do

ther

ure

ilit

es.

LA

P03

587

Alm

ahat

aS

itta

,A

S-H

109

Alm

ahat

aS

itta

,P

J-27

Asu

ka-

8813

91b

LE

W88

774b

NW

A76

6b

Oli

vin

eco

res

Pyr

oxe

ne

“in

tact

”

Pyr

oxe

ne

red

uce

dP

yro

xen

ere

du

ced

Gla

ssin

oli

vin

em

icro

po

rph

yry

Gla

ssin

ters

titi

alS

ilic

ain

inte

rsti

tial

glas

s

H10

9-O

Po

livi

nes

div

erse

H10

9-O

Pp

x/gl

ain

inte

rsti

tial

H10

9-L

PS

pyr

oxe

ne

core

s

H10

9-L

PS

pyr

oxe

ne

rim

s

Oli

vin

eco

res

Pyr

oxe

ne

core

sF

elsi

cgl

ass,

“S

i-p

oo

r”

Fel

sic

glas

s,S

i-ri

ch

Fel

sic

glas

s,d

iver

se

Fel

sic

glas

s,d

iver

se

N aver

aged

2413

416

14

220

424

99

192

343

58

SiO

238

.35

54.4

953

.83

56.5

249

.89

68.6

797

.59

39.9

153

.28

55.7

957

.35

39.1

955

.17

65.1

574

.83

70.0

371

.21

TiO

20.

020.

060.

220.

030.

780.

350.

040.

030.

270.

060.

080.

030.

050.

300.

43A

l 2O

30.

030.

951.

830.

237.

3418

.90

0.57

0.04

1.50

0.39

0.67

0.03

0.52

20.4

514

.40

17.6

718

.35

Cr 2

O3

0.60

1.20

1.06

1.00

2.23

0.03

0.01

0.52

1.28

0.80

0.36

0.69

1.02

MgO

37.9

024

.72

22.7

030

.60

18.3

40.

980.

0149

.96

22.7

431

.87

35.6

245

.95

30.6

21.

300.

870.

170.

64C

aO0.

354.

9512

.61

3.61

17.8

85.

230.

050.

2515

.23

2.81

2.99

0.26

2.42

2.80

1.53

6.79

2.99

Mn

O0.

400.

400.

510.

450.

420.

100.

050.

480.

380.

400.

350.

450.

440.

250.

270.

030.

04F

eO22

.93

13.1

16.

246.

973.

541.

010.

189.

114.

587.

221.

7414

.03

8.83

1.60

1.27

0.32

0.59

Na 2

O0.

020.

050.

110.

020.

153.

480.

210.

010.

200.

070.

050.

020.

068.

055.

533.

723.

30K

2O

0.01

0.08

0.01

0.01

0.01

0.01

0.01

0.01

0.01

0.35

0.40

0.02

0.61

Su

m10

0.62

99.9

399

.11

99.4

310

0.56

98.8

398

.73

100.

3199

.48

99.4

099

.20

100.

6599

.16

100.

2599

.53

99.6

899

.04

Sto

ich

iom

etry

:T

ota

lca

tion

sb

ase

do

nre

leva

nt

nu

mber

of

ox

yg

ens

(id

eal

py

rox

ene

=4

cati

on

sp

er6

ox

yg

ens;

oli

vin

ea

nd

spin

el=

3p

er4

ox

yg

ens;

sili

ca=

2p

er4

ox

yg

ens)

3.99

3.99

3.99

2.01

3.02

4.02

4.02

4.01

3.01

4.01

En

69.4

64.4

82.5

55.3

62.7

84.0

91.9

82.1

Fs

20.6

9.9

10.5

6.0

7.1

10.7

2.5

13.3

Wo

10.0

25.7

7.0

38.7

30.2

5.3

5.5

4.7

mg

74.6

677

.07

86.6

88.7

90.2

63.4

10.9

90.7

89.9

88.7

97.3

85.3

786

.08

59.1

554

.94

49.2

66.0

a“p

x/gl

”d

esig

nat

esfi

ne-

grai

ned

mat

eria

lsth

atco

uld

be

eith

ercr

ysta

llin

ep

yro

xen

eo

rgl

ass

that

isve

ryn

ear

top

yro

xen

ein

com

po

siti

on

.b

Dat

afo

rA

suk

a-88

1391

fro

mIk

eda

(199

9);

LE

W88

774

and

NW

A76

6fr

om

un

pu

bli

shed

UC

LA

/War

ren

anal

yses

.



To a zeroeth approximation, the LAR 04315 glass diver-sity can be modeled as a mixing of three components: thehigh-Si, high-Na glass that formed within pyroxene (i.e.,formed more or less directly by smelting), the olivine-blisterglass, and an Al-basaltic melt, perhaps representing a minormelt component that was present interstitially to the coarsemafic silicates as this material was stewing in the ureiliteasteroid mantle prior to the impact smelting episode. Byanalogy with the low-P anatectic melts discussed by Hirsch-mann et al. (1998), the initial interstitial melt may have al-ready been SiO2-rich even before the smelting processadded SiO2. The interstitial melt may also have been mod-ified by abortive equilibration to the extreme low FeOactivity associated with the rock’s pervasive smelting.LAR 04315 glasses typically have mg�90 mol% (Table 1).

Both Eqs. (1) and (2) imply that the molar proportional-ity of COx gas to Fe-metal in the products, i.e., to FeO inthe reactants, is close to 1. By thermodynamical calculation,at magmatic temperature T and low P near the graphite–COx buffer, the gas is preponderantly CO (Warren and Kal-lemeyn, 1992). Thus (factoring in the molecular weights ofCO, SiO2 and Fe), for each 1 wt% of FeO that is reduced,“smelting” will form �0.39 wt% of CO, along with0.78 wt% of Fe-metal and roughly 0.3 (olivine reduction)to 0.8 (pigeonite reduction) wt% of SiO2. In other words,for each cm3 of pre-smelt rock (with density of roughly3 g cm�3), reduction of 1 wt% of FeO yields 0.012 g, i.e.,0.00042 moles, of CO. By the gas law, the volume of0.00042 moles of gas at �1200 �C is 0.50 cm3 at 100 bar;and for other pressures the volume is simply 50/P (for P

in bar). For the putative scenario of reduction with no es-cape of CO gas, the implied volume fraction C of CO(i.e., the porosity) that forms as a result of reduction is

C ¼ ð50w=PÞ=ð1þ 50w=P Þ ð3Þ

where w is the proportion (as wt% of the bulk rock) of FeOthat undergoes reduction. Relationships among w, P and(indirectly) C are illustrated in Fig. 21. As discussed above,in LAR 04315 it appears that w was�2.7 wt% (1.6 wt% frompigeonite reduction, 1.1 wt% from olivine reduction). C–COdriven redox processing will not reduce olivine past mg

�82 mol% (at magmatic T) unless P is less than about100 bar (Warren and Kallemeyn, 1992). It follows that the fi-nal porosity of 9–12% (see above) must represent only a smallfraction, at most �1/10, of the total COx gas volume pro-duced by the C–CO driven redox processing of LAR 04315.Most of the gas must have escaped by outward flow.

However, 100 bars is only an ultra-conservative upperlimit on the pressure during the redox processing, becauseit would imply a depth of many kilometers within any aster-oid (e.g., �12 km deep or �35 km deep, in asteroids withdiameter of 500 or 200 km, respectively; cf. Wilson et al.,2008; Warren, 2010); inconsistent with the rapid coolingimplied by the gross mineralogical disequilibrium. It ismore probable that the final stage of redox processing withsynchronous rapid cooling took place at a depth of�1 km;implying a P of <17 bar (the P at �1 km depth in a 500-kmdiameter asteroid). Given w �2.7 wt% and assumingP < 17 bar, (3) and Fig. 21 imply that the final porosityof 9–12% must represent less than 2% of the total COx

gas volume produced by the C–CO driven redox processingof LAR 04315. If the cooling was at the typical end-stageureilite rate of >1 degree/hour (Miyamoto et al., 1985; cf.Goodrich et al., 2001; and Tribaudino, 2006; suggestingroughly 7–20 degree/hour), then the burial depth was prob-ably no more than a few meters, P was of order 0.01 bar,and the final porosity of 9–12% represents only �0.001%of the total COx gas volume produced. In short, almostall of the gas escaped, by outward flow that probably endedwith explosive expansion (Warren and Kallemeyn, 1992;Scott et al., 1993) into the vacuum of space.

This conclusion is almost equally strong for ureilites ingeneral. Typical ureilites have olivines roughly1 � 0.7 � 0.5 mm in size (e.g., Berkley et al., 1980) with re-duced rims whose average thickness is roughly 20–40 lm(Wittke et al., 2007); i.e., the perceptibly reduced rims arein general roughly 9–17% of the olivines. By stoichiometry,the amount of FeO reduced within any given rim is�0.95 wt% for each mol% of Fo reduction. Assuming thaton average the reduced rims have lower Fo than the coresby 5–10 mol% (5 mol% is probably a conservative lowerlimit), the rims have 4.8–9.5 wt% lower average FeO. Thus,a typical ureilite is implied to have had 0.4–1.6 wt%, andmostly likely close to 1 wt%, of its olivine FeO reduced.In an average ureilite most of the FeO is in olivine (Mit-tlefehldt et al., 1998), so we can conclude that for ureilitesin general, w was also not�1 wt%. This result, per (3), im-plies a gas yield (Fig. 21) that is for any plausible P largerthan the typical preterrestrial ureilite porosity, i.e., so lowthat it is hard to measure and seldom reported (Brittet al., 2010, found for �14 samples an average porosity

Fig. 21. Volume proportion of CO gas that forms by the smeltingof a ureilite’s FeO shown as a function of lithostatic pressure P.Numbers along curves denote w, the proportion, as wt% of thebulk rock, of FeO that is assumed to undergo reduction. In the caseof LAR 04315, with its Fo82 olivine, the smelting reaction wouldnever have proceeded unless P was lower than �100 bar. Theobserved porosity in the final ureilite must represent only a smallfraction of the total volume of CO gas that formed during theimpact smelting. Most of the gas, and probably (making a morerealistic assumption for P) all but a tiny fraction of it, streamed outand escaped.



of 3%, but their data have large uncertainties). If the redoxprocessing took place at the low P implied by the extremelyfast cooling rates (Miyamoto et al., 1985; Goodrich et al.,2001; Tribaudino, 2006), all but a tiny fraction of the gasmust have escaped.

Thus, the ureilites’ terminal redox episode, at least forCOx gas, was far from a closed-system process. The out-streaming COx gas may have carried with it much of what-ever melt (silicate or metallic-sulfide) was present in intersti-tial areas; i.e., whatever melt remained from the previousmantle-anatectic stage when the redox processing began,and/or, in cases such as LAR 04315, may have been addedas a by-product of impact smelting. However, ureilitepyroxene core incompatible-element compositions (Guanand Crozaz, 2000) indicate that these materials were al-ready extremely depleted during the mantle-anatectic stage;and bulk S and trace-element data (Warren et al., 2006)indicate that major fractions, at least, of the metal and sul-fides not previously drained out during the mantle-anatecticstage managed to survive the terminal redox processing.Nor, based on our INAA results, is the LAR 04315 bulkcomposition unusual. Our bulk-rock results for Sm, as arepresentative incompatible element, are 10–13 ng/g. Re-sults for representative chalcophile and siderophile elementsinclude 210 lg/g Zn and 310 ng/g Ir (cf. typical ureilites:Warren et al., 2006). Surface tension and high viscositymay have been important factors in allowing a fraction ofthe melt to cling as selvages to the walls of the larger voids,as the CO gas streamed outward through the void channels.

The great volume of out-streaming COx may have en-hanced the thoroughness of the catastrophic impact disrup-tion of the ureilite parent asteroid. All but a tiny fraction ofthe gas must have eventually jetted out of vents scatteredacross the surface of each remnant piece of the body. Inthe microgravity environment of the disrupted asteroid,explosive expansion of the gas as it approached the zero-pressure surface probably helped to disaggregate thematerial (the cohesiveness of which was probably almostuniformly low, because the impact occurred when tracesof anatectic melt were still almost ubiquitous). Presumablythe outflow tended to self-organize into fewer and fewercoalescing channels as the gases approached the surface.But the force of the largest such jets would have tendedto propel the remnant bodies in random directions, i.e.,mostly away from the local center of gravity; thus tendingto reduce the likelihood for the original body to quicklyor efficiently reassemble.

4.4. Almahata Sitta: similarities and differences

Our discussion of Almahata Sitta focuses on the distinc-tive fine-grained lithology, and in particular on the Haberer109 sample (AS-H109), which consists in large part of the do-mains of mineralogical consistency that we call the OP andLPS sublithologies. We interpret these sublithologies as re-licts of grains of olivine and low-Ca pyroxene, respectively,within a normal ureilite that was impact-smelted in a mannersimilar to LAR 04315. In the case of AS-H109, the shockmetamorphism was more intense. It probably caused finergranulation of the olivine, and it engendered an even higher

post-shock temperature, which led to a more intense impactsmelting of the pyroxene. There was also much more dis-placement of materials, including in some areas a stretchingof pyroxenes into schlieren-like structures (Fig. EA-11),and elsewhere complete distortion to the point where no ves-tiges of the original grain shapes are discernible.

The detailed mineralogy of AS-H109 prior to the majorimpact and consequent impact smelting is unknown. Thereare no fully “intact” grains of either pyroxene or olivine.Conceivably the pre-impact lithology was a polymict brec-cia. Provided it was either monomict or unbrecciated, themg ratios of the most ferroan pyroxene (84.5 mol%;Fig. 18) and olivine (84.9 mol%) represent upper limits onthe mg of the pre-smelting pyroxene and olivine. One inter-esting feature is the differentiation of its pyroxenes betweenthe two sublithologies. The pyroxene that is the dominant(�95%) phase within the 109LPS sublithology is almostexclusively low-Ca (Wo <9 mol%); the sole exception,Wo32, is within �30 lm of (or put another way, separatedby 2–3 LPS pyroxene grains from) a large, ragged pore thatseparates its LPS domain from an OP domain. In the AS-H109-OP sublithology, which as discussed above resemblesthe olivine microporphyries of LAR 04315 and LAP 03587,the interstitial pyroxene (or pyroxene-like glass) is consis-tently Ca-rich (Wo >23 mol%). Thus, it appears that shockmelting and mobilization of AS-H109 pyroxene was com-positionally selective. Without knowing the detailed pre-im-pact mineralogy, it is hard to constrain the mechanism ofthis selectivity. Very possibly, in addition to the low-Cagrains that survive in relict form as the LPS domains, thepre-impact meteorite contained a considerable proportionof high-Ca pyroxene. A sizeable minority of ureilites haveboth augite and low-Ca pyroxene (Mittlefehldt et al.,1998; Goodrich et al., 2004). Conceivably, the selectivemelting of Ca-pyroxene arose during the complex impactsmelting of an original pigeonite; and presumably a finiteproportion of CaO was added from reduced Mg-olivine.In LAR 04315 and LAP 03587, where the interstitial pyrox-enes or pyroxene-like glasses of the olivine microporphyriesalso mostly have Wo >25 mol%, there is no evidence of anypre-impact high-Ca pyroxene. However, olivine micro-porphyry is only a trace component of LAR 04315, andeven rarer in LAP 03587, while the OP sublithology consti-tutes �60% of the non-chaotic half of AS-H109.

As in LAR 04315, the smelting of low-Ca pyroxene in theAS-H109-LPS sublithology engendered, per (2), both silicaand Fe-metal. In detail, however, the smelted low-Ca pyrox-ene of 109LPS is different. The material is broken into dis-crete, angular �10–20 lm grains, each of which typicallyfeatures a distinctly reduced rim (dark in BSE images, suchas Fig. 17). In many cases, as in Fig. 17, the interior of thegrain appears to have undergone a later stage of smelting thatformed a sprinkling of tiny (�1 lm) Fe-metals. Our electron-probe analyses of the interiors of these grains do not detectsuch tiny inclusions, except as “FeO” ingredients in pyroxene(the analysis is effectively a defocused-beam analysis, exceptthe Fe-metals are so tiny that their high density probablyhas minimal effect on electron and X-ray transmission). Inreality the final state of the 109LPS low-Ca pyroxene is prob-ably more reduced than reflected by Fig. 18. (This “later



stage” of smelting does not necessarily imply a discontinuityin the thermal evolution. It may reflect some kinetic-diffusiveeffect.) The other Fe-metals in 109LPS are distinctly coarsecompared to the metals within LAR 04315’s smelted pyrox-enes, which suggests that in 109LPS, during the period ofmost intense smelting, a greater degree of mobilization andcoalescence of the Fe-metal occurred.

4.5. Origin of compositional diversity among Fe-metals

To account for the pattern of Si diversity within the Fe-metals of LAR 04315 and AS-H109 (Fig. 14), we speculatethat the smelting-associated decline in fO2, which caused Sito be reduced from silicate and enter into Fe-metal, oc-curred in the following progressive, multi-stage way. Thesmelting of the pyroxene (which yielded the “within smeltedpigeonites” and LPS metals in Fig. 14) happened very soonafter the shock that triggered both the smelting and the dis-ruption of the parent asteroid, while fO2 remained near itspre-disruption intensity. But cooling soon caused thepyroxenes to resolidify, thus tending to isolate the new Si-poor metals within them. A similar fate overcame the smallnew metals of the olivine microporphyry locales (“olivineovoids” and OP metals in Fig. 14), where the silicate mate-rials (preponderantly forsteritic olivines) had especiallyhigh crystallization temperatures. In contrast, even at a la-ter stage the large elongate voids in between grains of thecoarse pre-disruption texture, dilated by out-flowing smelt-ing gas, remained open; and their lining materials, consist-ing largely of Fe-sulfide and felsic melt (now glass),remained reactive. Meanwhile, fO2, in consequence ofdecreasing pressure as the body continued to disrupt, hadreached a much lower level. Thus, metals adjacent to thelarge elongate pores (“original interstices” in Fig. 14) ac-quired much higher Si than most of the metals that formedby smelting of solid mafic silicates.

The normal-ureilitic Ni contents (average 5.2 wt%:Table 3) of the tiny within-pyroxene metals of LAR04315 are also enigmatic. Probably all of this Ni did not

derive primarily from smelting of pyroxene. We infer thatduring a brief interval after smelting had formed these met-als but before the pyroxenes resolidified, and before muchSi had reduction had occurred, these metals managed toequilibrate with the large elongate metal grains (interstitialto the coarse pre-shock mafic silicates) that were the rock’smain repository of Ni. The Ni data thus tend to confirmthat during the smelting of LAR 04315 even the pyroxenecores were opened to rapid exchange with their rims as aconsequence of pervasive COx-fluid porosity.

5. CONCLUSIONS

1. As exemplified by the meteorites LAP 03587 and espe-cially LAR 04315, the terminal, impact-triggered redoxprocessing of the ureilites was in some cases more effec-tive at reducing FeO from pyroxene than in the usualmanner from the rims of olivines.

2. Unlike the olivine-rim reduction phenomenon that hasoften been termed smelting, the process that reducedthe pyroxene of LAR 04315 probably was true smelting;i.e., reduction occurring along with and facilitated bymelting. The LAR 04315 pyroxenes were reduced(impact-smelted) almost as efficiently in their centers asin their rims; they are now almost uniformly porous(�15%) and sprinkled with small Fe-metals and pocketsof glass. However, these pyroxenes also managed toretain to an enigmatically large extent their former opti-cal continuity. Overall, the preterrestrial porosity ofLAR 04315 was probably between 9% and 12%.

3. Fortunately, in a few minor areas LAR 04315 pyroxenehappened to avoid smelting, so we know that the pre-impact pyroxene was, in typical ureilite fashion, a pigeon-ite with a very tight compositional range nearEn74.1Wo10.2.

4. The mosaicized state of the olivine in LAR 04315confirms that it experienced a severe impact shock.However, LAP 03587, which contains some pyroxene

Table 3Electron microprobe analyses (wt%) of metal and sulfide phases in the LAR 04315 and Almahata Sitta ureilites.

Fe Ni Si Co P S Cr N NG Total

Fe-metals in LAR 04315

Average, large elongate grains in original interstices 93.42 4.56 2.30 0.32 0.14 0.01 0.03 15 5 100.79Average, tiny grains within smelted pigeonites 93.95 5.23 0.01 0.30 0.09 0.02 0.08 5 5 99.68Average, tiny grains in olivine ovoids areas 97.63 1.14 0.26 0.14 0.19 0.05 0.09 5 4 99.49Grain in felsic glass + silica area 91.46 6.58 2.04 0.30 0.10 0.05 0.01 2 1 100.53

Fe-metals in Almahata Sitta, AS-H109

Grains (mostly large) not interior to either LPS or OP 92.44 3.69 3.52 0.29 0.10 0.01 nd 19 10 100.06Interior LPS sublithology grains, Si-rich 91.55 4.08 3.61 0.29 0.09 0.02 nd 4 4 99.89Interior LPS sublithology grains, Si-poor 94.95 4.39 0.13 0.34 0.10 0.02 nd 7 7 99.94Interior OP sublithology grains 97.03 1.65 0.09 0.14 0.16 0.02 nd 8 7 99.09

Sulfides in LAR 04315

Round 100 lm grain in felsic glass + silica area 59.21 0.10 0.01 0.00 0.02 35.07 2.86 2 2 97.27Large elongate grain in original interstice 60.62 0.14 0.00 0.03 0.00 34.50 1.75 2 1 97.05

Large elongate grain in original interstice 61.07 0.14 0.00 0.00 0.00 34.71 2.03 2 1 97.95Large elongate grain in original interstice 55.16 0.11 0.00 0.00 0.00 35.16 6.30 2 1 96.74

Abbreviations: N, number of similar analyses averaged; NG, number of separate grains involved in the average; nd, not determined.



almost as impact-smelted and porous as that of LAR04315, features little-shocked euhedral graphites andno olivine mosaicism.

5. By-products of impact smelting in LAR 04315 includediverse felsic glasses, mostly with 58–66 wt% SiO2 and19–25 wt% Al2O3. Together with a similar proportionof silica, these glasses typically occur as selvages that linethe walls of some of the largest voids in the meteorite.

6. The fine-grained ureilitic lithology of Almahata Sittaprobably also formed by a kind of impact smelting. Inplaces, this lithology retains the relict structure of a typ-ical ureilite, although the former coarse-equant olivineshave been granulated and infiltrated by a melt close toCa-rich pyroxene in composition, and the former pyrox-enes (consistently low-Ca) have been granulated andreduced to yield minor interstitial silica and Fe-metal.

7. The concentration of smelting into pigeonite in LAR04315 and LAP 03587 probably arose because the mate-rials were already very hot (anatectic) before the impact.Although the impact heating was only moderate, themelting temperature of the pigeonite was about 300degrees lower than that of the accompanying olivine.

8. The unusual aspects of LAR 04315 and LAP 03587thus confirm previous indications that the terminal epi-sode of smelting and rapid cooling was triggered whenthe parent ureilite asteroid underwent a catastrophicallydisruptive impact while its mantle was in a hot, anatec-tic condition.

9. The proportion of COx gas generated by ureilite smeltingexceeded by at least a small factor the volume repre-sented as porosity in the final ureilites; by a very largefactor (of order 103 or greater), if literature inferencesof extremely fast cooling rates, implying slight burialdepths, are accurate. The outflow of so much gas mayhave significantly enhanced the purging of the ureilitemelt component; and explosive expansion and jettingof the gas may have enhanced the thoroughness of thecatastrophic impact disruption of the parent asteroid.

ACKNOWLEDGMENTS

We thank Anonymous and C. Goodrich for helpful, construc-tive reviews; also Tim McCoy and John Wasson for helpful advice;and Peter Jenniskens, John Kashuba, and the Meteorite WorkingGroup for donations and loans of samples. This work wassupported by NASA grants NNG06GE86G (P.H.W.) andNNG06GF95G (A.E.R.).

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.gca.2010.05.026.

REFERENCES

Ahrens T. J. and Johnson M. L. (1995) Shock wave data for rocks.In Rock Physics and Phase Relations (ed. T. J. Ahrens).American Geophysical Union, pp. 35–44.

Berkley J. L. (1986) Four antarctic ureilites: petrology andobservations on ureilite petrogenesis. Meteoritics 21, 169–187.

Berkley J. L. and Jones J. H. (1982) Primary igneous carbon inureilites: petrological implications. Proc. Lunar Planet. Sci.

Conf. 13, A353–A364.

Berkley J. L., Taylor G. J., Keil K., Harlow G. E. and Prinz M.(1980) The nature and origin of ureilites. Geochim. Cosmochim.

Acta 44, 1579–1597.

Bischoff A., Goodrich C. A. and Grund T. (1999) Shock-inducedorigin of diamonds in ureilites. Lunar Planet. Sci. 30 (abstr. no.

1100).

Bischoff A., Horstmann M., Laubenstein M. and Haberer S. (2010)Asteroid 2008 TC3 – Almahata Sitta: not only a ureiliticmeteorite, but a breccia containing many different achondriticand chondritic lithologies. Lunar Planet. Sci. 41 (abstr. no.

1763).

Bowen N. L., Schairer J. F. and Posnjak E. (1933) The systemCaO–FeO–SiO2. Am. J. Sci. 26, 193–284.

Britt D. T., Macke R. J., Kiefer W. and Consolmagno G. J. (2010)An overview of achondrite density, porosity and magneticsusceptibility. Lunar Planet. Sci. 41 (abstr. no. 1869).

Carlson W. D. (1988) Subsolidus phase equilibria on the forsterite-saturated join Mg2Si2O6–CaMgSi2O6 at atmospheric pressure.Am. Mineral. 73, 232–241.

Carpenter R. L., Edgar A. D. and Thibault Y. (2002) Origin ofspongy textures in clinopyroxene and spinel from mantlexenoliths, Hessian Depression, Germany. Min. Pet. 74, 149–

162.

Chen M. and El Goresy A. (2000) The nature of maskelynite inshocked meteorites: not diaplectic glass but a glass quenchedfrom shock-induced dense melt at high pressures. Earth Planet.

Sci. Lett. 179, 489–502.

Clayton R. N. and Mayeda T. K. (1988) Formation of ureilites bynebular processes. Geochim. Cosmochim. Acta 52, 1313–1318.

Corrigan C. and McCoy T. (2005) LAP 03587 thin sectiondescription. Antarct. Meteorite Newslett. 28(1), 21.

Downes H., Mittlefehldt D. W., Kita N. and Valley J. W. (2008)Evidence from polymict ureilite meteorites for a disruptedand re-accreted single ureilite parent asteroid gardened byseveral distinct impactors. Geochim. Cosmochim. Acta 72,

4825–4844.

Estep P. A., Kovach J. J., Waldstein P. and Karr, Jr., C. (1972)Infrared and Raman spectroscopic studies of structural varia-tions in minerals from Apollo 11, 12, 14, and 15 samples. Proc.

Lunar Sci. Conf. 3, 3047–3067.

Fei Y. (1995) Thermal expansion. In Mineral Physics and Crystal-

lography (ed. T. J. Ahrens). American Geophysical Union, pp.

29–44.

Fritz J., Greshake A. and Stoffler D. (2005) Micro-Ramanspectroscopy of plagioclase and maskelynite in martian mete-orites: evidence of progressive shock metamorphism. Ant. Met.

Res. 18, 96–116.

Goodrich C. A. and Harlow G. E. (2001) Knorringite-uvavoritegarnet and Cr-Eskola pyroxene in ureilite LEW 88774 (abstr.).Meteor. Planet. Sci. 36, A68.

Goodrich C. A., Fioretti A. M., Tribaudino M. and Molin G.(2001) Primary trapped melt inclusions in olivine in the olivine-augite-orthopyroxene ureilite Hughes 009. Geochim. Cosmo-

chim. Acta 65, 621–652.

Goodrich C. A., Jones J. H. and Berkley J. L. (1987) Origin andevolution of the ureilite parent magmas: multi-stage igneousactivity on a large parent body. Geochim. Cosmochim. Acta 51,

2255–2273.

Goodrich C. A., Scott E. R. D. and Fioretti A. M. (2004) Ureiliticbreccias: clues to the petrologic structure and impact disruptionof the ureilite parent asteroid. Chem. Erde 64, 283–327.



Goodrich C. A., Van Orman J. A., Domanik K. and Berkley J. L.(2009) Metal in ureilites: petrologic characterization. Lunar

Planet. Sci. 40 (abstr. no. 1132).

Goodrich C. A., Wlotzka F., Ross D. K. and Bartoschewitz R.(2006) NWA 1500: plagioclase-bearing monomict ureilite orungrouped achondrite? Meteor. Planet. Sci. 41, 925–952.

Guan Y. and Crozaz G. (2000) Light rare earth enrichments inureilites: a detailed ion microprobe study. Meteor. Planet. Sci.

35, 131–144.

Hanson B. and Jones J. H. (1998) The systematics of Cr3+ and Cr2+

partitioning between olivine and liquid in the presence of spinel.Am. Mineral. 83, 669–684.

He H., Jin X., Jing F. and Ahrens T. J. (1996) Characteristic ofdynamic tensile fracture in augite-peridotite. In Shock Com-

pression of Condensed Matter (eds. S. C. Schmidt and W. C.Tao). AIP Press, pp. 593–596.

Hezel D. C., Palme H., Nasdala L. and Brenker F. E. (2006) Originof SiO2-rich components in ordinary chondrites. Geochim.

Cosmochim. Acta 70, 1548–1564.

Hirschmann M. M., Baker M. B. and Stolper E. M. (1998) Theeffect of alkalis on the silica content of mantle-derived melts.Geochim. Cosmochim. Acta 62, 883–902.

Hood L. L. and Jones J. H. (1987) Geophysical constraints onlunar bulk composition and structure: a reassessment. Proc.

Lunar Planet. Sci. Conf. 17, E396–E410.

Ikeda Y. (1999) Petrology of the Asuka 881391 ureilite with specialreference to crystallization of interstitial melt. Meteor. Planet.

Sci. 34, 625–636.

Janots E., Gnos E., Hofmann B. A., Greenwood R. C., Franchi I.A. and Bischoff A. (2009) Jiddat al Harasis 422: the first ureiliticimpact melt breccia. Meteor. Planet. Sci. 44 (abstr. no. 5197).

Jenniskens P. et al. (2009) The impact and recovery of asteroid2008 TC3. Nature 458, 485–488.

Johnson J. R., Horz F., Lucey P. G. and Christensen P. R. (2002)Thermal infrared spectroscopy of experimentally shockedanorthosite and pyroxenite: implications for remote sensing ofMars. J. Geophys. Res. Planets 107, JE001517.

Kita N. T., Ikeda Y., Togashi S., Liu Y., Morishita Y. andWeisberg M. K. (2004) Origin of ureilites inferred from a SIMSoxygen isotopic and trace element study of clasts in the Dar alGani 319 polymict ureilite. Geochim. Cosmochim. Acta 68,

4213–4235.

Matsuda J., Jusumi A., Yajima H. and Syono Y. (1995) Noble gasstudies in diamonds synthesized by shock loading in thelaboratory and their implications on the origin of diamondsin ureilites. Geochim. Cosmochim. Acta 59, 4939–4949.

McCoy T. (2005) LAR 04315 thin section description. Antarct.

Meteorite Newslett. 28(2), 16.

Mittlefehldt D. W. and Hudon P. (2004) Northwest Africa (NWA)1500: Not a basaltic ureilite; not even a ureilite. Meteor. Planet.

Sci. 39 (abstr. no. 5086).

Mittlefehldt D. W., McCoy T. J., Goodrich C. A. and Kracher A.(1998) Non-chondritic meteorites from asteroidal bodies. InPlanetary Materials, vol. 36 (ed. J. J. Papike), pp. 4.1–4.195.Reviews in Mineralogy. Mineralogical Society of America.

Miyamoto M., Takeda H. and Toyoda H. (1985) Cooling historyof some Antarctic ureilites. Proc. Lunar Planet. Sci. Conf. 16,

D116–D122.

Nakamuta Y. and Aoki Y. (2000) Mineralogical evidence for theorigin of diamond in ureilites. Meteor. Planet. Sci. 35, 487–493.

Ogata H., Mori H. and Takeda H. (1991) Mineralogy of interstitialrim materials of the Y74123 and Y790981 ureilites and theirorigin. Meteoritics 26, 195–201.

Presnall D. C. (1995) Phase diagrams of Earth-forming minerals. InMineral Physics and Crystallography (ed. T. J. Ahrens).American Geophysical Union, pp. 248–268.

Rubin A. E. (1983) The Adhi Kot breccia and implications for theorigin of chondrules and silica-rich clasts in enstatite chon-drites. Earth Planet. Sci. Lett. 64, 201–212.

Rubin A. E. (2006) Shock, post-shock annealing, and post-annealing shock in ureilites. Meteor. Planet. Sci. 41, 3–153.

Saito J. and Takeda H. (1990) Mineralogy of five new Antarcticureilites, LEW86216, LEW85328, Y-791839, Y-75154, Y-8448,and the origin of their chemical variations of pyroxene. Proc.

NIPR Sympos. Antarctic Met. (Tokyo) 3, 132–146.

Sazanova L. V., Milyavskii V. V., Borodina T. I., Sokolov S. N.and Zhuk A. Z. (2007) Shock metamorphism of plagioclase andamphibole (experimental data). Izvestiya Phys. Solid Earth 43,

707–712.

Schiano P., Bourdon B., Clocchiatti R., Massare D., Varela M. E.and Bottinga Y. (1998) Low-degree partial melting trendsrecorded in upper mantle minerals. Earth Planet. Sci. Lett. 160,

537–550.

Scott E. R. D., Taylor J. G. and Keil K. (1993) Origin of ureilitemeteorites and implications for planetary accretion. Geophys.

Res. Lett. 20, 415–418.

Sharp T. G. and DeCarli P. S. (2006) Shock effects inmeteorites. In Meteorites and the Early Solar System II (eds. D.S. Lauretta and H. Y. McSween). University of Arizona, pp.

653–677.

Singletary S. J. and Grove T. L. (2003) Early petrogenetic processeson the ureilite parent body. Meteor. Planet. Sci. 38, 95–108.

Speziale S., Jiang F. and Duffy T. S. (2005) Compositionaldependence of the elastic wave velocities of garnets, olivinesand other mantle minerals at high pressures. In Earth’s Deep

Mantle: Structure, Composition, and Evolution (eds. R. van derHilst, J. D. Bass, J. Trampert and J. Matas). American

Geophysical Union, pp. 301–320.

Stoffler D., Keil K. and Scott E. R. D. (1955) Shock metamorphismof ordinary chondrites. Geochim. Cosmochim. Acta 55, 3845–

3867.

Steele A., Fries M., Jenniskens P. and Zolensky M. (2009)Characterisation of diamond in the Almahata Sita meteorite.41st DPS Meet. (abstr. no. 9.08).

Takeda H. (1987) Mineralogy of Antarctic ureilites and a workinghypothesis for their origin and evolution. Earth Planet. Sci.

Lett. 81, 358–370.

Takeda H., Mori H. and Tachikawa O. (1985) Cloudy pigeonite inYamato 790981 with reference to the origin of chemical trendsof ureilites (abstr.). Lunar Planet. Sci. 16, 841–842.

Thrush P. W., ed. (1968) A Dictionary of Mining, Mineral, and

Related Terms. U.S. Department of the Interior, Bureau ofMines, U.S. Government Printing Office, 1269 p.

Treiman A. H. and Berkley J. L. (1994) Igneous petrology of thenew ureilites Nova 001 and Nullarbor 010. Meteoritics 29, 843–

848.

Tribaudino M. (2006) Microtextures and crystal chemistry ofpigeonite in the ureilites ALHA77257, RKPA80239, Y-791538,and ALHA81101. Meteor. Planet. Sci. 41, 979–988.

Warren P. H. (2004) Complex, contrasting behavior of chromiumduring late-stage processes in ureilites. Lunar Planet. Sci. 35

(abstr. no. 2151).

Warren P. H. (2010) Ejecta-megaregolith accumulation onplanetesimals and large asteroids. Meteor. Planet. Sci. 45,in press.

Warren P. H. and Huber H. (2006a) Chromium-silicates, feldsparsand highly silicic glasses formed from felsic melts in post-depressurization ferroan ureilites, especially LEW 88774. Lunar

Planet. Sci. 27 (abstr. no. 2400).

Warren P. H. and Huber H. (2006b) Ureilite petrogenesis: a limitedrole for smelting during anatexis and catastrophic disruption.Meteor. Planet. Sci. 41, 835–849.



Warren P. H. and Kallemeyn G. W. (1992) Explosive volcanismand the graphite–oxygen fugacity buffer on the parent aster-oid(s) of the ureilite meteorites. Icarus 100, 110–126.

Warren P. H. and Rubin A. E. (2006) Melted-reduced pigeonite,vermicular silica, Si-rich glasses, and other impact-smeltingproducts in the LAR 04315 ureilite. Meteor. Planet. Sci. 41,

A184.

Warren P. H., Ulff-M/ller F., Huber H. and Kallemeyn G. W.(2006) Siderophile geochemistry of ureilites: a record of earlystages of planetesimal core formation. Geochim. Cosmochim.

Acta 70, 2104–2126.

Weber I., Bischoff A. and Weber D. (2003) TEM investigations onthe monomict ureilites Jalanash and Hammadah al Hamra 064.Meteor. Planet. Sci. 38, 145–156.

Wiik H. B. (1972) The chemical composition of the Haverometeorite and the genesis of ureilites. Meteoritics 7, 553–

557.

Wilson L., Goodrich C. A. and Van Orman J. A. (2008) Thermalevolution and physics of melt extraction on the ureilite parentbody. Geochim. Cosmochim. Acta 72, 6154–6176.

Wittke J. H., Bunch T. E. and Goodrich C. A. (2007) Classificationof secondary reduction textures in ureilites. Meteor. Planet. Sci.

42 (abstr. no. 5246).

Yamaguchi A. and Sekine T. (2000) Monomineralic mobilizationof plagioclase by shock: an experimental study. Earth Planet.

Sci. Lett. 175, 289–296.

Yanai K., Kojima H., Koeberl C., Graham A. L. and Prinz M.(1987) Photographic Catalog of the Antarctic Meteorites.National Inst. Polar Research, 298 p.

Zhang Y., Benoit P. and Sears D. W. G. (1995) The classificationand complex thermal history of the enstatite chondrites. J.

Geophys. Res. Planets 100, 9417–9438.

Zhu Y. (2008) K- and Si-rich glasses in harzburgite fromDamaping, north China. Island Arc 17, 560–576.

Ziegler K., Young E. D., Schauble E. A. and Wasson J. T. (2009)Silicon isotope fractionation between silicate and metal from anenstatite meteorite: implications for earth’s core formation.Meteor. Planet. Sci. 44 (abstr. no. 5435).

Associate editor: W. Uwe Reimold


Documents

Author's personal copy - UCLAcosmochemists.igpp.ucla.edu/2010-1.pdf · Author's personal copy Pyroxene-selective impact smelting in ureilites ... Received 1 March 2010; accepted in