another feldspathic fragmental breccia, 67975 [Marvin and. Warren, 1980 .... responsible for volcanic mare glasses (Butler and Meyer, 1976; Hughes et al., ...
GEOPHYSICAL RESEARCH LETTERS, VOL. 18,NO. 11,PAGES2081-2084, NOVEMBER 1991
ADDITIONAL
COMPLEXITY
ANORTHOSITES
IN THE LUNAR
AND SULFUR-RICH,
CRUST: PETROLOGY
OF SODIC
FERROAN NORITIC ANORTHOSITES
Marc D. Norman,G. JeffreyTaylor, andKlaus Keil
PlanetaryGeosciences, Departmentof GeologyandGeophysics, SOEST,Universityof Hawaii Abstract
Lunar
breccia
67016
contains
two
suites of
unusualhighlandsrocks:sulfur-rich,noritic anorthosites, and sodic anorthosites.Mineral compositionsof the sulfur-rich noritic anorthosites closely match those of ferroan anorthosites. Unusually large amounts of sulfides, and sulfidationreactionsinvolvingolivinein theseclastsprobably reflectendogenous volatiletransfer.The sodicanorthosites are mafic-poor,with mineralcompositions intermediatebetween thoseof pristineferroanandalkalianorthosites. Rareexamples of pristinerockswith similarmineralcompositions areknown, andmay representa distinctclassof lunarcrustalrocks. Introduction
An understandingof how the lunar crustformedrequires knowledgeof thecompositions anddistributions of rock types within the crust.Even after decadesof study,new rock types continueto be discovered,primarily throughstudiesof clasts from brecciasandsoils.This paperdescribesthepetrologyof two unusualtypes of anorthositicclastsfrom an Apollo 16 breccia,andbriefly discusses theirsignificance. 67016 is a typicalfeldspathicfragmentalbrecciafrom the rim of North Ray Crater. It is composedof abundantmelt brecciaandanorthositic clastsin a plagioclase-rich matrix.The diversepopulationof anorthositicclastsin this brecciareflects the complexcrustal stratigraphyexcavatedby the Nectaris impactevent [Norman, 1981; SpudisandPieters,1991]. Two groupsof anorthositic clastsrepresentrocktypesthathavenot beenwidelyrecognized in thelunarsamplecollection.Oneis a ferroannoriticanorthosite with mineralcompositions matching pristine ferroan anorthosites,unusually large amounts of troilite, and bulk compositionsvery similar to that of the averagelunar crust[NormanandTaylor, 1991]. The otheris a sodic anorthositewith mineral compositionsintermediate betweenferroananorthosites andthe Apollo 12 and 14 alkali anorthosites. The ferroan noritic anorthosites were included in
the 'plutonic'groupof Norman [1981]. The sodicanorthosites havenot beenrecognizedby previousstudiesof thisbreccia. The chemicalcompositions of thesetwo clasttypessuggest they havenot beencontaminated with significantamountsof regolith,meteoriticmaterialor KREEP [NormanandTaylor, 1991]. Therefore careful petrologicalstudy may be able to deciphersomesignatureof their igneousprecursors. Here we use major and minor elementmineral compositionsto assess further the petrologic affinities of these clasts. Mineral compositions were determinedby electronmicroprobeusing 15 kV accelerating potential,15-18 nA beamcurrentand20-30 second count times. Feldspars were analyzed using a
Copyright1991by the AmericanGeophysical Union.
defocusedbeam-•15 gm in diameter;analysesof maficphases useda focusedbeam.Full ZAF correctionswere appliedto the data.
Petrography andMineralCompositions Ferroan
noritic anorthosites
Ferroan noritic clasts have well-preserved crystalline textures that appear intermediate between igneous and metamorphic.Plagioclasegrainsare oftenirregularin shape, but relatively large (up to 2 mm long) euhedralgrainscan be found.Pyroxeneoccursas granularto tabulargrainsinterstitial to plagioclase.Olivine is very rare, occurringonly as relics within sulfide+pyroxene intergrowths.Theseclastshavebeen only lightly shocked:plagioclasedisplaysstraightto slightly wavy extinction and shocklamellae are rare to absentin the pyroxenes.The texture is suggestiveof a formerly coarsegrainedrock whichwasbrecciatedandsubsequently annealed nearthe solidus,or possiblyinfiltratedby melt. Mineral compositions in these clasts are very homogeneous. Plagioclaserangesfrom An94-97,with a strong mode at An96. Low-Ca pyroxene (Wo2En55-62) is the predominantmafic mineral, but augite (Wo43-46En38-40)is common.The few olivines we could analyze are Fo58-61. Sulfide,ilmenite,andchromiteareaccessory phases.Fe-metal appearsto be absent. An unusual feature
of these clasts is their abundance
of
troilite, which occurs(1) as fine-grainedintergrowthswith low-Ca pyroxene,(2) aspuretroilite veinsalongcracks,grain boundaries, and cleavage fractures, and (3) as discrete inclusionsor interstitialgrains.Sulfur-richvaporsappearto haveinvadedtheseclasts,reactingwith theolivineto produce the sulfide+pyroxeneintergrowths,and depositingtroilite alonggrainboundaries.A smallamountof thetroilite may be primary. Sodic anorthosites
We have discoveredan unusually sodic variety of anorthositein 67016 [Normanand Taylor, 1991]. The sodic
anorthositic clastsare all brecciatedbut only moderately shocked. Two of theseclastsareexceptional because they containmorehigh-Capyroxenethanlow-Capyroxene.Most original grain boundarieshave been destroyed,but some relativelylargegrains(e.g.,pyroxenes up to 0.8 mm long) remain,and someregionsappearto havebeencrushedmore or lessin situ.Plagioclase compositions rangefromAn91-94, averagingAn92.High-Capyroxene(Wo31-45En36-53), low-Ca pyroxene(Wo2-8En49-66), andolivine (Fo60-63)were found.
Paper number91GL02617
Oneclastretainsa moderately coarsegrainsize,with augite containing exsolution lamellaeof low-Capyroxeneup to -•25
0094-8534/91/91GL-02617503.00
gm wide. Plagioclaseoften containsorientedrodsof an oxide 2081
2082
Normanetal.: Petrologyof UnusualHighlandsRocks 1.2
mineral (probablyilmenite) which we interpretas exsolution lamellae. Minor phasesinclude ilmenite and chromite;metal and sulfideare absentor very rare.
Low-Ca pyroxene 1.0
T
N
Discussion 0.8
Comparisonwithpristine rocks:mineral compositions
Considerableprogresstowardunderstanding the lunarcrust was achievedwith the recognitionthat certaingroupsof rocks apparently retain coherent petrological and geochemical characteristicsacquiredduring endogenouslunar magmatic processes. Because these 'pristine' rocks (Warren and Wasson,1977) largelyescapedimpactmixing, theyprovidea
TiO2 wt%
G
0.6
0.4 F
0.2
useful focus for discussionsof the lunar crust. In this section,
we compare the major and minor element mineral compositionsin the 67016 noritic and sodicanorthositeswith thoseof knownpristinerocks. Co-existing plagioclaseand mafic silicate compositions (Figure 1) in the 67016 ferroannorific anorthositesillustrate the restrictedcompositionalrangeof mineralsin theseclasts, and their close mineralogical similarity to pristine ferroan anorthosites (FAN). The abundance and compositions of mafic minerals in these clasts suggestaffinities with the "relatively mafic, relatively ferroan" subgroup of FAN recognizedby Jameset al. [ 1989]. Mineral compositionsof
0.0
ß
,
,
0.0
,
I
,
,
,
ß
0.5
I
,
,
1.0
,
,
i
,
1.5
,
,
2.0
FeO/MgO
Figure2. FeO/MgO vs. TiO2 contentsof low-Ca pyroxenesin 67016 anorthositicclastscomparedto thosein knownpristine rocks. Symbols same as on Figure 1. Fields shownare for pristine troctolites (T), norites (N), gabbronorites(G), and ferroananorthosites (F). Data for pristinerocksfrom Berschet al. [1991].
the sodic anorthosites fall well outside the fields for known
pristinerocks(Figure 1). Althoughthe mafic mineralsin the sodic anorthositesare somewhat ferroan, their plagioclase compositionsare significantlymore sodicthan that of FAN, but not as sodicas the Apollo 12 and 14 alkali anorthosites. Similar mineral compositionswere found in the 15459 Fenoritc clasts[Lindstrometal., 1989] and in the augitc-rich anorthositicfragmentfrom 76504 [Warrenetal., 1986]. A potentiallyusefulbutunder-utilizedsourceof information is the minor element abundancesof mineral phasesin the pristine rocks. An extensive set of minor and trace element abundancedata for pyroxenes and olivines from known pristinerockshas been presentedby Berschet al. [1991].
Here,we usethesedataasdiscriminants to evaluatepossible petrogeneticaffinities of the 67016 noritic anorthositesand sodic anorthosites.
The Ti, Cr, and A1 contentsof low-Ca and high-Ca pyroxenesprovide an effective discriminant between ferroan
anorthositesand Mg-suite rocks (Figures 2, 3). Pristine gabbronorites aredistinguished fromboththeMg-suitenorites and the ferroananorthosites by their higherabundances of augitc,moreferroanpyroxenecompositions, and slightly more sodic plagioclase[Jamesetal., 1987]. Minor element
abundances of pyroxenesin gabbronorites typically are intermediate between those of ferroan anorthosites andtheMgsuiterocks(Figures2, 3).
100
1.0
clasts
67016 /•
High-Ca pyroxene
Mg-suite
80
En
Troctolites, norites/
•
I
'•
50 40 75
15459c
_
Cr203
•
opx 70 60
0.8 wt%
0.6
I ß /
76504,18 X•
0.4
AlkalianorthositesFerroan\•/
0.2
anorthosites
80
85
90
95
100 10
An plagioclase
0.0
Figure 1. A plot of averageplagioclaseandlow-Ca pyroxene
(opx)compositions forthe67016ferro •annoritic anorthosites (dots)andsodicanorthosites (squares),comparedto fieldsfor knownpristinerock groupsandthe76504 and 15459samples (X). Fieldsfor pristinerock groupsafterWarren etal. [ 1991].
ß ß ß ß i , , i , i , ,
0.5
1.0 TiO2
, , i , i i i
1.5
I i ß ß ß
2.0
2.5
wt %
Figure 3. TiO2 vs. Cr203 contentsof high-Capyroxenesin 67016 anorthositicclastscomparedto thosein knownpristine rocks.Symbolsand data sourcesare the sameas for Figure2.
Normanet al.: Petrologyof UnusualHighlandsRocks
Minor element concentrations in the 67016 ferroan noritic
anorthosites are very similarto thosein pristineFAN. This is truefor bothlow-Ca andhigh-Capyroxenes. Pyroxenes from the sodicanorthositeshave varied compositionswhich are broadly similar to those of FAN and the pristine gabbronorites.Low-Ca pyroxenesin the sodic anorthosites havethe low A1 andCr contentstypicalof FAN, buttheirTi contentsaremoresimilarto thoseof thegabbronorites. Minor elementcontentsof augitein theseclastsalsoresemblethosein FAN, but thereis considerable overlapbetweenthe fieldsfor FAN and gabbronorite.Augite in the sodicanorthosites does not show the broad range in Fe/Mg apparent in the gabbronorites, but doesincludeTi contentshigherthan those in FAN andmore similarto thoseof gabbronorites.
2083
large amount of sulfide present within the intergrowths (Roedder and Weiblen [1974] estimated 22% troilite in one
occurrence), and pyroxene compositions within the intergrowthsthat are more magnesianthan thosein the bulk sample.When the sulfide veins are also considered,it is clear that considerable
sulfur was introduced
into these rocks
followingtheircrystallization. Reconnaissance experimentson the sulfidationof olivine were discussedby Kullerud and Yoder (1963, 1964), and allow a preliminaryestimateof the conditionsnecessaryto form the sulfidesin the 67016 clasts.These experiments
suggestthat olivine reactsreadilywith S-bearingvapor,but thatvery largeproportions of sulfurto olivinearerequiredto avoid the formationof Fe-oxidesas reactionproducts.For example,at 800oc and 2000 bars,60 mole% sulfur (relativeto
Origin of sodicanorthosites
Pristinesampleswith mineralcompositions similarto those of the sodicanorthositeclastsin 67016 are rare. They include an anorthositicfragmentfrom 76504 [Warrenet al., 1986] and Fe-norite clasts in 15459 [Lindstrom et al., 1989]. Warren et
al. [1986] suggestedthat the sodic,augite-richanorthositein 76504 might be an evolvedmemberof the ferroananorthosite suite.Lindstromet al. [1989], however,proposedthat the Fenorites in 15459 probably were related to Mg-suite norites. Magnesianalkali gabbronorites suchas 61224 andthe clastsin another feldspathicfragmental breccia, 67975 [Marvin and Warren, 1980; James et al., 1987; Roedder and Weiblen,
1974], have pyroxenecompositions very similar to thosein our sodicanorthositicclasts,but the plagioclasein thesealkali gabbronoritestypically is more sodic (e.g., An82-89 in the magnesianalkali gabbronoriteclastsrs. An91-94in our sodic anorthosites). Our minor elementdata for pyroxenesin the 67016 sodic
anorthosites are somewhatequivocaland couldbe consistent with
affinities
with
either
ferroan
anorthosites
or
the
gabbronorites andFe-norites.Nonetheless thereappearsto be an overall similarityin mineralogicalcharacteristics between the 67016
sodic anorthosites, and the Fe-norites
and
gabbronorites. The predominance of augitein two of our sodic anorthositeclastssuggestsa kinshipwith gabbronorites.The somewhat variable mineral compositions of the sodic anorthositescontrast with the well equilibrated mineral compositionsof the ferroan noritic anorthositeclasts.This may be indicatingeithera fairly small,rapidlycooledbodyof magma or separate intrusions as the source of the sodic anorthosites.
Origin of sulfidesinferroan noritic anorthosites
olivine)were necessary to form only FeS, low-Ca pyroxene and SO2 asreactionproducts.At lower proportionsof sulfur to olivine, the reaction assemblageincluded abundant magnetite,hematite,andsilica,althoughoxygenfugacitywas not explicitlycontrolledin theseexperiments. Nonetheless, the 67016 intergrowthsprobablydid form in a high pS2/pO2 environment.Theseexperimentsalsofoundthattroilite formed at 800øC,whereaspyriteformedat 650oc. By analogywith these experiments, the presence of troilite in the noritic anorthositeswould be consistentwith the temperaturesof
---850øC which are indicated by co-existing pyroxene compositions in theseclasts.
Is thesulfurin theseclastsendogenous or exogenous? That is, wasit derivedby outgassing of thelunarinterior,or might the sulfurousvaporshave beendeliveredto theselunarcrustal rocks via the impact of a sulfur-rich meteorite? Enstatite chondrites are sulfur-rich, and themselves show evidence of
sulfidationof olivine to producesulfides,pyroxenes,and silica (Rubin, 1983, 1984; Fleet and MacRae, 1987). Sulfur mobilization associatedwith shock metamorphismwas documentedfor lunar regolith breccia 14315 by Ramdohr (1972). In 14315, the sulfur reacted with metallic Fe, ilmenite
andmafic silicates.Somesignificantdifferencesbetweenthe sulfidationobservedin 14315 and that presentin the 67016 noritic anorthosite clasts include: (1) in 14315, Ni also migratedwith the sulfur,formingpentlandite.Pentlanditehas not yet been observedin the 67016 clasts;neither are these clastsespeciallyNi-rich. (2) Sulfurmobilizationin 14315 was associatedwith significantshockdeformation,whereasthe 67016 clastsare only lightly shocked.(3) Reactionsin 14315 are incomplete,whereasin the 67016 claststhereactionshave gonevirtually to completion.(4) 14315 is a regolithbreccia, but 67016 contains only very minor amounts of regolith material.Lunar regolithcomponentsare known to be mobile
undernear-surface conditions. It is alS'6notablethat the The ferroan noritic anorthosite clasts are unusual for their
abundance of sulfidesandthe apparentreplacement of olivines by sulfidationreactions.Similar sulfide+silicateoccurrences have been observedin 67915 [Roedderand Weiblen, 1974] and 67075 (our unpublishedobservations),but no general explanationfor their origin has been offered. Roedderand Weiblen [1974] noted that the sulfide-rich intergrowthsin 67915 were associatedwith, and appearedto partiallyreplace olivines,and Norman [1981] suggested that the intergrowths formed by replacement of olivine by sulfide + low-Ca pyroxene. Evidence for a replacement origin for the intergrowthsinclude: an associationwith relic olivine, the
sulfidation
reactions in the ferroan noritie•anorthosite
clasts
were completedprior to brecciaformationhnddid not affect
otherclasts ormatrixolivine[Norman, 1•981]. The relativelylarge amountsof sulfurand apparentlyhigh temperatures inferredfor the sulfidationreactionsrecordedin the 67016 noritic anorthositeclastssuggestthat outgassingof
thelunarinterior maybe• moreattractive alternative. Mare basaltscontainlarge amountsof sulfurcomparedto terrestrial basalts(e.g., up to 2000 ppm in typical Apollo 17 high-Ti basalts),and significant amountsof sulfur apparentlywere released during the pyroclastic eruptions responsible for volcanicmare glasses(Butler and Meyer, 1976; Hugheset al.,
2084
Normanet al.: Petrology of UnusualHighlands Rocks
1989).It maybethatemplacement of magmas richin volatile compoundsinto the lunar crustor near the baseof the crust provided both the sulfur for the olivine breakdown,and the heat for themetamorphismof the ferroannoritic anorthosite clasts in 67016.
Kullerud G. and H.S. Yoder
Jr., Sulfide-silicate reactions,
Carnegie Institution of WashingtonYearbook, 218-222, 1964.
Lindstrom M.M.,
U.B. Marvin,
and D.W.
Middlefehldt,
Apollo 15 Mg- and Fe-Norites:a redefinitionof the Mgsuite differentiation trend, Proc. 19th Lunar Planet. Sci. Conclusions
Conf., 245-254, 1989. The Lunar and PlanetaryInstitute, Houston.
PetrolOgiccharacteristicsof anorthositicclastsfrom lunar breccia67016 provide evidencefor lunar crustalrock types which have not beenwidely recognized.Noritic anorthosites havemajor and minor elementmineralcompositions closely matching those of ferroan anorthosites,but are enrichedin
mafic phases.These clasts also contain unusually large amountsof sulfidesthat probablyreflect endogenous volatile transfer. Augite-bearing, sodic anorthositeshave mineral compositions outsidethe rangeof commonpristinerocks,and intermediate between those of ferroan and alkali anorthosites.
Rare examples of pristine rocks with similar mineral compositionshave been reportedpreviously.These samples may representa new classof lunarhighlandsrocks.
AcknowledgmentsWe thank Graham Ryder for a useful review, and Tom Hulseboschfor expertassistance with the electronmicroprobe.This work was supportedin part by the NationalAeronauticsandSpaceAdministration,grantNAG9454 to K. Keil. This is PlanetaryGeosciences PublicationNo. 668, and SOEST Publication No. 2679.
Marvin U.B. and P.H. Warren, A pristineeucrite-likegabbro from Descartes and its exotic kindred, Proc. Lunar Planet
Sci. Conf., 11,507-521, 1980.
Marvin U.B. andP.H. Warren,A pristineeucrite-likegabbro from Descartes and its exotic kindred, Proc. Lunar Planet
Sci. Conf., 11, 507-521, 1980. Norman M.D., Petrology of suevitic lunar breccia 67016, Proc. Lunar Planet. Sci. Conf., 12, 235-252, 1981. Norman M.D. and S.R. Taylor, Geochemistryof plutonic anorthositesfrom Apollo 16 breccia67016: lunar crustal evolution and the compositionof the Moon, Geochim. Cosmochim.Acta, in press,1991. RamdohrP., Lunar pentlandireand sulfidizationreactionsin microbreccia 14315,9, Earth and Planet. Sci. Lett., 15, 113-115, 1972.
RoedderE. and P.W. Weiblen, Petrologyof clastsin lunar breccia67915, Proc. Lunar Sci. Conf., 5, 303-318, 1974. Rubin A.E., The Adhi Kot brecciaand implicationsfor the origin of chondrulesand silica-rich clasts in enstatite chondrites, Earth and Planet. Sci. Lett., 64, 201-212, 1983.
References
Bersch M.G., Taylor G.J., Keil K., and Norman M.D., Mineralcompositions in pristinelunarhighlandrocksand the diversity of highland magmatism.Geophysical Research Letters, this volume.
ButlerP., Jr. andC. MeyerJr.,Sulfurprevailsin coatings on glassdroplets:Apollo 15 greenand brownglassesand Apollo 17 orangeandblack(devitrified)glasses, Proc. Lunar Sci. Conf., 7, 1561-1581, 1976.
Rubin A.E., The Blithfield meteoriteand the origin sulfiderich, metal-poorclastsandinclusionsin brecciatedenstatite chondrites, Earth and Planet. Sci. Lett., 67, 273-283, 1984.
SpudisP. and C. Pieters,Global and regionaldata aboutthe Moon, in Lunar Sourcebook,editedby G. Heiken, D.T. Vaniman, and B. French,p. 595-632, 1991. Cambridge University Press,New York. Warren P.H. and J.T Wasson, Pristine nonmare rocks and
the natureof the lunar crust,Proc. Lunar Sci. Conf., 8, 2215-2235, 1977.
HughesS.S., J.W. Delano,and R.A. Schmitt,Petrogenetic Warren P.H., D.N Shirley, and G.W. Kallemeyn, A modelingof 74220high-Tiorangevolcanicglasses andthe potpourriof pristineMoon rocks,includinga VI-tK mare
Apolloi 1 and17high-Timarebasalts, Proc.19thLunar
Planet. Sci. Conf. 175-188, 1989.
FleetM.E. and N.D. MacRae,Sulfidation of Mg-richolivine and the stability of niningerite in enstatitechondrites,
Geochi m. Cosmochim. Acta,51, 1511-1521,1987.
James O.B.,M.M. Lindstrom, andM.K. Flohr,Petrology andgeochemistry of alkaligabbronorites fromlunar
basalt and a unique augite-richApollo 17 anorthosite, Proc. Lunar Planet. Sci. Conf., 16, J. Geophys.Res., 91, Suppl., D319-D330, 1986. Warren P.H., E.A. Jerde, and G.W. Kallemeyn, Pristine Moon rocks:Apollo 17 anorthosites, Proc. Lunar Planet. Sci., 21, 51-61, 1991. The Lunar andPlanetaryInstitute, Houston.
brecCia 67975, Proc. Lunar Planet. Sci. Conf., 17, J.
GeoPhy s. Res.,92, Suppl.,E314-330,1987. James,o.B., M.M. Lindstrom, andM.K. Flohr,Ferroan anorthoSite from lunarbreccia64435:implications for the originandhistoryof lunarferroananorthosites. Proc. 19th Lunar Planet. Sci. Conf., 219-243. Kullerud G. and H.S. Yoder Jr., Sulfide-silicatereactions, CarnegieInstitutionof WashingtonYearbook,215-218, 1963.
M.D. Norman, G.J. Taylor and K. Keil, Planetary Geosciences,Dept. of Geology and Geophysics,SOEST, 2525 Correa Road, Honolulu, HI 96822 U.S.A.
(Received: May 23, 1991; revised: October 10, 1991;
accepted:October11, 1991)