Evolution, Lunar: From Magma Ocean to Crust Formation

1 EEvolution, Lunar: From MagmaOcean to Crust FormationJuliane Gross1,2,3and Katherine H. Joy41 Department of Earth and Planetary Sciences,Rutgers University, Piscataway, NJ, USA2 American Museum of Natural History, NewYork, NY, USA3 Lunar and Planetary Institute, Houston, TX,USA4 School of Earth and Environmental Sciences,University of Manchester, Manchester, UKIntroductionThe lunar Crust provides a record of the planetaryformation and early evolutionary processes andcontains a wealth of information about the originand evolution of the Earth-Moon system ( ,Taylor1982; NRC2007; Canup2008,2012;Cuk and Stewart2012; Young et ). Under-standing these processes is crucial for the recon-struction of the early evolutionary stages of theEarth, , the early geological evolution of aterrestrial planet, inner Solar System impact bom-bardment, and the solar and galactic environmentthroughout the last billion years (Ga) ( ,NRC2007; Crawford et ).

2 Our knowledge of the lunar highland Crust hasadvanced enormously. Studies of lunar meteor-ites; experimental and computational studies;remote sensing of mineralogy, chemistry, andtopography of the lunar surface; new gravitydata; geochronology; geochemistry, especiallyisotopic constraints; and the abundances andsource reservoirs of lunar volatiles have brought,and continue to bring, valuable insights to under-standing the Moon s geological evolution ( ,Shearer et ; Elkins-Tanton et ;Elardo et ; Zuber et ; Wieczoreket ; Borg et ).The Lunar Magma Ocean ParadigmThe early history and broad-scale petrogenesis ofthe Moon were glimpsed from thefirst mannedand unmanned missions to the Moon that returned 382 kg of lunar rocks and soils from the lunarsurface (Vaniman et ).

3 These were col-lected by the Apollo and Luna missions from thecentral and eastern nearside of the Moon ( ).Based on numerous studies of these returnedApollo and Luna samples, hypotheses of theMoon s differentiation and Crust Formation weremade. The most supported model for the forma-tion and evolution of the Moon is that of a giantimpact between proto-Earth and another largeplanetary body early in the Solar System history( , Canup2008,2012; Canup et ; Cukand Stewart2012), chemically mixing the twobodies ( , Young et ). This produced alunar Magma Ocean (LMO), although the globalextent of this Ocean and depth of melting thatoccurred is debated (Wood et ; Smithet ; Solomon1986; Binder1986;#Springer International Publishing AG 2016B.)

4 Cudnik (ed.),Encyclopedia of Lunar Science,DOI , Lunar: From Magma Ocean to Crust For-mation, Fig. 1 Clementine mission albedo map of theMoon in a cylindrical projection, overlain with locations ofthe Apollo sample return missions (as indicated bycrosses). (a) Map is overlain with the distribution andconcentration of regions with>2 ppm Th as mapped bythe Lunar Prospector mission (2 per pixel data calibration;Prettyman et ) showing surface expression ofKREEP-bearing lithologies. Major crustal terranes asmapped by Jolliff et al. (2000) are denoted as South Pole-Aitken (SPA) basin, feldspathic highland terrane (FHT)and Procellarum KREEP Terrane (PKT).

5 (b) Map is over-lain with the distribution and concentration of regions with< ppm Th (cyan-colored pixels) and< ppm Th(dark blue-colored pixels) (Prettyman et ). Thethreshold< ppm Th is taken as regolith with no or apositive Eu-anomalies (O Hara and Nui2015). It is impor-tant to note that the gamma-ray spectrometer data samplethe upper 30 cm of the lunar regolith, and mapped regolithswill include a wide mix of igneous rocks, reworked impactmelt breccias, and impact glass over a scale of 60 km perpixel, so we also highlight in with< ppm Th, taken as equivalent to significant positiveEu-anomalies akin to Apollo igneous FAN rocks(< ppm Th, Taylor2009), bearing in mind the averageuncertainty of the Lunar Prospector dataset (Prettymanet ).

6 An interesting point to note is that none ofthe average Apollo landing site soils have positive Eu-anomalies as they all contain soil components rich innegative Eu-anomalies-bearing KREEPy mafic impactmelt breccias and/or mafic mare basalt material2 Evolution, Lunar: From Magma Ocean to Crust FormationPritchard and Stevenson2000; Shearer et refs. therein; Delano2009).As this LMO cooled, crystals that precipitatedfrom the melt separated due to density differences( ). This has been modeled ( , Solomonand Longhi1977; Dreibus et ; Longhi1980,2003; Tonks and Melosh1990; Snyderet ; Meyer et ; Elkins-Tantonet ) and experimentally tested (Elardoet ; Rapp and Draper2012,2013).

7 Thesemodels consider two end-member forms: (1) frac-tional crystallization from start tofinish (the one-stage model ) and (2) a two-stage model inwhich early equilibrium crystallization occurredfollowed by fractional crystallization of the resid-ual Magma Ocean . In all models, Mg-rich olivineis the liquidus phase, followed by orthopyroxenecrystallization. These dense mafic (Mg-rich) min-erals sank and formed the mantle, enriching theresidual Magma in iron and more incompatibleelements ( , Th, Ti, K).After 75 80 % of the LMO had solidified,the calc-end-member of plagioclase feldspar(anorthite) began to crystallize. Recent studiessuggest that initialization of plagioclase crystalli-zation occurred at depths of>75 km in the LMO(Nakvasil et ).

8 Upon Formation , thislow-density plagioclase overcame a density con-trast with dense comagmatically crystallizingmafic phases by buoyantly rising and migratingtoward the lunar surface, forming a thick onionshell anorthosite Crust (Warren1990; Elkins-Tanton et ) ( ). This crystal separationalso removed the plagiophile compatible elementEu from the Magma Ocean melt, providing theanorthositic Crust with a positive chondrite-normalized (cn) Eu-anomalies ( , Eu/Eu*>1,where Eu/Eu*=(Eucn/(Smcn+Gdcn)^ )) androcks that have low concentrations of incompati-ble trace elements. Analyses of Apollo returnedsamples show that this primary feldspathic Crust iscomposed of anorthositic rock with plagioclaseAn# (molar Ca/[Ca + Na + K]) of 94 98, withmafic minerals that have a Mg# (molar Mg/[Mg +Fe]) ranging from 40 to 70, although varietieswithin that range exist (Warren et ; Jameset ; Jolliff and Haskin1995; Nyquistet ) ( ).

9 These rocks are called theEvolution, Lunar: From Magma Ocean to Crust For-mation, Fig. 2 Schematic sketch of the (a) differentiationand (b) crystallization of the global lunar Magma Ocean (LMO). (b) The LMO crystallized mafic cumulates ofolivine and pyroxene crystals. These cumulates sank intothe interior to form the lunar mantle (Shearer et refs. therein; Elardo et ). (c) After 75 to 80 %of LMO crystallization, plagioclase began to crystallize(Snyder et ) andfloated to the top of the LMOproducing the global anorthositic Crust , the light-coloredhighlands seen on the lunar surface today. The residualmagma, trapped between the anorthositic Crust and theunderlying ultramafic mantle, continued to crystallize andlate-stage melt became increasingly enriched in KREE-P.

10 For more details, see text (Modified from the classroomillustration of Jennifer Rapp (2013) from the Lunar andPlanetary Institute; Center for Lunar Science and Explora- )Evolution, Lunar: From Magma Ocean to Crust Formation3ferroan anorthosite suite (FAN; ). FANs arecommon at the Apollo 16 highland landing siteand are also present at the mare regions visited bythe other Apollo missions (Wood et ).The development of a thick onion shell anor-thosite Crust would have likely caused a largethermal blanketing effect on the lunar interior(Shearer et ). The rate of cooling in theremaining molten lunar interior (be it the dregs ofa Magma Ocean , more localized melt environ-ments, or Moon-wide partial melts) would haverapidly slowed, and the melt itself would havebecome increasingly FeO-rich and dense(Warren1990) and crystallized to yield abundantilmenite (FeTiO3).