15
Lunar Meteorites 1
Lunar Meteorites
1
Lunar Meteorites
2
The first recognized lunar meteorite (1981).
Note fusion crust, brecciated nature, and presence of significant anorthite.
Lunar Meteorites
3
~98% of lunar rocks are composed of just a few minerals: olivine, pyroxene, plagioclase, and ilmenite
Can compare chemistry of a found meteorite directly to chemistry of Apollo samples!
~16% of mass returned by Apollo and Luna missions.
>170 individual stones
Somewhere between ~50-90 impact events?
Fast Facts
How Do We Know They Are Lunar?
Textural attributes and ages.
Lunar Meteorites
4
[Kalahari 008/009]
Lunar Meteorites
5
Kalahari 009 is the largest single stone (13.5 kg).It is a basaltic breccia with no anorthositic clasts.
Kalahari 008 is paired with 009 (found 50 m apart) and is 598 g.It is an anorthositic breccia with little evidence for basaltic clasts.
Similar exposure ages...only ~230 years!
Lunar Isotopes
6
Moon and Earth fall on same line....or do they?
Lunar Isotopes
7
high-precision data sets (8–10) reveals that lunarsamples are elevated, on average, by 9 ppm com-pared with the terrestrial UWG-2 standard usedin all three studies (fig. S1). With an improvedanalytical technique (19), we are able tomeasurevariations in parts per million (0.00x‰) in D17O(see supplementary materials and methods), al-lowing us to detect the isotopic differences be-tween Earth and the Moon.We first attempted to use lunar meteorites
to determine the isotope composition of theMoon. Our data show that terrestrial weatheringmodified the D17O of the studied lunar meteor-ites, making identification of small variationsimpossible (fig. S2). Therefore, we have ana-lyzed fresh basalt samples from three Apollolanding sites that were provided by NASA. Wecompared the composition of the lunar basaltswith the composition of Earth mantle xenolithsand mantle-derived melts [mid-ocean ridge basalt(MORB)] (Table 1 and Fig. 1).The bulk silicate Earth (BSE) is constrained
to a D17O value of –0.101 T 0.002‰ [1s SEM, n =65 measurements (19)] from mantle xenolithsand MORB from seven localities around theworld [for definitions and analytical details, see(19, 20)]. Earth mantle minerals and MORB fallon a common mass-fractionation line, with aslope of q = 0.532 T 0.006, in the d!17O versus d!18Ospace. This slope is, within uncertainty, iden-tical to the high-temperature approximation forequilibrium oxygen isotope fractionation of0.5305 (21). The D17O values of Earth mantleminerals and MORB are identical within theuncertainty. This is in agreement with massfractionation upon melt extraction along a slopeof ~0.53 for high-temperature processes (19, 21).Because the same slope must apply to meltextraction from the lunar mantle, we can safelyuse the lunar basalts as analogs for the bulksilicate Moon (BSM) with respect to D17O.The three lunar basalts span a small range in
d18O and have an averageD17O = –0.089 T 0.002‰(1s SEM, n = 20). We suggest that this value isrepresentative of the BSM. Thus, the D17O of theMoon is 12 T 3 ppm (0.012 T 0.003‰) higher thanthat of Earth (Fig. 2). This unequivocally identifiesan isotopic difference between Earth and the Moonand supports the view that the Moon formed bya giant collision of the proto-Earth with Theia.The reevaluation of lunar O isotope data from
three previous studies (8–10) is consistent withour finding within the respective uncertainties(fig. S1) (20). Thus, our new data are not in con-flict with previous studies, and the slightly ele-vated D17O composition of the Moon was alreadypresent in these data sets (8–10).Carbonaceous chondrites and BSE differ
considerably, not only in D17O, but also in theirTi, Cr, and Ni isotope composition (22), makinga carbonaceous chondrite composition of Theiaunfeasible. Rather, Theia formed from the samelarge noncarbonaceous chondrite (22) reservoiras Earth, Mars, ordinary chondrites (OCs), en-statite chondrites (ECs), and other noncarbo-naceous chondrites and achondrites (22). Allnumerical simulations, except some with very
large impactors, predict that the Moon receivedfractionally more impactor material than Earth(negative dfT). Hence, the D17O of Theia wasmost likely higher than that of Earth and theMoon. The solar system bodies from the non-carbonaceous chondrite (22) reservoir with ahigher D17O than that of Earth are Mars and theparent asteroids of OC and R-chondrites (7).Admixing only 4% of material isotopicallyresembling Mars would be sufficient to explainthe observed 12 ppm difference between Earthand the Moon. For an OC or R-chondrite com-position, less than 2% would be required in the
Moon. Such small proportions are inconsistentwith most numerical models (2, 3, 15–17) thatgenerally suggest larger fractions of Theia in theMoon (Fig. 3A). This implies that the compo-sition of Theia was only slightly higher in D17Othan that of Earth.We have obtained new data on ECs that were
previously assumed to be identical to Earth inD17O (23, 24). Our data show a difference of59 T 8 ppm (1s SEM, n = 14) between Earthand EL (low iron) enstatite chondrites and 35 T10 ppm (1s SEM, n = 10) between Earth andmoremetal-richEH (high iron) enstatite chondrites.
RESEARCH | REPORTS
Fig. 1. Terrestrial (squares) and lunar(circles) samples and ECs (triangles) ind!18O versus D17O space.VSMOW, Viennastandard mean ocean water. Error barsdenote 1s SEM.
Fig. 2. D17O composition for terrestrialand lunar samples. A slope of 0.5305and zero intercept (VSMOW) is used tocalculate D17O (19, 20). Error bars are1s SEM. Solid vertical lines denote meanvalues for the BSE and the Moon. Grayshaded areas represent 1s SEM, anddotted lines represent 2s SEM.
1148 6 JUNE 2014 • VOL 344 ISSUE 6188 sciencemag.org SCIENCE
high-precision data sets (8–10) reveals that lunarsamples are elevated, on average, by 9 ppm com-pared with the terrestrial UWG-2 standard usedin all three studies (fig. S1). With an improvedanalytical technique (19), we are able tomeasurevariations in parts per million (0.00x‰) in D17O(see supplementary materials and methods), al-lowing us to detect the isotopic differences be-tween Earth and the Moon.We first attempted to use lunar meteorites
to determine the isotope composition of theMoon. Our data show that terrestrial weatheringmodified the D17O of the studied lunar meteor-ites, making identification of small variationsimpossible (fig. S2). Therefore, we have ana-lyzed fresh basalt samples from three Apollolanding sites that were provided by NASA. Wecompared the composition of the lunar basaltswith the composition of Earth mantle xenolithsand mantle-derived melts [mid-ocean ridge basalt(MORB)] (Table 1 and Fig. 1).The bulk silicate Earth (BSE) is constrained
to a D17O value of –0.101 T 0.002‰ [1s SEM, n =65 measurements (19)] from mantle xenolithsand MORB from seven localities around theworld [for definitions and analytical details, see(19, 20)]. Earth mantle minerals and MORB fallon a common mass-fractionation line, with aslope of q = 0.532 T 0.006, in the d!17O versus d!18Ospace. This slope is, within uncertainty, iden-tical to the high-temperature approximation forequilibrium oxygen isotope fractionation of0.5305 (21). The D17O values of Earth mantleminerals and MORB are identical within theuncertainty. This is in agreement with massfractionation upon melt extraction along a slopeof ~0.53 for high-temperature processes (19, 21).Because the same slope must apply to meltextraction from the lunar mantle, we can safelyuse the lunar basalts as analogs for the bulksilicate Moon (BSM) with respect to D17O.The three lunar basalts span a small range in
d18O and have an averageD17O = –0.089 T 0.002‰(1s SEM, n = 20). We suggest that this value isrepresentative of the BSM. Thus, the D17O of theMoon is 12 T 3 ppm (0.012 T 0.003‰) higher thanthat of Earth (Fig. 2). This unequivocally identifiesan isotopic difference between Earth and the Moonand supports the view that the Moon formed bya giant collision of the proto-Earth with Theia.The reevaluation of lunar O isotope data from
three previous studies (8–10) is consistent withour finding within the respective uncertainties(fig. S1) (20). Thus, our new data are not in con-flict with previous studies, and the slightly ele-vated D17O composition of the Moon was alreadypresent in these data sets (8–10).Carbonaceous chondrites and BSE differ
considerably, not only in D17O, but also in theirTi, Cr, and Ni isotope composition (22), makinga carbonaceous chondrite composition of Theiaunfeasible. Rather, Theia formed from the samelarge noncarbonaceous chondrite (22) reservoiras Earth, Mars, ordinary chondrites (OCs), en-statite chondrites (ECs), and other noncarbo-naceous chondrites and achondrites (22). Allnumerical simulations, except some with very
large impactors, predict that the Moon receivedfractionally more impactor material than Earth(negative dfT). Hence, the D17O of Theia wasmost likely higher than that of Earth and theMoon. The solar system bodies from the non-carbonaceous chondrite (22) reservoir with ahigher D17O than that of Earth are Mars and theparent asteroids of OC and R-chondrites (7).Admixing only 4% of material isotopicallyresembling Mars would be sufficient to explainthe observed 12 ppm difference between Earthand the Moon. For an OC or R-chondrite com-position, less than 2% would be required in the
Moon. Such small proportions are inconsistentwith most numerical models (2, 3, 15–17) thatgenerally suggest larger fractions of Theia in theMoon (Fig. 3A). This implies that the compo-sition of Theia was only slightly higher in D17Othan that of Earth.We have obtained new data on ECs that were
previously assumed to be identical to Earth inD17O (23, 24). Our data show a difference of59 T 8 ppm (1s SEM, n = 14) between Earthand EL (low iron) enstatite chondrites and 35 T10 ppm (1s SEM, n = 10) between Earth andmoremetal-richEH (high iron) enstatite chondrites.
RESEARCH | REPORTS
Fig. 1. Terrestrial (squares) and lunar(circles) samples and ECs (triangles) ind!18O versus D17O space.VSMOW, Viennastandard mean ocean water. Error barsdenote 1s SEM.
Fig. 2. D17O composition for terrestrialand lunar samples. A slope of 0.5305and zero intercept (VSMOW) is used tocalculate D17O (19, 20). Error bars are1s SEM. Solid vertical lines denote meanvalues for the BSE and the Moon. Grayshaded areas represent 1s SEM, anddotted lines represent 2s SEM.
1148 6 JUNE 2014 • VOL 344 ISSUE 6188 sciencemag.org SCIENCE
[Herwartz et al., 2014]
Slight difference in O isotopes between Moon and Earth.
Can use this to model type of impactor that formed Moon and mixing between Earth and Theia material.
Lunar Meteorites
8
KARNER ET AL.: PYROXENE FROM PLANETARY BASALTS 1577
1 Deposit item AM-06-028, Appendix Table and Appendix Figures A1–A8, which present analyses representing the com-positional range for the suites (coded by thin section). Deposit items are available two ways: For a paper copy contact the Business Offi ce of the Mineralogical Society of America (see inside front cover of recent issue) for price information. For an electronic copy visit the MSA web site at http://www.minsocam.org, go to the American Mineralogist Contents, fi nd the table of contents for the specifi c volume/issue wanted, and then click on the deposit link there.
presented in Appendix Table 1 on the AmMin web site.1 The ap-pendix also contains fi gures A1–A8, which include all pyroxene
analyses broken down by thin section for the systematics of Fe3+ vs. XFe, Na vs. XFe, Cr vs. XFe, Mn vs. XFe, V* vs. XFe, V* vs. Ca, Ti vs. XFe, and Al vs. XFe, where XFe = Fe/(Fe + Mg)atomic. The entire data set is available to any interested parties by request from the fi rst author.
50
40
30
20
10
100
90
80
70
60
50
100 90 80 70 60 50 40 30 20 10
Di
CaMgSi2O6
Hed
CaFe2+Si2O6
En
Mg2Si2O6
Fs
Fe2+2Si2O6
FIGURE 1. Pyroxene Di (CaMgSi2O6)-Hed (CaFeSi2O6)-En (Mg2Si2O6)-Fs (Fe2Si2O6) composite variation diagram for all pyroxene analyses from the Earth (blue), Moon (green), Mars (red), and Vesta (black).
0.00
0.04
0.08
0.12
0.16
0.20
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Fe3+
(af
u)
Earth
Mars
FIGURE 2. Fe3+ atoms per 6-oxygen formula unit (afu) vs. XFe [(Fe/Fe+Mg)atomic] variation diagram for terrestrial and martian pyroxene.
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Na
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 3. Na (afu) vs. XFe for pyroxene grains from the four different planetary bodies.
0.00
0.04
0.08
0.12
0.16
0.000 0.010 0.020 0.030 0.040
Na (afu)
Fe3+
(af
u)
Earth
Mars
FIGURE 4. Fe3+ (afu) vs. Na (afu) for pyroxene grains from Mars and Earth. The positive correlation between Fe3+ and Na shows that the charge-balance substitutional couple of M2Na-M1Fe3+ is very important in terrestrial pyroxenes.
0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Cr
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 5. Cr vs. XFe for pyroxene grains from the Earth, Moon, Mars, and Vesta.
y = 0.0132x + 0.0023
y = 0.0234x + 0.0016
y = 0.0235x + 0.0054
y = 0.0341x
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Fe2+
(afu)
Mn
(af
u)
EarthMoonMarsVesta
FIGURE 6. Mn vs. Fe2+ in atoms per 6-oxygen formula unit (afu) for pyroxene analyses from the Earth, Moon, Mars and Vesta. Best-fi t trend lines are indicated and their equations are given.
[Karner et al., 2006]
Moon is depleted in Na
Lunar Meteorites
9[Karner et al., 2006]
Moon is enriched in Cr
KARNER ET AL.: PYROXENE FROM PLANETARY BASALTS 1577
1 Deposit item AM-06-028, Appendix Table and Appendix Figures A1–A8, which present analyses representing the com-positional range for the suites (coded by thin section). Deposit items are available two ways: For a paper copy contact the Business Offi ce of the Mineralogical Society of America (see inside front cover of recent issue) for price information. For an electronic copy visit the MSA web site at http://www.minsocam.org, go to the American Mineralogist Contents, fi nd the table of contents for the specifi c volume/issue wanted, and then click on the deposit link there.
presented in Appendix Table 1 on the AmMin web site.1 The ap-pendix also contains fi gures A1–A8, which include all pyroxene
analyses broken down by thin section for the systematics of Fe3+ vs. XFe, Na vs. XFe, Cr vs. XFe, Mn vs. XFe, V* vs. XFe, V* vs. Ca, Ti vs. XFe, and Al vs. XFe, where XFe = Fe/(Fe + Mg)atomic. The entire data set is available to any interested parties by request from the fi rst author.
50
40
30
20
10
100
90
80
70
60
50
100 90 80 70 60 50 40 30 20 10
Di
CaMgSi2O6
Hed
CaFe2+Si2O6
En
Mg2Si2O6
Fs
Fe2+2Si2O6
FIGURE 1. Pyroxene Di (CaMgSi2O6)-Hed (CaFeSi2O6)-En (Mg2Si2O6)-Fs (Fe2Si2O6) composite variation diagram for all pyroxene analyses from the Earth (blue), Moon (green), Mars (red), and Vesta (black).
0.00
0.04
0.08
0.12
0.16
0.20
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Fe3+
(af
u)
Earth
Mars
FIGURE 2. Fe3+ atoms per 6-oxygen formula unit (afu) vs. XFe [(Fe/Fe+Mg)atomic] variation diagram for terrestrial and martian pyroxene.
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Na
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 3. Na (afu) vs. XFe for pyroxene grains from the four different planetary bodies.
0.00
0.04
0.08
0.12
0.16
0.000 0.010 0.020 0.030 0.040
Na (afu)
Fe3+
(af
u)
Earth
Mars
FIGURE 4. Fe3+ (afu) vs. Na (afu) for pyroxene grains from Mars and Earth. The positive correlation between Fe3+ and Na shows that the charge-balance substitutional couple of M2Na-M1Fe3+ is very important in terrestrial pyroxenes.
0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Cr
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 5. Cr vs. XFe for pyroxene grains from the Earth, Moon, Mars, and Vesta.
y = 0.0132x + 0.0023
y = 0.0234x + 0.0016
y = 0.0235x + 0.0054
y = 0.0341x
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Fe2+
(afu)
Mn
(af
u)
EarthMoonMarsVesta
FIGURE 6. Mn vs. Fe2+ in atoms per 6-oxygen formula unit (afu) for pyroxene analyses from the Earth, Moon, Mars and Vesta. Best-fi t trend lines are indicated and their equations are given.
Lunar Meteorites
10[Karner et al., 2006]
Moon has higher Fe/Mn values
KARNER ET AL.: PYROXENE FROM PLANETARY BASALTS 1577
1 Deposit item AM-06-028, Appendix Table and Appendix Figures A1–A8, which present analyses representing the com-positional range for the suites (coded by thin section). Deposit items are available two ways: For a paper copy contact the Business Offi ce of the Mineralogical Society of America (see inside front cover of recent issue) for price information. For an electronic copy visit the MSA web site at http://www.minsocam.org, go to the American Mineralogist Contents, fi nd the table of contents for the specifi c volume/issue wanted, and then click on the deposit link there.
presented in Appendix Table 1 on the AmMin web site.1 The ap-pendix also contains fi gures A1–A8, which include all pyroxene
analyses broken down by thin section for the systematics of Fe3+ vs. XFe, Na vs. XFe, Cr vs. XFe, Mn vs. XFe, V* vs. XFe, V* vs. Ca, Ti vs. XFe, and Al vs. XFe, where XFe = Fe/(Fe + Mg)atomic. The entire data set is available to any interested parties by request from the fi rst author.
50
40
30
20
10
100
90
80
70
60
50
100 90 80 70 60 50 40 30 20 10
Di
CaMgSi2O6
Hed
CaFe2+Si2O6
En
Mg2Si2O6
Fs
Fe2+2Si2O6
FIGURE 1. Pyroxene Di (CaMgSi2O6)-Hed (CaFeSi2O6)-En (Mg2Si2O6)-Fs (Fe2Si2O6) composite variation diagram for all pyroxene analyses from the Earth (blue), Moon (green), Mars (red), and Vesta (black).
0.00
0.04
0.08
0.12
0.16
0.20
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Fe3+
(af
u)
Earth
Mars
FIGURE 2. Fe3+ atoms per 6-oxygen formula unit (afu) vs. XFe [(Fe/Fe+Mg)atomic] variation diagram for terrestrial and martian pyroxene.
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Na
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 3. Na (afu) vs. XFe for pyroxene grains from the four different planetary bodies.
0.00
0.04
0.08
0.12
0.16
0.000 0.010 0.020 0.030 0.040
Na (afu)
Fe3+
(af
u)
Earth
Mars
FIGURE 4. Fe3+ (afu) vs. Na (afu) for pyroxene grains from Mars and Earth. The positive correlation between Fe3+ and Na shows that the charge-balance substitutional couple of M2Na-M1Fe3+ is very important in terrestrial pyroxenes.
0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.20 0.40 0.60 0.80 1.00
XFe (atomic)
Cr
(afu
)
Earth
Moon
Mars
Vesta
FIGURE 5. Cr vs. XFe for pyroxene grains from the Earth, Moon, Mars, and Vesta.
y = 0.0132x + 0.0023
y = 0.0234x + 0.0016
y = 0.0235x + 0.0054
y = 0.0341x
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Fe2+
(afu)
Mn
(af
u)
EarthMoonMarsVesta
FIGURE 6. Mn vs. Fe2+ in atoms per 6-oxygen formula unit (afu) for pyroxene analyses from the Earth, Moon, Mars and Vesta. Best-fi t trend lines are indicated and their equations are given.
Lunar Meteorites
11
Fe and Mn are in 2+ oxidation
state on Moon, thus they do not
fractionate during
geochemical processes on the Moon like they do on Earth.
[WUSTL meteorite site]
Magma Ocean or Not?
12[Gross et al., 2012]
3) Most feldspathic meteorites contain little KREEP, consistent with the Th distribution on the lunar surface (concentrated in the nearside Procellarium KREEP Terrain, PKT). This alone suggests that a hallmark of LMO models KREEP, is not globally distributed.
4) Clasts of Mg-suite rock are rare in lunar feldspathic meteorites. Nearly all clasts that plot in the Apollo Mg-suite field (Fig.2) are magnesian anorthosites and noritic anor-thosites that appear continuous with the range of other anorthositic clasts. Very few clasts in the feldspathic mete-orites fall along the Mg-suite trend of decreasing An and Mg# (Fig.2).
Lunar Farside Highlands: The lunar feldspathic me-
teorites suggest that magnesian anorthositic rock is a major component of the lunar highland crust and is the dominant material over most of the Moon. Only 1/3 of the feldspathic meteorites are ferroan, which is approximately the same as the proportion of the lunar highlands that is affected by the continuous Imbrium ejecta [3,16]. Only the Luna 20 and 24 missions returned samples from outside the continuous ejecta of Imbrium, and their highlands materials are dominated by magnesian anorthositic rocks, not ferroan anorthosites [17]. Thus, it seems reasonable to suggest that ferroan anorthosite, as well as KREEP and Mg-suite rocks, reflects processes lo-calized in the Imbrium area and that magnesian anor-thositic material is characteristic of the remaining 2/3 of the lunar crust. So, how did that remaining 2/3 form? By analogy with the Imbrium ejecta, and by impact modeling [27], that surface is likely underlain by ejecta from the South-Pole-Aitkin basin (SPA). SPA is the largest and oldest recognized lunar impact basin, and its ejecta blanket was kilometers thick over the whole lunar surface [27]. This ejecta, mostly of mid- and deep-crustal material, should be dominant at the lunar surface, except where covered by younger deposits (basin ejecta or basalt). Thus, it seems reasonable that the materials of most feldspathic highlands meteorites should be princi-
pally derived from mid- and deep crustal materials of the SPA target, and the remainder of highlands meteor-ites should consist mostly of Imbrium ejecta.
If the surface of the lunar highlands as we see it today is a continuous blanket of the Imbrium and SPA ejecta then how did the original lunar crust form and evolve? The long-standing alternative to the LMO hy-pothesis is serial magmatism – that the observed lunar crust is the product of multiple intrusions of basaltic magma, each differentiating during and after emplace-ment so that any primordial LMO crust is ob-scured[18,19]. In that model, plagioclase-rich cumulates from intrusions that rise into the crust as diapirs [18,20-22], while the complementary mafic layers sank back to the mantle [18]. Our data and the literature data are consistent with a modified version of this model, in which layered intrusions are emplaced close to the surface, differentiate and rise continuously over time. Each diapir is expected to have its own unique chemis-try, Mg# range and plagioclase composition, depend-ing on the physical and chemical characteristics of its source region and the duration of ascent and fractiona-tion of interstitial melt within the diapirs [23]. During the SPA impact event those diapirs, close to the lunar surface, were then distributed onto and over the lunar surface. Acknowledgements: Supported in part by NASA Cosmo-chemistry Grant NNX08AH78G to AHT and an NLSI/CLSE subcontract to J.Gross. References: [1] Wood, J.A. et al. (1970) Proc. 1st Lunar Sci. Conf., 965-988. [2] Shearer, C.K., et al. (2006) Rev. Min. Geochem. 60, 365-518. [3] Warren, P.H. (1990) Am. Min. 75, 46-58. [4] Elkins-Tanton, L.T et al. (2011) EPSL 304, 326–336. [5] Borg, L. et al. (1999) Geochim. Cosmochim. Acta, 63, 2679-2691. [6] Lucey, P.G. (2004) GRL 31, L08701. [7] Prettyman, T.H. et al. (2006) J. Geoph. Res. 111, E12007 [8i] Greenhagen, B.T. et al. (2010) Science 329, 1507-1509. [9] Warren, P.H. (1985) Ann. Rev. Earth Planet. Sci. 13, 201-240. [10] Korotev, R.L. (2005) Chemie der Erde 65, 297–346. [11] Korotev, R.L. (2011) <http://meteorites.wustl.edu/lunar/moon_meteorites_list_alumina.htm>. [12] Goodrich, C.A., et al. (1984) Proc. 15th Lunar Planet. Sci. Conf., C87-C94. [13] Kallemeyn, G.W. & Warren, P.H. (1983) Geophys. Res. Let. 10, 833-836. [14] Korotev, R.L. et al. (2009) Meteor. Planet. Sci. 44, 1287–1322. [16] Spudis, et al. (2011) 42nd LPSC, Abstract #1365. [17] Taylor G.J., et al. (1973) Geochim. Cosmochim. Acta 37, 1087-1106. [18] Longhi, J. & Ashwall, L.D. (1985 Proc. 15th Lunar Planet. Sci. Conf., C571-C584. [19] Longhi, J. (2003) J. Geophys. Res. 108, #5083. [20] Ashwal, L.D. (1993) Anorthosites. Springer, 422 p. [21] Longhi, J., et al. (1999) J. Petrol. 40, 339-362. [22] Vander Auwera, et al. (2006) Lithos 89, 326-352. [23] Haloda, J et al. (2009) Geo-chim. Cosmochim. Acta, 73, 3450-3470. [24] Korotev R.L. et al. (2010) LPSC 41, abstr. Xxxx. [25] Ohtake et al, 2010; Nature; [26] Gross J. et al. (2012) LPSC 43. [27] Petro and Pieters (2008).
Fig. 2: A) Graph of anorthite (mol%) in plagioclase versus Mg# in mafic minerals in lunar samples. A) Anorthosite clasts in ALHA81005 (red symbols) and NWA2996 (blue symbols); B) Fields of anorthosite clasts in some feldspathic lunar meteorites and Luna 24, each color represents a different meteorite.
9021.pdfSecond Conference on the Lunar Highlands Crust (2012)
Feldspathic meteorites suggest magnesian (not ferroan) anorthosite is a major component of highland crust. Only Luna 20 & 24 sampled areas outside of Imbrium ejecta, and they are dominated by magnesian anorthosite.
Magma Ocean or Not?
13[Gross et al., 2012]
3) Most feldspathic meteorites contain little KREEP, consistent with the Th distribution on the lunar surface (concentrated in the nearside Procellarium KREEP Terrain, PKT). This alone suggests that a hallmark of LMO models KREEP, is not globally distributed.
4) Clasts of Mg-suite rock are rare in lunar feldspathic meteorites. Nearly all clasts that plot in the Apollo Mg-suite field (Fig.2) are magnesian anorthosites and noritic anor-thosites that appear continuous with the range of other anorthositic clasts. Very few clasts in the feldspathic mete-orites fall along the Mg-suite trend of decreasing An and Mg# (Fig.2).
Lunar Farside Highlands: The lunar feldspathic me-
teorites suggest that magnesian anorthositic rock is a major component of the lunar highland crust and is the dominant material over most of the Moon. Only 1/3 of the feldspathic meteorites are ferroan, which is approximately the same as the proportion of the lunar highlands that is affected by the continuous Imbrium ejecta [3,16]. Only the Luna 20 and 24 missions returned samples from outside the continuous ejecta of Imbrium, and their highlands materials are dominated by magnesian anorthositic rocks, not ferroan anorthosites [17]. Thus, it seems reasonable to suggest that ferroan anorthosite, as well as KREEP and Mg-suite rocks, reflects processes lo-calized in the Imbrium area and that magnesian anor-thositic material is characteristic of the remaining 2/3 of the lunar crust. So, how did that remaining 2/3 form? By analogy with the Imbrium ejecta, and by impact modeling [27], that surface is likely underlain by ejecta from the South-Pole-Aitkin basin (SPA). SPA is the largest and oldest recognized lunar impact basin, and its ejecta blanket was kilometers thick over the whole lunar surface [27]. This ejecta, mostly of mid- and deep-crustal material, should be dominant at the lunar surface, except where covered by younger deposits (basin ejecta or basalt). Thus, it seems reasonable that the materials of most feldspathic highlands meteorites should be princi-
pally derived from mid- and deep crustal materials of the SPA target, and the remainder of highlands meteor-ites should consist mostly of Imbrium ejecta.
If the surface of the lunar highlands as we see it today is a continuous blanket of the Imbrium and SPA ejecta then how did the original lunar crust form and evolve? The long-standing alternative to the LMO hy-pothesis is serial magmatism – that the observed lunar crust is the product of multiple intrusions of basaltic magma, each differentiating during and after emplace-ment so that any primordial LMO crust is ob-scured[18,19]. In that model, plagioclase-rich cumulates from intrusions that rise into the crust as diapirs [18,20-22], while the complementary mafic layers sank back to the mantle [18]. Our data and the literature data are consistent with a modified version of this model, in which layered intrusions are emplaced close to the surface, differentiate and rise continuously over time. Each diapir is expected to have its own unique chemis-try, Mg# range and plagioclase composition, depend-ing on the physical and chemical characteristics of its source region and the duration of ascent and fractiona-tion of interstitial melt within the diapirs [23]. During the SPA impact event those diapirs, close to the lunar surface, were then distributed onto and over the lunar surface. Acknowledgements: Supported in part by NASA Cosmo-chemistry Grant NNX08AH78G to AHT and an NLSI/CLSE subcontract to J.Gross. References: [1] Wood, J.A. et al. (1970) Proc. 1st Lunar Sci. Conf., 965-988. [2] Shearer, C.K., et al. (2006) Rev. Min. Geochem. 60, 365-518. [3] Warren, P.H. (1990) Am. Min. 75, 46-58. [4] Elkins-Tanton, L.T et al. (2011) EPSL 304, 326–336. [5] Borg, L. et al. (1999) Geochim. Cosmochim. Acta, 63, 2679-2691. [6] Lucey, P.G. (2004) GRL 31, L08701. [7] Prettyman, T.H. et al. (2006) J. Geoph. Res. 111, E12007 [8i] Greenhagen, B.T. et al. (2010) Science 329, 1507-1509. [9] Warren, P.H. (1985) Ann. Rev. Earth Planet. Sci. 13, 201-240. [10] Korotev, R.L. (2005) Chemie der Erde 65, 297–346. [11] Korotev, R.L. (2011) <http://meteorites.wustl.edu/lunar/moon_meteorites_list_alumina.htm>. [12] Goodrich, C.A., et al. (1984) Proc. 15th Lunar Planet. Sci. Conf., C87-C94. [13] Kallemeyn, G.W. & Warren, P.H. (1983) Geophys. Res. Let. 10, 833-836. [14] Korotev, R.L. et al. (2009) Meteor. Planet. Sci. 44, 1287–1322. [16] Spudis, et al. (2011) 42nd LPSC, Abstract #1365. [17] Taylor G.J., et al. (1973) Geochim. Cosmochim. Acta 37, 1087-1106. [18] Longhi, J. & Ashwall, L.D. (1985 Proc. 15th Lunar Planet. Sci. Conf., C571-C584. [19] Longhi, J. (2003) J. Geophys. Res. 108, #5083. [20] Ashwal, L.D. (1993) Anorthosites. Springer, 422 p. [21] Longhi, J., et al. (1999) J. Petrol. 40, 339-362. [22] Vander Auwera, et al. (2006) Lithos 89, 326-352. [23] Haloda, J et al. (2009) Geo-chim. Cosmochim. Acta, 73, 3450-3470. [24] Korotev R.L. et al. (2010) LPSC 41, abstr. Xxxx. [25] Ohtake et al, 2010; Nature; [26] Gross J. et al. (2012) LPSC 43. [27] Petro and Pieters (2008).
Fig. 2: A) Graph of anorthite (mol%) in plagioclase versus Mg# in mafic minerals in lunar samples. A) Anorthosite clasts in ALHA81005 (red symbols) and NWA2996 (blue symbols); B) Fields of anorthosite clasts in some feldspathic lunar meteorites and Luna 24, each color represents a different meteorite.
9021.pdfSecond Conference on the Lunar Highlands Crust (2012)
KREEP, Mg-suite, and ferroan anorthosite rocks reflect processes localized to Imbrium? Serial magmatism to form variety of layered intrusions that are then spread over planet by SPA basin-forming process?
Magma Ocean or Not?
14
Lunar meteorites exhibit different trace element patterns compared to Apollo and Luna samples: heterogeneous mantle sources?
Note the range in Eu depletion....varying degrees of plag. xtlzn?
samples, but 14 contain magnesian anorthosites (with Mg#s up to 90), and little or no ferroan anorthosite, Mg-suite rocks or KREEP (Fig. 2B).
These meteorite data show that ferroan anortho-
sites are not globally distributed. The Apollo landing sites are all strongly influenced by the continuous ejec-ta of the Imbrium basin [3,16]. Only the Luna 20 and 24 missions returned samples from outside the contin-uous ejecta of Imbrium, and their highlands materials are dominated by magnesian anorthositic rocks, not ferroan anorthosites [17]. Thus, it seems reasonable to suggest that ferroan anorthosite, as well as KREEP and Mg-suite rocks, reflects processes localized in the Im-brium area and cannot be extrapolated to the whole Moon or to a global LMO. The lunar meteorites provide further support for this idea. Of the 19 feldspathic high-lands meteorites with adequate data, approximately only 1/4 contain ferroan anorthosite; this proportion is ap-proximately the same as the proportion of the lunar highlands that is affected by continuous Imbrium ejecta.
How did the lunar plagioclase-rich crust, as we see it today, form? The long-standing alternative to the LMO hypothesis is serial magmatism – that the observed lunar crust is the product of multiple intrusions of basaltic magma, each differentiating during and after emplace-ment so that any primordial LMO crust is ob-scured[18,19]. In that model, heat from the Moon’s inte-rior allowed plagioclase-rich cumulates from the intru-sions to rise into the crust as diapirs [18,20-22], while the complementary mafic layers sank back to the mantle [18].
Our data and the literature data are consistent with a modified version of this model, in which layered intrusions are emplaced and differentiate continuously over time. The range of crystallization ages of ferroan anorthosites [2,5], inconsistent with a single LMO, is a natural consequence of this serial diapirism in which anorthosite diapirs form and rise continuously over time. Each diapir is expected to have its own unique
chemistry, Mg# range and plagioclase composition, depending on the physical and chemical characteristics of its source region and the duration of ascent and frac-tionation of interstitial melt within the diapirs (Fig 3). The sources of mare basalts and magnesian suite rocks would form as mixtures of primitive mantle with the sinking diapirs of mafic material [18]. Portions of each mafic diapir would bear the element signatures of pla-gioclase co-crystallization (i.e. Eu depletion), of late ilmenite cumulates, and of late magmatic differentiates (KREEP). Each potential source area for mare basalt would have a unique mafic diapir input, and thus a unique degree of Eu depletion, Ti enrichment, and KREEP enrichment, contributing to the compositional diversity of mare basalts. Thus it is not surprising that we see a range of Eu depletions [23] (Fig. 4).
Acknowledgements: Supported in part by NASA Cos-mochemistry Grant NNX08AH78G to AHT. References: [1] Wood, J.A. et al. (1970) Proc. 1st Lunar Sci. Conf., 965-988. [2] Shearer, C.K., et al. (2006) Rev. Min. Geochem. 60, 365-518. [3] Warren, P.H. (1990) Am. Min. 75, 46-58. [4] Elkins-Tanton, L.T et al. (2011) EPSL 304, 326–336. [5] Borg, L. et al. (1999) Geochim. Cosmochim. Acta, 63, 2679-2691. [6] Lucey, P.G. (2004) GRL 31, L08701. [7] Prettyman, T.H. et al. (2006) J. Geoph. Res. 111, E12007 [8i] Greenhagen, B.T. et al. (2010) Science 329, 1507-1509. [9] Warren, P.H. (1985) Ann. Rev. Earth Planet. Sci. 13, 201-240. [10] Korotev, R.L. (2005) Chemie der Erde 65, 297–346. [11] Korotev, R.L. (2011) <http://meteorites.wustl.edu/lunar/moon_meteorites_list_alumina.htm>. [12] Goodrich, C.A., et al. (1984) Proc. 15th Lunar Planet. Sci. Conf., C87-C94. [13] Kallemeyn, G.W. & Warren, P.H. (1983) Geophys. Res. Let. 10, 833-836. [14] Korotev, R.L. et al. (2009) Meteor. Planet. Sci. 44, 1287–1322. [16] Spudis, et al. (2011) 42nd LPSC, Abstract #1365. [17] Taylor G.J., et al. (1973) Geochim. Cosmochim. Acta 37, 1087-1106. [18] Longhi, J. & Ashwall, L.D. (1985 Proc. 15th Lunar Planet. Sci. Conf., C571-C584. [19] Longhi, J. (2003) J. Geophys. Res. 108, #5083. [20] Ashwal, L.D. (1993) Anorthosites. Springer, 422 p. [21] Longhi, J., et al. (1999) J. Petrol. 40, 339-362. [22] Vander Auwera, et al. (2006) Lithos 89, 326-352. [23] Haloda, J et al. (2009) Geo-chim. Cosmochim. Acta, 73, 3450-3470.
Fig. 2: A) Graph of anorthite (mol%) in plagioclase versus Mg# in mafic minerals in lunar samples. A) Anorthosite clasts in ALHA81005 (red symbols) and NWA2996 (blue symbols); B) Fields of anorthosite clasts in some feldspathic lunar meteorites and Luna 24, each color represents a different meteorite.
Fig. 3: Range of REE in selected lunar basalts. After [23].
2306.pdf43rd Lunar and Planetary Science Conference (2012)
[Gross et al., 2012]
Lunar Meteorites
15
3: Timing of impact events as recorded by argon isotopes in Apollo 16 (black curve) and 17 (blue curve) and lunar meteorites (red curve). (a) Full y-axis scale and (b) limited y-axis scale showing details of the lunar meteorite dataset. To calculate these curves, the age and error were combined in bins of 0.5 Gyr (500 Myr), which is representative of the average error in 40Ar/39Ar age determination for the Apollo 16 samples. The normalized Gaussian curve calculated for each age bin (age column) was obtained by taking into consideration the width of the Gaussian curve calculated and the measured uncertainties. Each column was added and the result normalized to the total of all analyses per sample. Argon plateau data for Apollo 16 and 17 samples from supplementary dataset of Shuster et al. (2010) (see details for S2 and references therein) with additional data from Borchardt et al. (1986). Individual impact melt clasts and bulk rock (MacAlpine Hills 88105, Dar al Gani 400, Dar al Gani 262, Queen Alexandra Range 93069, Dhofar 026, 280 and 303) impact melt lunar meteorites taken from data compiled by Fernandes et al. (2013) with additional argon data from Yamato-86032 (Nyquist et al. 2006), Sayh al Uhaymir 169 (Gnos et al. 2004), North West Africa 482 (Daubar et al. 2002) and Dhofar 489 (Takeda et al. 2006).
[Joy & Arai, 2013]
