Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 Lunar bulk chemical composition: a post-Gravity Recovery and Interior Laboratory reassessment rsta.royalsocietypublishing.org G. Jeffrey Taylor1 and Mark A. Wieczorek2 1 Hawai‘i Institute of Geophysics and Planetology, University of Research Cite this article: Taylor GJ, Wieczorek MA. 2014 Lunar bulk chemical composition: a post-Gravity Recovery and Interior Laboratory reassessment. Phil. Trans. R. Soc. A 372: 20130242. http://dx.doi.org/10.1098/rsta.2013.0242 One contribution of 19 to a Discussion Meeting Issue ‘Origin of the Moon’. Subject Areas: geochemistry, geophysics Keywords: refractory elements, volatiles, bulk composition, crust, mantle, Gravity Recovery and Interior Laboratory Author for correspondence: G. Jeffrey Taylor e-mail: [email protected] Hawai‘i, Honolulu, HI 96822, USA 2 Institut de Physique du Globe de Paris, Sorbonne Paris Cité, Université Paris Diderot, Case 7071, Lamarck A, 35 rue Hélène Brion, Paris Cedex 13 75205, France New estimates of the thickness of the lunar highlands crust based on data from the Gravity Recovery and Interior Laboratory mission, allow us to reassess the abundances of refractory elements in the Moon. Previous estimates of the Moon fall into two distinct groups: earthlike and a 50% enrichment in the Moon compared with the Earth. Revised crustal thicknesses and compositional information from remote sensing and lunar samples indicate that the crust contributes 1.13–1.85 wt% Al2 O3 to the bulk Moon abundance. Mare basalt Al2 O3 concentrations (8–10 wt%) and Al2 O3 partitioning behaviour between melt and pyroxene during partial melting indicate mantle Al2 O3 concentration in the range 1.3–3.1 wt%, depending on the relative amounts of pyroxene and olivine. Using crustal and mantle mass fractions, we show that that the Moon and the Earth most likely have the same (within 20%) concentrations of refractory elements. This allows us to use correlations between pairs of refractory and volatile elements to confirm that lunar abundances of moderately volatile elements such as K, Rb and Cs are depleted by 75% in the Moon compared with the Earth and that highly volatile elements, such as Tl and Cd, are depleted by 99%. The earthlike refractory abundances and depleted volatile abundances are strong constraints on lunar formation processes. 2014 The Author(s) Published by the Royal Society. All rights reserved. Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 1. Introduction The thickness of the lunar crust has been revised several times since the first seismic data were analysed during the Apollo missions. These revisions were in part a result of using improved computational approaches to analyse the seismic data, but also the collection of improved topographic and gravitational data from subsequent orbital missions. In this section, we review the history of how the thickness of the lunar crust has been estimated, and then provide mass fractions of the Moon’s major geochemical reservoirs using the most recent GRAIL-derived crustal thickness models. The initial analyses of the Apollo seismic data implied that the crust was about 55–60 km thick in the vicinity of the Apollo 12 and 14 landing sites. Toksöz et al. [14] reported a sharp seismic discontinuity in the P-wave velocity profile at 55 km depth, and a later independent analysis by Nakamura [15] obtained a concordant value of 58 ± 8 km. A study by Goins et al. [16] provided tentative evidence for a crustal thickness of 75 km in the highlands at the Apollo 16 site. These three seismic analyses were the basis of numerous bulk Moon compositional studies that were published over two decades. A summary of how measurements of the seismically determined and average crustal thickness of the Moon have evolved with time is given in figure 1. Laser altimeter data were collected along the equatorial ground tracks of the Apollo 15 and 17 command and service modules, showing that the elevations over the nearside hemisphere were about 4 km lower than those over the farside [17]. As these large-scale variations in topography were not associated with any significant gravity anomaly, the crust should be largely isostatically compensated at these scales. If the variations in surface elevation were compensated by a crustal root (Airy compensation), the farside crust would be thicker than that of the nearside and the ......................................................... 2. Thickness of the crust rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 A firm knowledge of the bulk chemical composition of the Moon is important for unravelling the details of lunar origin. For example, it is well known that the Moon is depleted in volatile elements compared with the Earth, suggesting loss during lunar origin. Researchers have debated whether the Moon is enriched in refractory elements (Th, U, Zr, rare earth elements, etc.) compared with the Earth, with some authors arguing for very similar abundances [1–5] and others suggesting that the Moon is enriched in refractory elements by at least 50% compared with the Earth [5–11]. (In this paper, ‘refractory’ will be used to denote elements that condense in a gas of solar composition at temperatures greater than 1300 K at a pressure of 10−4 bar [12]. Note that condensation temperatures are only a guide to volatility as they apply to equilibrium in a gas of solar composition at low pressure, conditions vastly different from those in the hypothetical protolunar disc.) These two drastically different views of lunar refractory abundances were driven largely by estimates of the thickness of the feldspathic crust (see §2). However, data from heat flow probes installed during the Apollo 15 and 17 missions also seemed to indicate high concentrations of U and Th, at least locally, and these measurements were used by some to also favour a global enrichment in refractory elements. Global γ-ray data from the Lunar Prospector mission helped identify the confines of the trace-element-rich province that has been named the Procellarum KREEP Terrane (PKT [9]). One of the heat-flow measurements was made in this province, with the other just exterior to it, calling into question the idea that the Moon was enriched globally in U and Th. Indeed, the asymmetric distribution of Th, as well as its unknown depth dependence, makes it difficult to determine its concentration in the crust and mantle, as seen by variations in estimates of the bulk Moon Th abundance [5,9,11,13]. In this paper, we reassess the abundances of refractory elements on the basis of revised estimates of crustal thickness derived from gravity data obtained from the Gravity Recovery and Interior Laboratory (GRAIL) mission. We conclude that the Moon is not enriched in refractory elements in comparison to the Earth. We then use the revised lunar refractory abundances and element correlations to determine the abundances of volatile elements in the bulk Moon. 2 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 80 average crustal thickness 3 70 crustal thickness (km) 40 30 seismic crustal thickness A16 70 60 A12/14 Apollo zone 50 Apollo zone 40 A16 A15 A12 A14 Apollo zone 30 Apollo zone 1970 1975 1980 1985 1990 1995 2000 publication date 2005 2010 Figure 1. Seismically determined crustal thickness estimates (bottom) and average crustal thickness estimates as a function of publication date. As analyses of the Apollo seismic data have improved with time, and as orbital missions have improved our knowledge of the lunar gravitational field and topography, estimates of the Moon’s crustal thickness have decreased by almost a factor of two over a 40 year time period. average crustal thickness of the Moon would be even greater than that obtained from the seismic analyses [17]. After accounting for lateral variations in crustal density, as estimated from orbital γ-ray spectrometer data, Haines & Metzger [18] showed that the farside crust should be on average 22 km thicker than the nearside crust. In using the seismic constraint of Goins et al. [16], they reported an average crustal thickness of 73 km. The Clementine and Lunar Prospector missions, launched respectively in 1994 and 1998, obtained the first near-global topographic maps of the Moon [19] and allowed for the construction of a vastly improved gravitation field, especially over the nearside hemisphere of the Moon [20]. With these data, it was possible to create the first global crustal thickness models by assuming densities for the crust and mantle, combined with knowledge of the crustal thickness at one locale. Neumann et al. [21] obtained an average crustal thickness of 61 km by requiring their models to match the seismically determined 55 km thickness at the Apollo 12 and 14 sites. Using an improved inversion approach, Wieczorek & Phillips [22] obtained an average thickness of 66 km with a 60 km constraint at the Apollo 12 and 14 sites. However, if the crust of the Moon were isostatically compensated by crustal thickness variations, Wieczorek & Phillips [23] showed that the relation between the lunar geoid and topography implied an average crustal thickness of 43 ± 20 km (revised to 49 ± 16 km [24]). At the low elevations of the Apollo 12 and 14 sites, the crustal thickness should have been about 28 km. Rather than call into question the accuracy of the Apollo seismic constraints, these authors instead interpreted this finding to mean that the lunar crust was stratified in density, with a low-density anorthositic layer overlying a denser noritic layer. Two independent re-analyses of the Apollo seismic data were performed following the Lunar Prospector mission, with both suggesting that the lunar crust was considerably thinner than previously thought. Khan et al. [25] suggested that the crustal thickness was 45 ± 5 km in the Apollo zone (which is most representative of the Apollo 12 and 14 landing sites given the distribution of seismic events used in their analysis). In a subsequent analysis that tested the hypothesis of having a crust thinner than the Apollo-era analyses, they showed that models with ......................................................... 50 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 60 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 With the GRAIL crustal thickness constraints, we calculate the mass fractions of the major lunar geochemical terranes. We report results that make use of the 30 km seismic constraint [10] with a crustal porosity of 12% and provide results for three other GRAIL models in table 1. If the radius of the core is 380 km with a density of 5200 kg m−3 [32], then the mass fractions of the core, mantle and crust are 1.6%, 94% and 4.4%, respectively. Based on orbital geochemical data, as well as analyses of lunar samples, the composition of the crust is known to vary among three major provinces that have distinct geologic evolutions. Following Jolliff et al. [9], these correspond to the Procellarum KREEP Terrane, the Feldspathic Highlands Terrane (FHT) and the interior of the South Pole–Aitken (SPA) impact basin. The FHT was originally divided into an inner anorthositic region and an outer, somewhat more mafic region, but we ignore this distinction here. The PKT is distinguished by having elevated abundances of the geochemical component called KREEP (for potassium, rare earth elements and phosphorous), which includes the important heat-producing elements uranium, thorium and potassium. The FHT is composed largely of anorthositic materials that are believed to represent the Moon’s primary crust that formed by flotation in a magma ocean. Finally, the interior of the SPA impact basin possesses a mafic geochemical anomaly, along with moderate enhancements in incompatible elements. The composition of this basin’s floor may represent either lower crustal materials, or a differentiated impact-melt pool [33]. ......................................................... 3. Mass fractions of major geochemical terranes 4 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 thicker crusts had unrealistic P-wave velocity profiles and that a crustal thickness of 38 ± 3 was preferred [26]. Lognonné et al. [10] performed an independent reprocessing of the Apollo seismic data, including the picking of new arrival times, and suggested that the crustal thickness in the Apollo zone was 30 ± 2.5 km. Several global crustal thickness models were constructed following the revision of the Apollo seismic results. Wieczorek et al. [24] obtained an average crustal thickness of 53 km using the 45 km seismic constraint at the Apollo 12 and 14 landing sites. From a joint seismic and global crustal thickness inversion, Chenet et al. [27] found the average thickness of the crust to lie between 32 and 45 km, with a best fit value of 40 km. This study also provided seismically determined crustal thickness estimates at numerous meteoroid impact locations and provided revised crustal thicknesses at the Apollo 12, 14, 15 and 16 sites of 33 ± 5, 31 ± 7, 35 ± 8 and 38 ± 7 km, respectively. In an inversion by Hikida & Wieczorek [28], they required only that the minimum crustal thickness be less than 1 km and obtained an average crustal thickness of 43 km, with a thickness of about 40 km near the Apollo 12 and 14 landing sites. Using a similar approach, but with improved global topography and gravity models from the Kaguya mission, Ishihara et al. [29] obtained an average crustal thickness of 53 km, with a thickness of about 47 km near the Apollo 12 and 14 sites. Almost all of the global crustal thickness models discussed above assumed that the density of the crust was somewhere between 2800 and 2900 kg m−3 , which corresponds to the grain density of typical anorthositic rocks. High-resolution gravity data from the GRAIL mission, however, have shown that the bulk density of the lunar crust is considerably lower, 2550 kg m−3 [30], corresponding to a porosity of about 12%. As discussed in Wieczorek et al. [30], this porosity probably extends into the uppermost mantle. Taking into account lateral variations in crustal density as implied by remote sensing data, Wieczorek et al. [30] varied the mantle density in order to find a global crustal thickness model that fit the recent seismic constraints and that possessed minimum crustal thicknesses close to zero. The thinnest crust was found to occur beneath the Crisium and Moscoviense impact basins, precisely where remote sensing data suggest that the mantle may have been excavated [31]. When using the seismic crustal thickness constraints of 30 and 38 km at the Apollo 12 and 14 landing sites with 12% porosity in the crust, the average crustal thickness was found to be 34 and 43 km, respectively. Two additional models were constructed using a 7% porosity for the entire crust that yielded nearly identical results, but with somewhat different mantle densities. Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 Table 1. Crustal thickness models used in assessing crustal composition. model 2 model 3 model 4 surface area (%) 77.53 77.53 77.53 77.53 .......................................................................................................................................................................................................... mean elevation (km) 0.58 0.58 0.58 0.58 volume (m3 ) 1.06 × 1018 1.09 × 1018 1.31 × 1018 1.32 × 1018 crustal thickness (km) 36.66 37.81 45.87 45.99 mass (kg) 2.71 × 10 2.95 × 10 3.37 × 10 3.57 × 1021 3.75 4.08 4.66 4.94 2566 2711 2566 2712 .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... 21 21 21 .......................................................................................................................................................................................................... silicate mass fraction (%) .......................................................................................................................................................................................................... −3 average density (kg m ) .......................................................................................................................................................................................................... PKT .......................................................................................................................................................................................................... surface area (%) 16.56 mean elevation (km) −1.49 16.56 16.56 16.56 .......................................................................................................................................................................................................... −1.49 −1.49 −1.49 .......................................................................................................................................................................................................... volume (m ) 1.63 × 10 crustal thickness (km) 26.36 mass (kg) 4.22 × 10 4.57 × 10 5.53 × 10 5.80 × 1020 silicate mass fraction (%) 0.58 0.63 0.77 0.80 −3 2594 2742 2594 2742 3 17 1.67 × 10 17 2.13 × 10 17 2.12 × 1017 .......................................................................................................................................................................................................... 27.01 34.75 34.46 .......................................................................................................................................................................................................... 20 20 20 .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... average density (kg m ) .......................................................................................................................................................................................................... SPA .......................................................................................................................................................................................................... surface area (%) 5.91 mean elevation (km) −3.37 5.91 5.91 5.91 .......................................................................................................................................................................................................... −3.37 −3.37 −3.37 .......................................................................................................................................................................................................... volume (m ) 4.51 × 10 4.54 × 10 6.26 × 10 6.08 × 1016 crustal thickness (km) 20.48 20.59 28.52 27.70 mass (kg) 1.18 × 10 1.26 × 10 1.64 × 10 1.68 × 1020 silicate mass fraction (%) 0.16 0.17 0.23 0.23 −3 2618 2766 2619 2767 3 16 16 16 .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... 20 20 20 .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... average density (kg m ) .......................................................................................................................................................................................................... The areas spanned by the three major crustal provinces, shown in figure 2, are estimated using thorium and iron abundances obtained from the Lunar Prospector γ-ray spectrometer. The PKT is here defined by those regions having more than 4 ppm Th as determined from the 2◦ lowaltitude Th map [34]. As some regions within this province have been covered by thin layers of low thorium basalts, we also include FeO-rich regions (more than 13 wt%) where the thorium abundances are greater than 2.75 ppm. The interior of the SPA basin is defined by those regions having more than 6.5 wt% FeO using the 5◦ data [35]. As shown in figure 2 and table 1, the FHT, PKT and SPA basin comprise, respectively, 77.5%, 16.6% and 5.9% of the Moon’s surface area. For GRAIL model 1 (thin crust, high porosity), the average crustal thicknesses of these three provinces are 36.7, 27.9 and 20.5 km, which correspond to bulk Moon silicate mass fractions of 3.75%, 0.58% and 0.16%, respectively. 4. Al2 O3 concentration in the bulk silicate Moon The most straightforward way to assess the bulk abundance of refractory elements in the Moon is to determine the contributions to the bulk lunar Al2 O3 abundance from the crust and mantle. We make the common cosmochemical assumption that refractory elements are not fractionated ......................................................... highlands .......................................................................................................................................................................................................... rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 model 1 5 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 (a) nearside Feldsp eldspathic thic Highlands Ter errane ane South outh Pol ole-Ait itken en Ter errane ane Figure 2. Distribution of the three prominent lunar terranes: the FHT (white), the PKT (yellow) and the SPA Terrane (blue). This image is shown in two Lambert azimuthal equal-area projections centred over the nearside (a) and farside (b) hemispheres. The distribution of mare basalts is shown by thin black lines. relative to each other. Taylor [36] discusses this concept and shows that it is reasonable in the case of Mars. Determining the bulk lunar Al2 O3 requires knowing the masses of crust and mantle, which we report above (table 1), and the mean concentrations in the crust and mantle, which we evaluate in this section. (a) Crust The total abundance of Al2 O3 in the crust depends on the relative sizes of the three major crustal terranes shown in figure 2, and on their Al2 O3 contents. We divide each of the three terranes into three layers, though not all are necessarily present, and in some cases might be mixed. We list thorium concentrations, a refractory trace element, in each crustal column in figure 3 for reference, but do not try to compute bulk Th concentrations because the depth dependence of this element in each terrane is not easily constrained. (i) Upper layer A somewhat-mafic layer is exposed on the surface virtually everywhere in the highlands. Lunar meteorites [37] and orbital spectral observations [31] indicate that the uppermost crust is less feldspathic than the underlying anorthosite layer. Meteorite regolith breccias from the lunar highlands (defined as those with Al2 O3 more than 25 wt%) average 28 ± 1 wt% (compilation of published bulk meteorite analyses taken from [38]). Global spectral reflectance data for the highlands are in the range 4–6 wt% FeO, which corresponds to 24–28 wt% Al2 O3 using the linear inverse correlation between FeO and Al2 O3 in lunar rocks [39]. We adopt an Al2 O3 content of approximately 28 wt% for the upper layer. A somewhat lower value might be suitable, but this upper value maximizes Al2 O3 in the bulk crust to test the idea of lunar refractory enrichment. Its thickness over the FHT is approximately 10 km [31], but the revised, lower crustal thickness suggests a thinner value, which we assume to be approximately 5 km. We use a thickness of between 5 and 10 km in our mass-balance calculations. Our analysis of published [40,41] remote FeO measurements suggests that it is much thinner, and probably absent, in the PKT and SPA regions. In our mass-balance calculations of Al2 O3 , we allow for up to a 5 km thickness in the PKT and SPA terranes. 6 ......................................................... eldspathic thic Feldsp Highlands Ter errane ane farside rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 Procellarum ellarum KREEP Ter errane ane (b) Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 highlands 0–5 km SPA 21.7– 31.7 km 0–5 km 17.9– 27.9 km 0–5 km crustal thickness (km) area of moon (%) volume of moon (%) silicate mass fraction (%) 36.7 77.5 4.80 3.75 27.9 16.6 0.74 0.58 0–5 km upper layer 0–5 km ferroan anorthosite 10.5– 20.5 km LKFM 20.5 5.9 0.21 0.16 Figure 3. Key parameters for the three crustal terranes. Crustal thicknesses are derived from the GRAIL crustal thickness model 1. Thicknesses and compositions of layers within each stratigraphic column are based on assessments of remote sensing and sample data. Parameters for different crustal thicknesses are given in table 1. (ii) Anorthosite layer The presence of almost pure anorthosite in basin rings [31,42] suggests that the upper somewhatmafic layer is underlain by a substantial thickness of anorthosite. Spectral data indicate that the anorthosite contains less than 2 wt% FeO, corresponding to more than 31 wt% Al2 O3 , and contains no detectable mafic silicates. We assume an Al2 O3 content of 34 wt% to account for the presence of at least some olivine and pyroxene in the anorthosites as shown by Apollo samples [43]; for comparison, pure anorthite contains approximately 36 wt%. Anorthosite occupies most of the crustal column in the feldspathic highlands (figure 3), with its thickness depending on the thickness of layers above and below it, and on the total crustal thickness. Sample studies suggest that anorthosite is rare, but not absent in the PKT (anorthosites were collected at the Apollo 15 site), so we allow for an anorthosite layer up to 5 km thick in that terrane. Remote sensing observations do not unequivocally identify anorthosite in the SPA except in its outermost regions, but we allow for a layer of 5 km thick, recognizing that it could be essentially absent. (iii) Mafic lower layer It is plausible that a more mafic layer resides below the anorthosite layer, as first argued by Ryder & Wood [44] on the basis of the compositions of mafic impact melt breccias of the type ‘LKFM basalt’ (low-K Fra Mauro [45]) collected at the Apollo 15 and 17 landing sites. Using LKFM impact melt rocks as a guide, we estimate that the layer contains approximately 19 wt% Al2 O3 . Based on the abundance of LKFM at the landing sites in the PKT, including in ejecta from small and large craters, it seems likely that most of the crustal column in the PKT is composed of rock with such basaltic compositions [9]. We assume that mafic rock dominates the crust in the SPA basin, too, though with different abundances of incompatible elements. We wish to calculate the total contribution the crust makes to the bulk lunar composition of refractory elements. The four models shown in table 1 (corresponding to the four GRAIL crustal thickness models [30]), combined with the three layers possible in each terrane (figure 3 and above discussion) result in a large combination of possible Al2 O3 concentrations, so we present the lowest and highest values. The lowest (1.13 wt%) is obtained for the case of model 1 (thinnest crust ......................................................... PKT 7 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 5–10 km Al2O3 (wt%) Th (ppm) 28 0.5 34 0.05 19 5 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 3.5 8 maximum Al2O3 in mantle source 2.5 10 wt% Al2O3 in mare basalt 10% partial melting 2.0 9 wt% Al2O3 in mare basalt 5% partial melting 1.5 minimum Al2O3 in mantle source 1.0 0.2 0.3 0.4 0.5 pyroxene/(pyroxene + olivine) 0.6 0.7 Figure 4. Estimate of Al2 O3 in the mantle mare basalt source regions. Curves are shown for mare basalt magmas with 9 and 10 wt% Al2 O3 and for 5% and 10% partial melting. The horizontal axis shows the ratio of pyroxene to total mafic silicates. The more pyroxene in the source, the higher the Al2 O3 because it partitions into pyroxene in preference to olivine. Using the lowest and highest Al2 O3 values on the lines, we infer that the mantle contributes between 1.3 and 3.1 wt% to the bulk lunar Al2 O3 content. with the lowest mass fraction of total lunar silicate), with no upper layer or anorthosite in the PKT and SPA terrane, and a 10 km upper layer and 5-km lower layer in the feldspathic highlands. The highest crustal Al2 O3 contribution (1.85 wt%) is for model 4 (thickest crust with the highest silicate mass fraction), with no lower mafic layer and a 5-km upper layer in the highlands, and 5-km thick layer of anorthosite in the SPA terrane and PKT. This range in crustal Al2 O3 contribution is drastically different from other previous estimates that used a thicker crust (hence higher crustal mass fractions), which ranged as high as 3.3 wt% [11]. Considering that the Earth has a bulk Al2 O3 of 4.0 wt% (e.g. average of estimates in [46–48]), having more than 3 wt% Al2 O3 in the crust would make it unavoidable that the bulk Moon is enriched in refractory elements. (b) Mantle We determined the Al2 O3 concentration in the mantle from the Al2 O3 contents in mare basaltic lavas, the Al2 O3 contents in residual pyroxene obtained from multiple-saturation partial melting experiments ([49] and references cited therein) and estimates of the olivine/pyroxene ratio in mare basalt mantle source regions. Residual pyroxene in the multiple-saturation experiments average 3 ± 1 wt% Al2 O3 (1-σ ) with no apparent trend with pressure. From the range of Al2 O3 in mare basalts (8–10 wt%), assuming 5–10% melting, and that pyroxene constitutes 30–70% of the residual mafic assemblage (a range consistent with some interpretations of geophysical data [50]), we estimate that the Al2 O3 in their mantle sources range from 1.3 (for 8 wt% in the magma and 5% partial melting) to 3.1 (for 10 wt% in the magma and 10% partial melting), as summarized in figure 4. Strictly speaking, these Al2 O3 concentrations apply only to the portion of the mantle that melted to produce mare basalts. We have essentially no data on the mantle source regions of the Mg-suite and KREEP basalts, which contain more Al2 O3 than do mare basalts. An additional uncertainty in the composition of the mantle is that our data are based on mare basalts, which have been estimated to have formed at depths ranging from 150 to 500 km ......................................................... 3.0 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 Al2O3 of mantle source (wt%) mare basalt source regions in mantle Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 9 4 3 bulk M o on Al O 2 3 1.5 times E a rth 0.9 range in mantle 2 bulk M oon Al 2 O3 same as Earth 1 range in crust 0 0.5 1.0 1.5 2.0 2.5 Al2O3 (wt%) crust contribution Figure 5. Contributions to bulk Moon abundance of Al2 O3 as a function of Al2 O3 in the mantle and crust. Crustal contributions from consideration of models 1–4 (table 1) and thicknesses of lithologic layers (figure 3) range from 1.13 to 1.85 wt% Al2 O3 . Mantle concentrations are from figure 4. Dark lines represent 1.0 and 1.5 times the Earth abundance, with light lines showing increments of 0.1. The full range of possibilities for crustal and mantle concentrations of Al2 O3 straddle the line for which bulk lunar Al2 O3 is the same as in the bulk Earth. Strong enrichment relative to the Earth is ruled out, although enrichments of 10–20% are possible. [49] based on multiple-saturation experiments. These depths may be minima due to possible formation at a range of pressures and to fractional crystallization during migration to the lunar surface [51], so the mare basalt source regions could perhaps be representative deeper parts of the mantle. Some analyses of the Apollo seismic data are consistent with the mantle not having any seismic discontinuities [50,52], suggesting a uniform composition. Nevertheless, the composition of the lower mantle is uncertain, so it is worth examining how a reasonable range in Al2 O3 concentrations would affect the bulk mantle, and hence bulk Moon, composition. If the core is 380 km in radius, the crust is 34 km thick and the upper mantle extends to 500 km beneath the base of the crust, the lower mantle would constitute 35% of the total mantle. If the lower mantle contains 6 wt% Al2 O3 (the bulk Moon abundance if the Moon is enriched in refractory elements by 50%), the mantle contribution to the bulk Moon would be 2.9–4.1 wt% for an upper mantle Al2 O3 of 1.3–3.1 wt%, respectively. As can be inferred from figure 5, the upper part of this range could indicate a refractory-enriched Moon, although not enriched by 50%. In the more likely case that the lower mantle contains 4 wt% Al2 O3 , the amount in the bulk silicate Earth [46–48], the total mantle contribution changes from 1.3 to 3.1 wt% to 2.2 to 3.4 wt%, not significantly different from the ranges shown in figure 5. Finally, the lower mantle could be depleted in Al2 O3 if a large fraction of the initial, deep cumulates from the magma ocean did not rise to higher levels during mantle overturn. If the lower mantle contains only 1 wt% Al2 O3 , the mantle contribution would drop to 1.2–2.4 wt%, making a firmer case for no enrichment relative to the Earth. (c) Bulk silicate Moon We combine the crustal and mantle contributions in figure 5. The intersection of the mantle and crust ranges defines the probable lunar composition. The negative-sloping heavy lines represent the expected relations for bulk compositions exactly like that of Earth and 1.5 times terrestrial ......................................................... 1.5 1.4 1.3 1.2 1.1 1.0 5 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 Al2O3 (wt%) mantle contribution 6 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 Volatile elements are informative about planetary formation and the source materials for the planets. In the case of the Moon, they might hold especially useful information about the processes involved in formation of the Moon. As noted above, refractory elements have condensation temperatures greater than 1300 K (at a pressure of 10−4 bar [12]). The condensation temperature is really the temperature at which half the mass of an element is in the solid and is properly called the 50% condensation temperature. Moderately volatile elements (e.g. K and Rb) have 50% condensation temperatures in the range 1230–800 K; highly volatile elements (e.g. Tl and Cd) have condensation temperatures in the range 750 K (Bi) to 530 K (Tl). (Hg has a 50% condensation temperature of only 250 K, but far fewer measurements of Hg concentrations are available than for the other highly volatile elements.) If our model is correct and the Moon has terrestrial abundances of refractory elements, then we can determine the absolute abundances of volatile elements. We do this by searching for correlations between two elements with very similar geochemical behaviour, one refractory and the other volatile. For example, we can establish the K concentration in the Moon by its correlation with an incompatible refractory element such as La or Th. If the correlation is strong, we multiply the ratio of the volatile to the refractory element (e.g. K/La) by the abundance of the refractory, assuming that the lunar refractory abundance is the same as in the Earth. For terrestrial values, we have used those determined by McDonough & Sun [47]. For comparisons between refractory and volatile incompatible elements (these concentrate in magma in preference to major minerals), it is useful to use as wide a range in element concentrations as possible. Thus, we use concentrations in mare basalts (mostly low concentrations) and KREEP basalts (high concentrations) to define the volatile/refractory element ratio most reliably. Although ferroan anorthosites and Mg-suite rocks contain useful information as well, their cumulate nature masks the incompatible behaviour of trace elements. In addition, far fewer highland rocks have been analysed for volatile trace elements. The alkali elements K, Rb and Cs in mare and KREEP basalts are plotted against La in figure 6. In spite of the scatter at low concentrations, the linear fits are impressive because of the large range in concentrations and large number of data points. The correlations and the terrestrial bulk La value (0.648 ppm) allow us to estimate the bulk concentrations in the Moon (table 2). Uncertainties are calculated from the quality of the fit using the York method [54]. The 2-σ uncertainties for the Moon/Earth ratios incorporate the uncertainties in the terrestrial values [47]. Other moderately volatile elements do not have strong correlations with a refractory element. The highly volatile elements Tl (figure 7) and Cd are incompatible and correlate with the incompatible refractory element La (R2 , the square of the correlation coefficient, is 0.9 and 0.6, respectively). Br correlates adequately with Tl. It would be preferable to determine the Br abundance from a refractory element rather than Tl, which is already an estimate from its correlation with La, but no such correlations were observed. Bi is an incompatible element, but it does not correlate well (R2 < 0.5) with any refractory element. We estimate its abundance by determining its ratio with Th, using the mean abundance in rocks where both have been ......................................................... 5. Absolute abundances of volatile elements 10 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 values; lighter lines represent increments 0.1 times terrestrial abundances. The line representing no enrichment relative to the Earth roughly bisects the box defined by our estimates of Al2 O3 in the crust and mantle, indicating that it is likely that the Moon is not enriched in Al and by inference other refractory elements compared with the Earth. Nevertheless, the data allow for a modest enrichment or depletions of up to ±20% (relative), which we take as an estimate of the uncertainty of the estimate. When the crust was thought to be thicker and more aluminous than our analysis indicates (more than 2.5 wt% Al2 O3 ), enrichment relative to the Earth was far more likely. The similar aluminium abundances in the Earth and the Moon indicate that processes during formation of the Moon did not cause discernible enrichment of refractory elements in the Moon. This is not the case for volatile elements as shown below. Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 105 11 concentration (ppm) K 102 y = 0.24x R2 = 0.94 10 1 y = 0.009x R2 = 0.878 Rb 10–1 Cs 10–2 1 10 100 1000 La (ppm) Figure 6. Correlations of incompatible volatile alkali elements with incompatible refractory La. Knowing that refractory elements are similar in the Earth and the Moon allows us to determine K, Rb and Cs concentrations in the bulk Moon. The quality of the linear fits to the data (R2 values) decrease with decreasing concentration, perhaps indicating greater analytical uncertainty, particularly for Cs. Data from [53]. Table 2. Concentrations of moderately volatile and volatile elements in the bulk silicate Moon. element temp. (K)a conc. moderately volatile elements 2-σ methodb ratioc K (ppm) 1006 36.9 0.9 S K/La 0.205 0.01 0.068 Rb (ppm) 800 0.13 0.005 S Rb/La 0.24 0.02 0.056 Cs (ppb) 799 5.0 0.3 S Cs/La 0.28 0.03 0.026 Moon/Earth 2-σ Moon/CI .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... highly volatile elements .......................................................................................................................................................................................................... Sb (ppb) 0.02 0.01 S Sb/Tl 0.003 0.003 0.00013 .......................................................................................................................................................................................................... Bi (ppb) 746 0.02 0.006 A Bi/Th 0.008 0.007 0.00018 Zn (ppm) 726 0.40 0.15 A Zn/Sc 0.007 0.004 0.0013 Cd (ppb) 652 0.31 0.14 S Cd/La 0.009 0.004 0.00055 Br (ppb) 546 0.79 0.05 S Br/Tl 0.019 0.01 0.00274 Tl (ppb) 532 0.030 0.004 S Tl/La 0.010 0.005 0.00025 .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... .......................................................................................................................................................................................................... a 50% condensation temperature at 1 × 10−4 bars [12]. b S indicates use of slope of correlation line. A indicates average of ratio of abundances to another element. c Ratio indicates elements used to determine volatile element abundance. measured. Zn is a weakly incompatible to compatible element with behaviour similar to Sc (in the absence of sulfide phases), so we estimate its abundance by taking the ratio of the average concentrations of Zn and Sc. Lunar abundances of selected refractory and volatile elements are shown in figure 8 and table 2. We normalize them to terrestrial values as this is most informative about lunar formation by a giant impact. Though normalization to CI chondrites could be useful, planets do not necessarily have CI abundances of volatile elements, making it difficult to track changes in ......................................................... 103 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 y = 67.60x R2 = 0.95 104 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 6 12 y = 0.057x R2 = 0.87 Tl (ppb) 3 KREEP basalts 2 1 mare basalts 0 0 10 20 30 40 La (ppm) 50 60 70 80 Figure 7. Incompatible highly volatile element Tl plotted against incompatible refractory element La in lunar mare and KREEP basalts. La from [53]; Tl from [55–58]. R2 values are high, but driven by the highest values on the plot. 1 Moon/Earth 0.1 0.01 0.001 lunar abundances normalized to Earth abundances 0.0001 Sr Ba Be Eu Mg K Rb Cs Sb Bi Zn Cd Br Tl Figure 8. Abundances of volatile elements in the Moon, normalized to terrestrial abundances. Refractory elements shown for reference and assumed to be the same as in the Earth. Lunar depletions compared with the Earth are 0.24 for moderately volatile K, Rb and Cs, and 0.01 for highly volatile elements (values are geometric means). Uncertainties for refractory and moderately volatile elements are within the symbols. their abundances during formation of the Moon. As has been known since early analysis of Apollo samples, lunar abundances of volatiles are depleted relative to the Earth abundances. The moderately volatile elements are curiously inversely related to their condensation temperatures, but not strongly so (figure 8). Their mean abundance is 0.24 times the Earth. The alkalis are progressively depleted in order of condensation temperature when normalized to CI chondrites (table 2). ......................................................... 4 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 5 Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 On the basis of revised crustal thicknesses from GRAIL, our model calculations indicate that refractory elements are not enriched in the Moon compared with their abundances in the bulk silicate Earth. The similarity in abundances of refractory elements implies that processes in the protolunar disc did not lead to relative fractionation of refractory elements or to a general enrichment during condensation and partial crystallization of the disc before the Moon accreted from it. Chemical models of disc processes [59–61] need to account for the approximate similarity with terrestrial abundances of refractory elements. The volatile elements might allow us to constrain processes in the protolunar disc after a giant impact, ignoring the possibility that volatiles could have been lost from the magma ocean after the Moon formed. The moderately volatile elements have lunar concentrations only about 25% (table 2 and figure 8) of their concentrations in the Earth. Assuming that most of the material making up the Moon derived from the Earth, then 75% of the initial concentrations of moderately volatile elements were lost during the Moon-forming impact and subsequent evolution of the protolunar disc. About 99% of the highly volatile elements were lost during lunar formation, with little fractionation among them. Thus, there appears to be a rough correlation of percentage loss with volatility (moderately volatile versus highly volatile elements), but losses within each group were accompanied by only limited fractionation of one element from another. Preliminary results [62,63] address volatile loss from the protolunar disc through a combination of mixing with the proto-Earth (similar to that proposed by Pahlevan & Stevenson [59]) and subsequent hydrodynamic escape of the volatile-rich vapour portion of the disc. The concentrations of volatiles in the Moon would depend on their solubility in the magma portion of the protolunar disc (assuming minimal loss from the magma ocean). The discovery of H components (H2 , OH and H2 O) in the lunar interior [64] adds another dimension to the picture of lunar volatiles. H components are even more volatile than the highly volatile elements, yet may be present in higher abundance than the volatile elements. For example, estimates for the concentrations of H components (expressed as equivalent H2 O) in the source regions of pyroclastic glasses [65] and mare basalts [66] are approximately 100 ppm. These are similar to terrestrial Mid-Ocean Ridge basalt (MORB) source regions, 80–180 ppm [67], and about 20% of minimum bulk Earth concentration [68]. (Enriched MORB mantle sources have higher H2 O concentrations, 200–950 ppm [67].) Volatile elements and H2 O are not well correlated. In spite of the overall strong depletion of volatile elements in the Moon, pyroclastic glasses have concentrations of highly volatile elements that are close to those in MORBs. By contrast, low-Ti mare basalts, many of which appear to have similar H2 O concentrations as do the pyroclastic glasses, have only 1% of terrestrial abundances of highly volatile elements. It appears that H components and highly volatile elements are decoupled. Whether the lack of correlation between H components and volatile elements is due to the nature of the impactor (e.g. ice-rich), early additions to the Moon, or processes inside the complicated and evolving protolunar disc is not yet known. An additional complication is that H concentrations do not appear to be uniform inside the Moon, as shown by the significantly lower ......................................................... 6. Implications for lunar origin 13 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 The highly volatile elements are severely depleted [55]. The geometric mean of their Earthnormalized abundances is 0.01. Although the uncertainties are high, there is a hint of an inverse relation with condensation order, as observed for the alkali elements. Additional analyses of lunar samples are needed to establish whether this weak trend is statistically significant or not. It is possible that the highly volatile elements are uniformly depleted relative to the Earth. Wolf & Anders [55] and Taylor et al. [11] drew attention to the uniformity in CI-normalized concentrations in a large suite of low-Ti mare basalts, but did not determine abundances by element correlations. Our results indicate that CI-normalized abundances among the highly volatile elements differ by close to a factor of 10. By contrast, Earth-normalized abundances are similar to each other, although present at only 1% of terrestrial abundances. Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 Brad Jolliff and an anonymous reviewer were helpful, constructive and insightful. Funding statement. Support was provided by NASA grant no. NNX12L10G (G.J.T.). M.A.W. was supported by a grant from the French space agency, CNES. References 1. Wänke H, Dreibus G. 1982 Chemical and isotopic evidence for the early history of the Earth–Moon system. In Tidal friction and the Earth‘s rotation (eds P Brosche, J Sünderman), pp. 322–344. Berlin, Germany: Springer. 2. Ringwood AE, Seifert S, Wänke H. 1987 A komatiite component in Apollo 16 highland breccias: implications for the Ni-Co systematics and bulk composition of the Moon. Earth Planet. Sci. Lett. 81, 105–117. (doi:10.1016/0012-821X(87)90149-X) 3. Jones JH, Delano JW. 1989 A three-component model for the bulk composition of the Moon. Geochim. Cosmochim. Acta 53, 513–527. (doi:10.1016/0016-7037(89)90402-X) 4. O’Neill HStC. 1991 The origin of the Moon and the early history of the Earth—a chemical model. Part 1: the Moon. Geochim. Cosmochim. Acta 55, 1135–1157. (doi:10.1016/ 0016-7037(91)90168-5) 5. Warren PH. 2005 ‘New’ lunar meteorites: implications for composition of the global lunar surface, lunar crust, and the bulk Moon Meteorit. Planet. Sci. 40, 477–506. (doi:10.1111/j.19455100.2005.tb00395.x) 6. Morgan JW, Hertogen J, Anders E. 1978 The Moon: composition determined by nebula processes. Moon Planets 85, 465–478. (doi:10.1007/BF00897296) 7. Taylor SR. 1982 Planetary science: a lunar perspective, p. 481. Houston, TX: Lunar and Planetary Institute. 8. Kuskov OL, Kronrod VA. 1998 Constitution of the Moon: 5. Constraints on composition density temperature and radius of a core. Phys. Earth Planet. Inter. 107, 285–306. (doi:10.1016/ S0031-9201(98)00082-X) 9. Jolliff BL, Gillis JJ, Haskin L, Korotev RL, Wieczorek MA. 2000 Major lunar crustal terranes: surface expressions and crust-mantle origins. J. Geophys. Res. 105, 4197–4216. (doi:10.1029/ 1999JE001103) 10. Lognonné P, Gagnepain-Beyneix J, Chenet H. 2003 A new seismic model of the Moon: implications for structure, thermal evolution and formation of the Moon. Earth Planet. Sci. Lett. 211, 27–44. (doi:10.1016/S0012-821X(03)00172-9) 11. Taylor SR, Taylor GJ, Taylor LA. 2006 The Moon: a Taylor perspective. Geochim. Cosmochim. Acta 70, 5904–5918. (doi:10.1016/j.gca.2006.06.262) 12. Lodders K. 2003 Solar system abundances and condensation temperatures of the elements. Astrophys. J. 591, 1220–1247. (doi:10.1086/375492) 13. Hagerty JJ, Lawrence DJ, Hawke BR, Gaddis LR. 2009 Thorium abundances on the Aristarchus plateau: insights into the composition of the Aristarchus pyroclastic glass deposits. J. Geophys. Res. 114, E04002. (doi:10.1029/2008JE003262) 14. Toksöz MN, Dainty AM, Solomon SC, Anderson KR. 1974 Structure of the Moon. Rev. Geophys. 12, 539–567. (doi:10.1029/RG012i004p00539) 15. Nakamura Y. 1983 Seismic velocity structure of the lunar mantle. J. Geophys. Res. 88, 677–686. (doi:10.1029/JB088iB01p00677) 16. Goins NR, Dainty AM, Toksöz MN. 1981 Structure of the lunar crust at highland site Apollo station 16. Geophys. Res. Lett. 8, 29–32. (doi:10.1029/GL008i001p00029) 17. Kaula WM, Schubert G, Lingenfelter RE, Sjogren WL, Wollenhaupt WR. 1974 Apollo laser altimetry and inferences as to lunar structure. In Proc. 5th Lunar and Planetary Science Conf., vol. 3. Geochim. Cosmochim. Acta., Suppl. 5, 3049–3058. ......................................................... Acknowledgements. We thank our colleagues on the GRAIL science team for helpful discussions. Reviews by 14 rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 H2 O concentrations in rocks related to KREEP basaltic magmas [69], including KREEP basalts and Mg- and alkali-suite rocks. The details of lunar formation have not been worked out thoroughly, but all models need to satisfy central geochemical properties of the Moon: similar abundances of refractory elements in Moon and the Earth, and lunar depletion in moderately (25% compared with the Earth) and highly (1% compared with the Earth) volatile elements. Still to be determined are the details of the depletions and volatile element distributions (including H components) in the lunar interior. Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 15 ......................................................... rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 18. Haines EL, Metzger AE. 1980 Lunar highland crustal models based on iron concentrations: isostasy and center-of-mass displacement. In Proc. 11th Lunar and Planetary Science Conf., March 1980, pp. 689–718. Houston, TX: NASA. 19. Smith DE, Zuber MT, Neumann GA, Lemoine FG. 1997 Topography of the Moon from the Clementine lidar. J. Geophys. Res. 102, 1591–1611. (doi:10.1029/96JE02940) 20. Konopliv AS, Asmar SW, Carranza E, Sjogren WL, Yuan DN. 2001 Recent gravity models as a result of the Lunar Prospector mission. Icarus 150, 1–18. (doi:10.1006/icar.2000.6573) 21. Neumann GA, Zuber MT, Smith DE, Lemoine FG. 1996 The lunar crust: global structure and signature of major basins. J. Geophys. Res. 101, 16 841–16 863. (doi:10.1029/96JE01246) 22. Wieczorek MA, Phillips RJ. 1998 Potential anomalies on a sphere: applications to the thickness of the lunar crust. J. Geophys. Res. 103, 1715–1724. (doi:10.1029/97JE03136) 23. Wieczorek MA, Phillips RJ. 1997 The structure and compensation of the lunar highland crust. J. Geophys. Res. 102, 10 933–10 943. (doi:10.1029/97JE00666) 24. Wieczorek MA et al. 2006 The constitution and structure of the lunar interior. Rev. Mineral. Geochem. 60, 221–364. (doi:10.2138/rmg.2006.60.3) 25. Khan A, Mosegaard K, Rasmussen KL. 2000 A new seismic velocity model for the Moon from a Monte Carlo inversion of the Apollo lunar seismic data. Geophys. Res. Lett. 27, 1591–1594. (doi:10.1029/1999GL008452) 26. Khan A, Mosegaard K. 2002 An enquiry into the lunar interior: a nonlinear inversion of the Apollo lunar seismic data. J. Geophys. Res. 107, 3. (doi:10.1029/2001JE001658) 27. Chenet H, Lognonné P, Wieczorek M, Mizutani H. 2006 Lateral variations of lunar crustal thickness from the Apollo seismic data set. Earth Planet. Sci. Lett. 243, 1–14. (doi:10.1016/j.epsl.2005.12.017) 28. Hikida H, Wieczorek MA. 2007 Crustal thickness of the Moon: new constraints from gravity inversions using polyhedral shape models. Icarus 192, 150–166. (doi:10.1016/j. icarus.2007.06.015) 29. Ishihara Y et al. 2009 Crustal thickness of the Moon: implications for farside basin structures. Geophys. Res. Lett. 36, L19202. (doi:10.1029/2009GL039708) 30. Wieczorek MA et al. 2013 The crust of the Moon as seen by GRAIL. Science 6120, 671–675. (doi:10.1126/science.1231530) 31. Yamamoto S et al. 2010 Possible mantle origin of olivine around lunar impact basins detected by SELENE. Nat. Geosci. 3, 533–536. (doi:10.1038/ngeo897) 32. Garcia RF, Gagnepain-Beyneix J, Chevrot S, Lognonné P. 2011 Very preliminary reference Moon model. Phys. Earth Planet. Inter. 188, 96–113. (doi:10.1016/j.pepi.2011.06.015) 33. Vaughan WM, Head JW, Wilson L, Hess PC. 2013 Geology and petrology of enormous volumes of impact melt on the Moon: a case study of the Orientale basin impact melt sea. Icarus 223, 749–765. (doi:10.1016/j.icarus.2013.01.017) 34. Lawrence DJ, Feldman WC, Barraclough BL, Binder AB, Elphic RC, Maurice S, Miller MC, Prettyman TH. 2000 Thorium abundances on the lunar surface. J. Geophys. Res. 105, 20 307– 20 331. (doi:10.1029/1999JE001177) 35. Prettyman TH, Hagerty JJ, Elphic RC, Feldman WC, Lawrence DJ, McKinney GW, Vaniman DT. 2006 Elemental composition of the lunar surface: analysis of gamma-ray spectroscopy data from Lunar Prospector. J. Geophys. Res. 111, E12007. (doi:10.1029/2005JE002656) 36. Taylor GJ. 2013 The bulk composition of Mars. Chemie der Erde 73, 401–420. (doi:10.1016/j.chemer.2013.09.006) 37. Korotev RL, Jolliff BL, Ziegler RA, Gillis JJ, Haskin LA. 2003 Feldspathic lunar meteorites and their implications for compositional remote sensing of the lunar surface and the composition of the lunar crust. Geochim. Cosmochim. Acta 67, 4895–4923. (doi:10.1016/j.gca.2003.08.001) 38. Righter K, Gruener J. 2013 The lunar meteorite compendium. See http://curator.jsc.nasa. gov/antmet/lmc/index.cfm. 39. Haskin L, Warren P. 1991 Lunar chemistry, ch. 8. In Lunar sourcebook (eds GH Heiken, DT Vaniman, BM French), pp. 357–474. Cambridge, UK: Cambridge University Press. 40. Haskin LA, Gillis JJ, Korotev RL, Jolliff BL. 2000 The materials of the lunar Procellarum KREEP Terrane: a synthesis of data from geomorphological mapping, remote sensing, and sample analyses. J. Geophys. Res. 105, 20 403–20 415. (doi:10.1029/1999JE001128) 41. Lawrence DJ, Feldman WC, Elphic RC, Little RC, Prettyman TH, Maurice S, Lucey PG, Binder AB. 2002 Iron abundances on the lunar surface as measured by the Lunar Prospector gammaray and neutron spectrometers. J. Geophys. Res. 107, E125130. (doi:10.1029/2001JE001530) Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 16 ......................................................... rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 42. Hawke BR, Peterson CA, Blewett DT, Bussey DBJ, Lucey PG, Taylor GJ, Spudis PD. 2003 Distribution and modes of occurrence of lunar anorthosite. J. Geophys. Res. 108(E6), 5050. (doi:10.1029/2002JE001890) 43. Warren PH, Taylor GJ. 2013 The Moon, ch 2.8. In Treatise on geochemistry 1 (ed. A Davis), pp. 213–250. Oxford, UK: Elsevier-Pergamon. 44. Ryder G, Wood JA. 1977 Serenitatis and Imbrium impact melts–Implications for large-scale layering in the lunar crust. In Proc. 8th Lunar and Planetary Science Conf. Geochim. Cosmochim. Acta, Suppl. 8, 655–668. 45. Korotev RL. 2000 The great lunar hot spot and the composition and origin of the Apollo mafic (‘LKFM’) impact-melt breccias. J. Geophys. Res. 105, 4317–4345. (doi:10.1029/1999JE001063) 46. Jagoutz E, Palme H, Baddenhausen H, Blum K, Cendales M, Dreibus G, Spettel B, Lorenz V, Wänke H. 1979 The abundances of major, minor and trace elements in the Earth’s mantle as derived from primitive ultramafic nodules. In Proc. 10th Lunar and Planetary Science Conf. Geochim. Cosmochim. Acta, Suppl. 10, 2031–2050. 47. McDonough WF, Sun S-S. 1995 The composition of the Earth. Chem. Geol. 120, 223–253. (doi:10.1016/0009-2541(94)00140-4) 48. Taylor SR, McLennan SM. 2009 Planetary crusts, p. 378. New York, NY: Cambridge University Press. 49. Longhi J. 1992 Experimental petrology and petrogenesis of mare volcanics. Geochim. Cosmochim. Acta 56, 2235–2251. (doi:10.1016/0016-7037(92)90186-M) 50. Khan A, Maclennan J, Taylor SR, Connolly JAD. 2006 Are the Earth and the Moon compositionally alike? Inferences on lunar composition and implications for lunar origin and evolution from geophysical modeling. J. Geophys. Res. 111, E05005. (doi:10.1029/ 2005JE002608) 51. Asimow PD, Longhi J. 2004 The significance of multiple saturation points in the context of polybaric near-fractional melting. J. Petrol. 43, 2349–2367. (doi:10.1093/petrology/egh043) 52. Longonné P, Johnson CL. 2007 Planetary seismology. Treatise Geophys. 10, 69–122. (doi:10.1016/B978-044452748-6.00154-1) 53. Meyer CM. 2013 The Lunar Sample Compendium. See http://curator.jsc.nasa.gov/lunar/lsc/ index.cfm. 54. York D. 1969 Least squares fitting of a straight line with correlated error. Earth Planet. Sci. Lett. 5, 320–324. (doi:10.1016/S0012-821X(68)80059-7) 55. Wolf R, Anders E. 1980 Moon and Earth: compositional differences inferred from siderophiles, volatiles and alkalis in basalts. Geochim. Cosmochim. Acta 44, 2111–2124. (doi:10.1016/0016-7037(80)90208-2) 56. Gros J, Takahashi H, Hertogen J, Morgan JW, Anders E. 1976 Composition of the projectiles that bombarded the lunar highlands. In Proc. 7th Lunar and Planetary Science Conf. Geochim. Cosmochim. Acta, Suppl. 7, 2403–2425. 57. Ebihara M, Wolf R, Warren PH, Anders E. 1992 Trace elements in 59 mostly highland moon rocks. In Proc. 22nd Lunar and Planetary Science Conf., pp. 417–426. Houston, TX: Lunar and Planetary Institute. 58. Morgan JW, Ganapathy R, Higuchi H, Krahenbuhl U, Anders E. 1974 Lunar basins: tentative characterization of projectiles from meteoritic elements in Apollo 17 boulders. In Proc. 5th Lunar and Planetary Science Conf. Geochim. Cosmochim. Acta, Suppl. 5, 1703–1736. 59. Pahlevan K, Stevenson DJ. 2007 Equilibration in the aftermath of the lunar-forming giant impact. Earth Planet. Sci. Lett. 262, 438–449. (doi:10.1016/j.epsl.2007.07.055) 60. Pahlevan K, Stevenson DJ, Eiler JM. 2011 Chemical fractionation in the silicate vapor atmosphere of the Earth. Earth Planet. Sci. Lett. 301, 433–443. (doi:10.1016/j.epsl.2010.10.036) 61. Visscher C, Fegley Jr B. 2013 Chemistry of impact-generated silicate melt-vapor debris disks. Astrophys. J. Lett. 767, L12. (doi:10.1088/2041-8205/767/1/L12) 62. Desch SJ, Taylor GJ. 2011 A model of the Moon’s volatile depletion. In 42nd Lunar and Planetary Science Conf., Houston, TX, 14–18 March, abstract no. 2005. Houston, TX: Lunar and Planetary Institute. 63. Desch SJ, Taylor GJ. 2013 Isotopic mixing due to interaction between the protolunar disk and the Earth’s atmosphere. In 44th Lunar and Planetary Science Conf., Houston, TX, 18–22 March, abstract no. 2566. Houston, TX: Lunar and Planetary Institute. 64. Saal AE, Hauri EH, Cascio ML, Orman JAV, Rutherford MC, Cooper RF. 2008 Volatile content of lunar volcanic glasses and the presence of water in the Moon’s interoir. Nature 454, 192–195. (doi:10.1038/nature07047) Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017 17 ......................................................... rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130242 65. Hauri EH, Weinreich T, Saal AE, Rutherford MC, Van Orman JA. 2011 High pre-eruptive water contents preserved in lunar melt inclusions. Science 333, 213–215. (doi:10.1126/ science.1204626) 66. Tartèse R, Anand M, Barnes J, Starkey NA, Franchi IA, Sano Y. 2013 The abundance, distribution, and isotopic composition of Hydrogen in the Moon as revealed by basaltic lunar samples: implications for the volatile inventory of the Moon. Geochim. Cosmochim. Acta 122, 58–74. (doi:10.1016/j.gca.2013.08.014) 67. Luth RW. 2005 Chapter 2.07: mantle volatiles—distribution and consequences. In Treatise on geochemistry 2 (ed. RW Carlson), pp. 319–361. Amsterdam, The Netherlands: Elsevier. 68. Mottl M, Glazer B, Kaiser R, Meech K. 2007 Water and astrobiology. Chemie der Erde 67, 253–282. (doi:10.1016/j.chemer.2007.09.002) 69. Robinson KL, Taylor GJ. 2014 Heterogeneous distribution of water in the Moon. Nat. Geosci. 7, 401–408. (doi:10.1038/NGEO2173)