Laboratory Spectroscopic Detection of Hydration in Pristine Lunar Regolith ∗ Matthew R.M

Earth and Planetary Science Letters 390 (2014) 157–164

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Laboratory spectroscopic detection of hydration in pristine lunar regolith ∗ Matthew R.M. Izawa a, ,EdwardA.Cloutisa, Daniel M. Applin a, Michael A. Craig b, Paul Mann a, Matthew Cuddy a a Hyperspectral Optical Sensing for Extraterrestrial Reconnaissance Laboratory, Dept. Geography, University of Winnipeg, 515 Portage Avenue, Winnipeg, Manitoba R3B 2E9, Canada b Department of Earth Sciences/Centre for Planetary Science and Exploration, University of Western Ontario, 1151 Richmond Street, London, Ontario N6A 5B7, Canada article info abstract

Article history: Reflectance spectroscopy of Apollo lunar soil samples curated in an air- and water-free, sealed Received 21 August 2013 environment since recovery and return to Earth has been carried out under water-, oxygen-, CO2-and Received in revised form 30 October 2013 organic-controlled conditions. Spectra of these pristine samples contain features near 3 μm wavelength Accepted 7 January 2014 similar to those observed from the lunar surface by the Chandrayaan-1 Moon Mineralogy Mapper (M3), Available online xxxx Cassini Visual and Infrared Mapping Spectrometer (VIMS), and Deep Impact Extrasolar Planet Observation Editor: C. Sotin and Deep Impact Extended Investigation (EPOXI) High-Resolution Instrument (HRI) instruments. Spectral Keywords: feature characteristics and inferred OH/H2O concentrations are within the range of those observed by lunar regolith spacecraft instruments. These findings confirm that the 3 μm feature from the lunar surface results from + lunar hydration the presence of hydration in the form of bound OH and H2O. Implantation of solar wind H appears to reflectance spectroscopy be the most plausible formation mechanism for most of the observed lunar OH and H2O. Apollo samples © 2014 Elsevier B.V. All rights reserved.

1. Introduction are much lower than the values of up to 0.3–0.5 wt.% derived from remote spectral analysis (Clark, 2009; Pieters et al., 2009; Sunshine et al., 2009; McCord et al., 2011). Therefore, some ad- The prevailing view following analysis of lunar materials re- turned by the US Apollo and Soviet lunar programs was that the ditional hydroxyl and water must be contained in lunar regolith, moon is virtually anhydrous, with the possible exception of local probably as highly disordered hydroxyl groups at or near grain cold traps, e.g., near the poles (Urey, 1952; Watson et al., 1961; surfaces, and as strongly hydrogen-bonded adsorbed water. The Nozette et al., 1996; Feldman et al., 1998; Colaprete et al., 2010). objective of this study was to provide laboratory confirmation of Recently, the anhydrous model has been upended by the detec- remote detections of hydrogen in the form of hydroxyl groups tion of spectral features near 3 μm in wavelength, attributable to and/or bound molecular water in lunar regolith. To accomplish hydroxyl and water by the M3 (Pieters et al., 2009), VIMS (Clark, this goal, we have examined six Apollo regolith samples that have 2009) and HRI (Sunshine et al., 2009) instruments. In parallel to remained unopened since collection and return to Earth, under the remote detection of hydration in the lunar regolith, microana- oxygen-, water-, CO2, and organic-controlled conditions using in- lytical studies of lunar materials using ion microprobe and infrared frared reflectance spectroscopy. Our results show that these re- spectromicroscopy showed unexpectedly high levels of H in lunar golith samples, recovered from just below the lunar surface, con- minerals and volcanic glasses (Saal et al., 2008; McCubbin et al., tain adsorbed water and hydroxyl. The band depths, shapes, and 2010; Hui et al., 2013; Saal et al., 2013) and soil agglutinate par- feature positions observed here are similar to those observed for ticles (Liu et al., 2012). Some of the microanalytically-detected H the lunar surface by M3, VIMS and HRI. Implantation of hydrogen in lunar materials has been ascribed to solar wind bombardment by the solar wind has been suggested as the most likely source (Liu et al., 2012), and the remainder is ascribed to internal water, though the total concentrations inferred for regolith materials, for lunar hydroxyl and water, a process first suggested by Zeller et al. (1966). Following the M3, VIMS and HRI observations, McCord ranging up to hundreds of ppm in H2Oequivalent(Liu et al., 2012) et al. (2011) presented systematic arguments for the solar wind implantation model. Some contribution of internal (magmatic) hy- * Corresponding author. dration may also be remotely detectable in some areas of the lunar E-mail address: [email protected] (M.R.M. Izawa). surface (Klima et al., 2013).

Table 1 HOSERlab in-house 100 W quartz–tungsten–halogen light source Sample data for Apollo lunar regolith samples investigated in this study. directed through an open-air aluminum pipe at 30° from normal. Sample Generic Specific Parent Container Weight The spectrometer is equipped with a Designs & Prototypes cus- (g) tomized optical assembly that allows a spot size of ∼7mmto Apollo 11 10 084 2008 162 99425 0.536 be viewed by the instrument. The spectrometer uses a Michelson Apollo 12 12 030 185 7 99440 0.511 interferometer which contains the infrared optics, beam splitter, Apollo 15 15 041 71 20 PV 0.500 Apollo 16 61 221 176 25 925444 0.517 and the scanning mirror assembly. The internal mirrors are servo Apollo 16 64 801 176 4 925454 0.510 driven at a constant speed to produce the interferograms. Spec- − Apollo 17 71 061 14 5 98990 0.520 tral resolution for all measurements in this study was 4 cm 1. The thermo-electrically cooled detector (77 K) consists of InSb over mercury cadmium telluride (MCT). The servo, electronics, and 2. Samples and analytical methods wavelength calibration are referenced to a temperature controlled laser diode. Samples were placed on an adjustable stage and were Contamination of the Apollo samples with terrestrial water placed into focus, aided by a removable Modicam 10 video camera would compromise the goal of this study; therefore special proce- (Camea, Brno, Czech Republic) directed into the optical assembly. dures were implemented to preserve their integrity. Spectra were Reflectance spectra were acquired relative to a 35 mm diameter ® collected in a Plas-Labs 818 GBB glovebox with interior dimensions Infragold diffuse gold-coated standard (Labsphere, North Sutton, NH). Reference spectra were acquired with the same viewing ge- 97 × 152 × 79 cm; (Plas-labs Inc., Lansing, MI) under a dry N2 at- ometry and illumination as the samples and with the same num- mosphere, where dry N2 was passed through a series of Drierite filters. The glovebox also contained Drierite and Chemisorb to re- ber (100) of co-adds. move any remaining water and CO2 prior to opening of the sample containers. Humidity and O2 levels were monitored continuously 3. Results and discussion with a BW Technologies Gas Alert O2 Extreme oxygen meter, and an Extech Instruments RH20 humidity/temperature datalogger, and were 1% throughout the course of the measurements. Samples Spectra of pristine lunar regolith have spectral features due were only opened within the glovebox. Additional checks on sam- to water and hydroxyl that are similar to those detected by M3, ple integrity were provided by the inclusion of nanophase iron VIMS and HRI, displaying a broad asymmetrical feature centered (50 nm nominal particle size) and calcium sulfide in the glove box, near ∼3 μm and extending from ∼2.70 μm to ∼4.45 μm (Fig. 1). which are extremely sensitive to oxidation, hydration, or adsorp- A weak feature between ∼4.5 μm and ∼4.8 μm is ascribable to tion of water. Nanophase iron and CaS samples were monitored the first overtones of the silicate Si–O fundamental vibrations. The spectroscopically throughout the course of the experiment and by rise in reflectance towards longer wavelengths is due primarily to X-ray diffractometry (XRD) before and after and no evidence of rising reflectance approaching the principle Christiansen feature, oxidation, hydration or water adsorption within the glovebox envi- which occurs near ∼7.5–9.0 μm for many silicates (e.g., Conel, ronment was observed. For the XRD analysis, aliquots of nanophase 1969; Prost, 1973; Logan et al., 1978; Cooper et al., 2002). Be- iron and calcium sulfide were removed (through an independently cause the Infragold® standard and the regolith powders were held purged airlock) from the glovebox at the beginning of the ex- within the same environment, and a standard spectrum was mea- periment and immediately analyzed using a Bruker D8 Discover sured prior to each sample run, it is expected that temperature X-ray diffractometer with Co Kα1,2 radiation (λ = 1.78897 Å) op- differences between the sample and standard were small, of the erating at 40 kV and 40 mA. After the entire experiment, aliquots order of a few °C. Nevertheless, there may be a minute con- of nanophase iron and calcium sulfide that had remained exposed tribution of thermal excess affecting these spectra, the primary within the glovebox for the duration were again removed and an- effect of which would be to reduce the band depths. The pre- alyzed by XRD. cise location of the band centers of the hydration features, near An important issue which is not within the control of the ex- 2.957 μm are indicative of a high degree of hydrogen bonding, perimenters is that of pre-analysis contamination of the samples, as would be expected for adsorbed water remaining in material either during their return to Earth or subsequently, either during exposed to the hostile lunar surface environment (Clark, 2009). the 40–43 years of curation or during shipment to the laboratory. Continuum-removed spectra show that in detail, the 3-μm band The samples investigated in this study are splits of material that in these samples consists of a sharp inflection near ∼2.80 μm, at- (to the best of our knowledge) were not exposed to air or water tributable to structural hydroxyl, and a pair of features with local during either return or curation. The specific sample identities and minima near ∼2.95 μm and ∼3.14 μm attributable to bound wa- parentage of the materials investigated here are summarized in Ta- ter and hydroxyl (Fig. 2). These feature locations are very close ble 1. The samples were delivered to HOSERLab in sealed sample to the values of ∼2.81 μm, ∼2.95 μm and ∼3.14 μm observed vials (packed under high-purity N2 atmosphere at NASA Johnson by HRI (Sunshine et al., 2009). The feature near 3.14 μm is sim- Space Center) held in heavy plastic, within airtight steel containers ilar to that due to structurally bound water in silicate minerals further wrapped in heavy plastic, all held within a steel briefcase. (Farmer, 1974). The apparent band depths (calculated as depth be- There was no sign of damage to any of the packaging materials low a linear continuum anchored on the short-wavelength side when they were removed from the case within the purged glove- of the band, following the method described by McCord et al., box. We also opened the successive layers of packaging material 2011)at2.8 μm and 3.0 μm in our spectra of Apollo samples 3 with the O2 and RH meters nearby, to observe any spike in either are very similar to those observed by M near the lunar equator O2 concentration or relative humidity at the moment of open- (Fig. 3A). ing (which could indicate the presence of contamination), no such Quantification of the water content of the lunar regolith sam- spike was observed. ples in this study is subject to many uncertainties due to the Reflectance spectra from 2.0to5.2 μm were collected with mineralogical and textural heterogeneity of the samples. Using a Designs & Prototypes Model 102F spectrophotometer (Designs the Normalized Optical Path Length (NOPL) method of (Milliken & Prototypes, Simsbury, CT). Samples were illuminated with an and Mustard, 2005) we calculate between ∼0.08 and ∼0.16 wt.% M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164 159

Fig. 1. Reflectance spectra of pristine lunar regolith samples from 2.0 to 5.2 μm showing absorption features due to OH and H2O. The “3-μm” feature is a broad envelope ranging from ∼2.70 μm to ∼4.45 μm wavelength (cf. Fig. 1C of Sunshine et al., 2009 and Fig. 2B of Clark, 2009). This feature is a mixture of numerous O–H and H–O–H vibrational overtones and combinations, which are poorly resolved because of the vast diversity of possible O–H co-ordination environments in surficial and structural OH and adsorbed H2O. The weak feature between ∼4.5μmand∼4.8 μm is due to the first overtones of the silicate Si–O fundamental vibrations. (A) Spectra in absolute reflectance, shifted vertically by +0.25 each for clarity. (B) Spectra normalized to the reflectance averaged over all bands between 2617 nm and 2697 nm (2.617–2.697 μm; the range over which Pieters et al. (2009) calculated their approximate continuum).

+ H2O(∼800–1600 ppm) in our samples (Fig. 4), somewhat less of H , forming hydroxyl, some of which may form molecular wa- than the value of 0.3 wt.% calculated for high-latitude regions ter, if hydroxyl generation approaches the saturation rate (McCord (Sunshine et al., 2009). Although we report water concentrations et al., 2011). The presence of similar hydration features in the here, it must be acknowledged that the methods used both here diverse lunar regolith lithologies studied here suggests that the and by Sunshine et al. (2009) do not explicitly include hydroxyl, mechanisms of hydroxyl and water formation on the lunar surface which is likely to be the predominant source of the spectral are similar in different surface materials. Nevertheless, there are features near 3 μm. The water concentration values reported here some minor but signiﬁcant differences. There is a correlation be- are best regarded as upper bounds. tween the depth of the ∼3 μm feature (where ∼3 μm refers to The observations reported in this study are consistent with so- the continuum-removed band centers of the absorption feature at lar wind hydrogen implantation as the predominant mechanism the center locations listed in Table 2) and the surface maturity pa- for the formation of lunar hydration, with initial implantation rameter Is/FeO, the ratio of the value of the intensity in arbitrary 160 M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164

Fig. 2. Continuum-removed spectra of lunar regolith samples reveal details of band structure. These spectra are similar to those measured by Cassini VIMS (cf. Fig 2B of Clark, 2009) and Deep Impact/EPOXI (cf. Figs. 1D and 4A–D of Sunshine et al., 2009). All spectra show an inﬂection near 2.80 μm and local minima near 2.96 μm and 3.14 μm (marked by thin vertical lines), which are similar values to those in continuum-removed spectra of the lunar surface measured by Deep Impact/EPOXI (2.81 μm, 2.95 μm and 3.14 μm). units of the ferromagnetic resonance at g ∼ 2.1 to the FeO con- between C concentration and ∼3 μm feature depth may reﬂect in- centration (Morris, 1978)(Fig. 5A), consistent with hydroxyl and put from a variety of sources. Apollo 12 sample 12 030 appears water generation at or near the lunar surface. There is also a anomalously rich in C compared to the depth of the ∼3μmfea- weaker correlation between ∼3 μm feature depth and carbon con- ture (Fig. 5B), and is also anomalously immature (Fig. 5C) indi- tent (Fig. 5B), as well as a correlation between carbon content and cating that processes that incorporate carbon into the lunar re- the soil maturity parameter (Fig. 5C). Carbon is delivered to the lunar surface by the solar wind and by a variety of exogenous golith do not always lead to enhanced hydration or increased matter including cometary material, carbonaceous chondrites and soil maturity. There is no obvious correlation between Al2O3 con- interplanetary dust particles, therefore the observed correlation tent and ∼3μmfeaturedepth(Fig. 5D). To the extent that Al2O3 M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164 161

Fig. 3. Comparison of spectral properties derived from lunar regolith spectra reported in this study with observations from the Chandrayaan-1 M3 spectrometer (Pieters et al., 2009; McCord et al., 2011). (A) Comparison of the apparent band depths at ∼2.8μmand∼3.0 μm (in percent reﬂectance). The depths of the features were calculated from our observations to be most comparable with those shown by McCord et al. (2011), that is, as the depth below a straight-line continuum extrapolated from the short-wavelength edge of the band (cf. Fig. 9 and Section 6.2 of McCord et al., 2011). Light-toned shaded regions represent the range of 2.8 and 3.0μmfeaturedepthsfor M3 observations within 30° of the lunar equator (McCord et al., 2011), the dark-toned area is the region of overlap for both band depth parameters. Sample 61 121-176 (Apollo 16) falls slightly above the range of 3.0μmbanddepthsforM3 observations within 30° latitude of the lunar equator, but remains well within the range over all latitudes (up to ∼20%). (B) Comparison of relative 3-μm feature strength, calculated from our spectra following the method of Pieters et al. (2009), that is, relative band depth = 1 − (Rb/Rc ),whereRb is the average of channels at 2896 nm and 2936 nm, and Rc is the approximate continuum given by the average of channels at 2617 nm, 2657 nm, and 2697 nm (cf., Fig. 2C of Pieters et al., 2009). We have averaged over our channels in the ranges cited to produce numbers directly comparable to those of Pieters et al. (2009).

content corresponds to the presence of anorthosite-rich high- to hydration in diverse lithologies and the correlation between sur- lands material, this conﬁrms previous suggestions (Clark, 2009; face maturity and band depth are most consistent with solar wind Pieters et al., 2009; Sunshine et al., 2009; McCord et al., 2011) hydrogen implantation as the predominant source of lunar hydra- that composition (in terms of maria vs. highland material) does tion. Observations of similar hydration features over much of the not strongly affect the equilibrium OH/H2O content of lunar re- lunar surface (Pieters et al., 2009; McCord et al., 2011) also support golith. The fact that the anorthosite-rich Apollo 16 samples may the solar wind model. Traces of indigenous lunar hydrogen trapped plot on a different linear array in the space of calculated H2O con- in minerals and glasses may also make a minor contribution to tent vs. ∼3 μm band depth (Fig. 4) suggests the possibility that the hydration spectral features in some regions (Saal et al., 2008; higher-albedo anorthositic materials may be suppressing the mea- Hui et al., 2013; Klima et al., 2013), but the lack of strong sured ∼3 μm band depths, leading to a systematic underestimation compositional correlation between hydration features in labora- of lunar water content in highland areas. tory data (this study) or remote sensing argues against a major The surface and near-subsurface lunar regolith is certainly hy- spectral signature of lunar magmatic or metasomatic hydrogen drated, although the contributions of various mechanisms of hy- (Pieters et al., 2009; Sunshine et al., 2009; Clark et al., 2011; dration remain unquantiﬁed. The similarity of spectral features due McCord et al., 2011). 162 M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164

Fig. 4. Concentrations of H2O derived using the Normalized Optical Path Length (NOPL) (Milliken and Mustard, 2005) vs. continuum-removed band depths for the 3 μm feature for lunar regolith samples in this study. The H2O concentrations derived here are close to the modeled abundance of up to ∼770 ppm reported by Pieters et al. (2009) based on M3 data, and overlap the upper range of the estimate of 10–1000 ppm given by Clark (2009) based on VIMS spectra. It is possible that the anorthosite-rich Apollo 16 samples (diamonds) plot on a different linear array than the other samples (circles) suggesting the spectrally bright components may be suppressing the measured ∼3 μm band depths leading to a systematic underestimation of lunar water content in highland areas. Water concentrations obtained from these laboratory spectra are below the estimate of ∼0.3 wt.% for high-latitude regions reported by Sunshine et al. (2009), possibly reﬂecting lower concentrations of H2O and OH at the lower-latitude Apollo sampling sites. Sunshine et al. (2009) used the Effective Single-Particle Absorption Thickness (ESPAT) parameter of Milliken and Mustard (2005). Water concentrations derived by the ESPAT and NOPL techniques are comparable within the uncertainties of each method (Milliken and Mustard, 2005, 2007).

Table 2 Spectral characteristics and derived quantities.

Sample 10 084-2008 12 030-185 15 041-71 61 221-176 64 801-176 71 061-14 Apollo 11 Apollo 12 Apollo 15 Apollo 16 Apollo 16 Apollo 17 Band center (3rd order polynomial ﬁt, μm) 2.960 2.949 2.965 2.955 2.965 2.955 Band depth (non-continuum removed) 0.027 0.031 0.030 0.052 0.050 0.025 Band depth (continuum removed) 0.128 0.087 0.120 0.101 0.120 0.076 b 1 − Rb /Rc 0.176 0.114 0.059 0.134 0.155 0.100 Apparent deptha at 2.8 μm 0.017 0.021 0.023 0.033 0.034 0.017 Apparent deptha at 3.0 μm 0.026 0.031 0.032 0.051 0.050 0.024 Water from NOPL (wt.%) 0.16 0.10 0.14 0.09 0.12 0.08

a Values are derived following the methods in McCord et al. (2011), that is, depth below a straight-line continuum anchored to the short-wavelength side of the band. b The quantity 1 − Rb/Rc is calculated following Pieters et al. (2009), that is, averaging channels over the interval from 2.617–2.697 μm to approximate the continuum (Rc ) and over the interval 2.896–2.936 μm for the band intensity (Rb ).

4. Summary and conclusions Authors contributions

Laboratory spectra of pristine lunar regolith reported here E.A.C. designed the experiment. M.R.M.I., D.M.A., P.M. & M.C. conducted the measurements. M.R.M.I., M.A.C., D.M.A. & E.A.C. an- conﬁrm remote sensing observations of hydroxyl and water in alyzed the data. M.R.M.I. wrote the paper with contributions from the lunar near-surface. Our results cannot conclusively exclude all other authors. small contributions from terrestrial contamination or other lunar H reservoirs such as volatile-rich minor phases, as hydrous meteorites, comets and interplanetary dust particles also deliver Acknowledgements hydrogen to the lunar surface, some of which may become in- corporated into the regolith (Anders et al., 1973). Traces of in- M.R.M.I. gratefully acknowledges funding from the NSERC CRE- digenous (magmatic) lunar hydrogen trapped in minerals and ATE Canadian Astrobiology Training Program and the Mineralogi- glasses may also contribute to spectral features (Saal et al., 2008; cal Association of Canada. The University of Winnipeg’s HOSERLab Hui et al., 2013). Nevertheless, based on the similarity between was established with funding from the Canada Foundation for In- laboratory and remote spectra and the correlations between spec- novation, the Manitoba Research Innovations Fund and the Cana- tral and soil maturity parameters (Is/FeO and C concentration, dian Space Agency, whose support is gratefully acknowledged. This both of which are proxies for near-surface exposure), we conclude study was supported by research grants from NSERC, the Canadian that the predominant source of the observed 3-μm region spec- Space Agency and the University of Winnipeg. Thanks to Thomas tral features are hydroxyl and water native to the lunar surface, McCord and Paul Lucey for highly constructive reviews, and to most of which were created by solar wind hydrogen implanta- Christophe Sotin for editorial handling. This research has made use tion. of the NASA Astrophysical Data System. M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164 163

Fig. 5. Correlation plots of spectral parameters and sample characteristics previously reported in the literature for different splits of the same parentage as the samples studied here (Frondel et al., 1971; LSPET, 1973, 1974; Morris, 1978; Korotev, 1987; Korotev and Kremser, 1992; Korotev and Gillis, 2001). (A) Continuum-removed band depth of the ∼3 μm feature in Apollo lunar samples is correlated with lunar soil maturity parameter Is/FeO. The soil maturity parameter is the ratio between ferromagnetic resonance intensity at g ∼ 2.1(g is the dimensionless magnetic moment) to the FeO concentration (Morris, 1976, 1978). The relationship between increased hydration feature depth and maturity parameter supports solar wind H implantation as the predominant source of lunar surface OH and H2O. (B) The general correlation between C concentration and continuum-removed ∼3 μm feature depth may reﬂect input from solar wind implantation and a variety of infalling matter including cometary material, carbonaceous chondrites and interplanetary dust particles, therefore, Apollo 12 sample 12 030 appears anomalously rich in C compared to the continuum-removed band depth, indicating that the addition of carbon to the lunar regolith is not always accompanied by hydration. (C) There is a general correlation between soil maturity parameter and C concentration for these samples, but the relationship between C concentration, maturity and hydration is not simple, possibly due to overlapping or competing processes involved (solar wind implantation, infall of exogenous matter, and both creation and destruction/removal of OH/H2O). Sample 12 030 is again anomalous, with a very high C concentration for its maturity. (D) Although some correspondence between highland material rich in Al (primarily due to anorthitic plagioclase) and hydration spectral features has been noted (Sunshine et al., 2009; McCord et al., 2011), the lack of a strong correlation between continuum-removed ∼3μmfeaturebanddepthand Al2O3 content in these soil samples may indicate that a compositional or mineralogical control on OH/H2O generation is not the predominant mechanism responsible for the observed 3 μm feature enhancement in lunar spectra.

References Frondel, C., Klein, C., Ito, J., 1971. Mineralogical and chemical data on Apollo 12 lunar ﬁnes. In: 2nd Lunar and Planetary Science Conference, pp. 1407–1415. Anders, E., Ganapathy, R., Krähenbühl, U., Morgan, J.W., 1973. Meteoritic material on Hui, H., Peslier, A.H., Zhang, Y., Neal, C.R., 2013. Water in lunar anorthosites and the Moon. The Moon 8, 3–24. evidence for a wet early Moon. Nat. Geosci. 6, 177–180. Clark, B.E., Binzel, R.P., Howell, E.S., Cloutis, E.A., Ockert-Bell, M.E., Christensen, P., Klima, R.L., Cahill, J.T., Hagerty, J.J., Lawrence, D.J., 2013. Remote detection of mag- Barucci, M.A., DeMeo, F.E., Lauretta, D.S., Connolly, H.C., Soderberg, A., Hergen- matic water in Bullialdus Crater on the Moon. Nat. Geosci. 6, 737–741. rother, C., Lim, L., Emery, J., Mueller, M., 2011. Asteroid (101955) 1999 RQ36: Korotev, R.L., 1987. Mixing levels, the Apennine Front soil component, and com- Spectroscopy from 0.4 to 2.4 μm and meteorite analogs. Icarus 216, 462–475. positional trends in the Apollo 15 soils. In: 17th Lunar and Planetary Science Clark, R.N., 2009. Detection of adsorbed water and hydroxyl on the Moon. Sci- Conference. Houston, pp. E411–431. ence 326, 562–564. Korotev, R.L., Gillis, J.J., 2001. A new look at the Apollo 11 regolith and KREEP. J. Geo- Colaprete, A., Schultz, P.H., Heldmann, J.L., Wooden, D., Shirley, M., Ennico, K., Her- phys. Res. 106, 12339–12353. malyn, B., Marshall, W.G., Ricco, A., Elphic, R.C., Goldstein, D., Summy, D., Bart, Korotev, R.L., Kremser, D., 1992. Compositional variations in Apollo 17 soils and their G.D., Asphaug, E., Korycansky, D.L.D., Sollitt, L., 2010. Detection of water in the relationships to the geology of the Taurus–Littrow site. In: 22nd Lunar and Plan- LCROSS ejecta plume. Science 330, 463–468. etary Science Conference. Houston, pp. 275–301. Conel, J.E., 1969. Infrared emissivities of silicates: Experimental results and a cloudy Liu, Y., Guan, Y., Zhang, Y., Rossman, G.R., Eiler, J.M., Taylor, L.A., 2012. Direct mea- atmosphere model of spectral emission from condensed particulate mediums. surement of hydroxyl in the lunar regolith and the origin of lunar surface water. J. Geophys. Res. 74, 1614–1634. Nat. Geosci. 5, 779–782. Cooper, B.L., Salisbury, J.W., Killen, R.M., Potter, A.E., 2002. Midinfrared spectral fea- Logan, L.M., Hunt, G.R., Salisbury, J.W., Balsamo, S.R., 1978. Compositional implica- tures of rocks and their powders. J. Geophys. Res. 107, 5017–5034. tions of Christiansen frequency maximums for infrared remote sensing applica- Farmer, V.C., 1974. The Infrared Spectra of Minerals. Mineralogical Society Mono- tions. J. Geophys. Res. 78, 4983–5003. graphs. The Mineralogical Society, 527 pp. LSPET, 1973. The Apollo 16 lunar samples: Petrographic and chemical description. Feldman, W.C., Maurice, S., Binder, A.B., Barraclough, B.L., Elphic, R.C., Lawrence, D.J., Science 197, 23–34. 1998. Fluxes of fast and epithermal neutrons from lunar prospector: evidence LSPET, 1974. Apollo 17 lunar samples: Chemical and petrographic description. Sci- for water ice at the lunar poles. Science 281, 1496–1500. ence 182, 659–690. 164 M.R.M. Izawa et al. / Earth and Planetary Science Letters 390 (2014) 157–164

McCord, T.B., Taylor, L.A., Combe, J.-P., Kramer, G., Pieters, C.M., Sunshine, J.M., Clark, P.,Klima,R.L.,Kramar,R.,Kumar,S.,Livo,E.,Lundeen,S.,Malaret,E.,McCord, R.N., 2011. Sources and physical processes responsible for OH/H2O in the lunar T.B., Mustard, J.F., Nettles, J., Petro, N., Runyon, C., Staid, M.I., Sunshine, J.M., Tay- soil as revealed by the Moon Mineralogy Mapper (M3). J. Geophys. Res. 116, lor, L.A., Tompkins, S., Varanasi, P., 2009. Character and spatial distribution of 3 116–155. OH/H2O on the surface of the Moon seen by M on Chandrayaan-1. Science 326, McCubbin, F.M., Steele, A., Hauri, E.H., Nekvasil, H., Yamashita, S., Hemley, R.J., 2010. 568–572. Nominally hydrous magmatism on the Moon. Proc. Natl. Acad. Sci. USA 27, Prost, R., 1973. The influence of the Christiansen effect on I.R. spectra of powders. 11223–11228. Clays Clay Miner. 21, 363–368. Milliken, R.E., Mustard, J.F., 2005. Quantifying absolute water content of minerals Saal, A.E., Hauri, E.H., Cascio, M.L., van Orman, J.A., Rutherford, M.C., Cooper, R.F., using near-infrared reflectance spectroscopy. J. Geophys. Res. 110, E12001. 2008. Volatile content of lunar volcanic glasses and the presence of water in Milliken, R.E., Mustard, J.F., 2007. Estimating the water content of hydrated minerals the Moon’s interior. Nature 454, 192–195. using reflectance spectroscopy. I. Effects of darkening agents and low-albedo Saal, A.E., Hauri, E.H., Van Orman, J.A., Rutherford, M.J., 2013. Hydrogen isotopes materials. Icarus 189, 550–573. in lunar volcanic glasses and melt inclusions reveal a carbonaceous chondrite Morris, R.V., 1976. Surface exposure indicies of lunar soils: A comparative FMR heritage. Science 14, 1317–1320. study. In: 7th Lunar and Planetary Science Conference, pp. 315–335, abstract. Sunshine, J.M., Farnham, T.L., Feaga, L.M., Groussin, O., Merlin, F., Milliken, R.E., Morris, R.V., 1978. The surface exposure/maturity of lunar soils – Some concepts A’Hearn, M.F., 2009. Temporal and spatial variability of lunar hydration as ob- and Is/FeO compilation. In: 9th Lunar and Planetary Science Conference. Perga- served by the Deep Impact spacecraft. Science 326, 565–568. mon Press, Houston, pp. 2287–2297. Urey, H.C., 1952. The Planets: Their Origin and Development. Yale University Press, Nozette, S., Lichtenberg, C.L., Spudis, P., Bonner, R., Ort, W., Malaret, E., Robinson, M., New Haven, CT. Shoemaker, E.M., 1996. The Clementine bistatic radar experiment. Science 274, Watson, K., Murray, B.C., Brown, H., 1961. The behavior of volatiles on the lunar 1495–1498. surface. J. Geophys. Res. 66, 3033–3045. Pieters, C.M., Goswami, J.N., Clark, R.N., Annadurai, M., Boardman, J., Buratti, B., Zeller, E.J., Ronca, L.B., Levy, P.W., 1966. Proton-induced hydroxyl formation on the Combe, J.-P., Dyar, M.D., Green, R., Head, J.W., Hibbitts, C.A., Hicks, M., Isaacson, lunar surface. J. Geophys. Res. 71, 4855–4860.