The Apollo-derived tenet of an anhydrous Moon has been contested following measurement of water in several lunar samples that require water to be present in the lunar interior. However, significant uncertainties exist regarding the flux, sources and timing of water delivery to the Moon. Here we address those fundamental issues by constraining the mass of water accreted to the Moon and modelling the relative proportions of asteroidal and cometary sources for water that are consistent with measured isotopic compositions of lunar samples. We determine that a combination of carbonaceous chondrite-type materials were responsible for the majority of water (and nitrogen) delivered to the Earth–Moon system. Crucially, we conclude that comets containing water enriched in deuterium contributed significantly <20% of the water in the Moon. Therefore, our work places important constraints on the types of objects impacting the Moon ∼4.5–4.3 billion years ago and on the origin of water in the inner Solar System.
Lunar sample studies have focused on the measurement of H (reported as equivalent amounts of OH or H2O) in lunar volcanic glass beads1,2,3, melt inclusions1,2,4 and apatite in lunar basalts and highlands samples5,6,7,8. From this compilation of work the bulk silicate Moon (BSM) is estimated to contain between ∼10 and 300 p.p.m. H2O4,9,10. A few of these studies reported the identification of a primordial H-isotopic signature2,7,8, corresponding to Earth-like δD values between ∼−200 and ∼+180‰ (where δD=[(D/H)sample/(D/H)SMOW—1] × 1,000; SMOW: standard mean ocean water). Therefore, the similar nature of water (the term water is hereafter used to refer to H2O equivalent) in the Earth and Moon suggests that water locked up inside of the Moon was either inherited directly from the proto-Earth or delivered to the Earth–Moon system shortly after the formation of the Moon (Fig. 1) through impacting carbonaceous chondrite (CC)-type asteroids, since bulk D/H ratios for most CC-types11 are consistent with those of terrestrial12,13 and lunar water (Fig. 2).
Models for the differentiation of the Moon invoke the presence of a lunar magma ocean (LMO) and place the lifetime of this LMO to between ∼10 and 200 Ma after the Moon-forming event (as reviewed by Elkins-Tanton et al.14), which necessitates that volatile accretion occurred relatively early in the Moon’s geological history (Fig. 1). The final dregs of the LMO, enriched in incompatible elements such as potassium (K), rare-earth-elements (REEs) and phosphorous (P)—collectively termed urKREEP—should have contained the vast majority of the lunar water, as OH would have been incompatible in the crystallizing mantle cumulates and the crust10. Lunar samples with a significant KREEP component, such as Mg-suite rocks and KREEP basalts7,8, are characterized by H-isotopic compositions that are similar to those estimated for the mantle source regions of picritic glasses and mare basalts2,3,6, suggesting that the crystallization of the LMO did not result in significant fractionation of the H isotopes of water. To retain volatiles in the LMO, a conductive lid is required, to have prevented significant loss of water via degassing of H-bearing species, and is consistent with the formation of an early anorthosite crust a few kilometres thick in the first few thousand years of the Moon’s history14. It is likely that the lunar crust then grew by continuous underplating of more anorthositic materials over a period of between ∼10 and ∼200 million years until the LMO completely solidified, leading to the mantle cumulate overturn (see review by Elkins-Tanton et al.14). This time period between lunar accretion and the overturn of mantle cumulates permits any water (and other volatiles) accreted to the LMO to be incorporated into the mantle source regions of mare basalts and picritic glasses (>∼100 km deep15), and the urKREEP liquid provided that impactors were able to breach the lunar crust. Crystallization of an insulating crust likely started only after ∼80% LMO crystallization14. During its initial growth the crust was likely hot and tenuous and perhaps even re-melted if tidal heating played an important role14, meaning not only that impactors could have reached the LMO melt pool but also that many early impact structures could have been erased during crustal underplating and modification. Therefore, volatile delivery through the earliest lunar crust into the LMO could have balanced out any potential loss of volatiles (accreted or indigenous) by degassing (before 80% solidification).
Studies that have modelled the flux of impactors to the Moon posit that it was likely that the time period between lunar accretion and differentiation received relatively fewer impact events than during the late heavy bombardment (LHB) (for example, Morbidelli et al.16), but importantly this was not a quiescent time. Indeed, there are ∼12 preserved Pre-Nectarian lunar basins that have excavations depths around or greater than 34–45 km (ref. 17), which corresponds to the current estimates for lunar crustal thickness based on observations by the GRAIL spacecraft18. This demonstrates that large impacts occurred early in the Moon’s history (albeit infrequently compared with the LHB epoch19). In summary, the time interval for water delivery to the LMO, hereafter referred to as the late accretion window ‘LAW’, is conceptually restricted to between time zero (Moon formation) and up to ∼200 Ma thereafter. Importantly, the bulk of lunar water was delivered during LAW, which is distinct from a much younger basin-forming epoch on the Moon that is commonly referred to as the LHB, lunar cataclysm, or inner Solar System cataclysm16,20,21,22.
It has been suggested that water could have been accreted together with the Moon-forming material in the so-called ‘cold start’9. Whilst it is plausible that some water may have been retained since the Moon’s formation23, it is highly unlikely that all of BSM water (up to 300 p.p.m. H2O) was accreted in this manner and survived LMO processing. Therefore, we envisage two scenarios that can account for the bulk water inventory of the Moon, the first that water was accreted solely during the LAW and no water was inherited from the Moon-forming material (scenario 1), and the second that a portion (we consider up to 25%) of the BSM water was accreted with Moon-forming materials and was then supplemented by water delivery during the LAW (scenario 2).
In this study we utilize estimates of the water content of the BSM4,9,10 and its H-isotopic composition2,3,6,7,8 together with the current database for the bulk water contents of chondritic meteorites (Table 1, see Methods section), to determine the amount of water added to the early Moon by impacting objects. We consider a range of possible accretion scenarios during LAW to investigate the implications for the mass of water accreted and populations of accreting objects. This is constrained by the estimates for the proportion of lunar mass that was likely added post-lunar core-formation from studies of highly siderophile elements (HSEs) in lunar basalts (up to 0.02% lunar mass24,25) and highly volatile elements in mare basalts and picritic glasses (up to 0.4% lunar mass4,9). Our results indicate that carbonaceous chondrite types of impactors were the most important contributors of water to the lunar interior and that comets (with heavy water, Fig. 2) played a very small role.
Mass of asteroidal materials added during the LAW
Based on the BSM water estimates of McCubbin et al.10 (∼10–100 p.p.m. H2O), and Chen et al.4 and Hauri et al.9 (∼100–300 p.p.m. H2O), we have calculated the amount of material supplied by chondritic impactors to accrete ∼10, 100 and 300 p.p.m. bulk H2O for LMO depths of 400 and 1,000 km (to match the various LMO depths estimated in previous studies3,14,26,27), and have recast these estimates into the context of how much lunar mass each would represent (Table 2 and Fig. 3). The types of carbonaceous chondrites are defined in the caption accompanying Fig. 2.
In scenario 1, for example, between 3.94 × 1016 kg (∼1 p.p.m. H2O) and 2.05 × 1018 kg H2O (∼52 p.p.m. H2O) could have been accreted by CI-type CCs into a 400-km-deep LMO to remain below the addition threshold of <0.02% lunar mass (Table 1). In fact, using this <0.02% lunar mass threshold prohibits the addition of 300 p.p.m. H2O by asteroids alone (Table 2). On the other hand, the accretion of up to 0.4% lunar mass in some cases, allows for the addition of H2O in amounts greater than the high-end of BSM H2O estimates (Table 2).
Importantly, because equilibrated ordinary (OC) and enstatite chondrites (EC) contain very little water (Table 1) the mass of such types of chondritic material added to accrete even 10 p.p.m. H2O greatly exceeds the mass constraints imposed by HSE and highly volatile element abundances (Table 2). As a consequence, although ECs are considered an ideal geochemical match for the BSM and the bulk silicate Earth, and Os isotope signatures may suggest a large contribution of OC- and EC-type objects in materials accreted to the Earth–Moon system during late accretion28,29, it is clear that these types of objects are unlikely to have contributed significant amounts of H2O to the lunar interior. The accretion during the LAW of CC material is consistent with the impact of CC-type asteroids during the LHB20. HSE data have been interpreted to mean that EC and OC impactors also played an important role during the LHB30,31, which may reflect a change in impactor populations over the first ∼500 million years of the Moon’s evolution and, thus, dynamical mixing of material in the Solar System. The same data have also been interpreted, however, to represent a mixture of OC and CC impactors together with differentiated iron impactors29,30,31,32, in which case CC asteroids remained an important source of impactors.
Assessing the contribution of cometary water during LAW
The first scenario in this study assumes that the Moon accreted completely dry or that any primordial water was completely lost very early in the Moon’s history, and that late-accretion delivered all of the water in the BSM. Considering a BSM of 100 p.p.m. H2O (the most recent BSM water estimates overlapping at this abundance4,9,10) and a LMO depth of 400 km (see Methods section), a simple two-component mixing model (equation 1) permits a number of combinations for the mixing of water from bulk CCs (CI, CO, CV, CM type) and deuterium-depleted Kuiper belt (DDK) comets that all result in bulk BSM water δD values between ∼−50 and ∼+80‰ (Fig. 4), consistent with existing isotopic limits2,3,6,7,8 (see Methods).
Where; DR is the resultant D/H ratio, H2Ot refers to the total mass of water added (kg), fH2O is the proportion (fraction of total) of water accreted by a given type of impactor denoted with an A or B, characterized by D/H ratios DA or DB, respectively. The resultant isotopic ratios were subsequently converted into delta notation to ease comparisons.
However, because cometary water, excluding comets 103P and 45P, has significantly elevated δD values compared with chondrites (see Methods section), the amount of cometary water that can be added to the lunar interior is far more restricted. In such cases, ∼2% H2O from Oort cloud comets can be mixed with 98% H2O derived from CI-chondrites, or ∼1% H2O is allowed from D-enriched Kuiper belt (DEK) comets (Table 3). Mixing of cometary water with predominantly CV-, CM- or CO-type CCs, respectively, also result in minor cometary contributions of ∼6% for DEK comets and ∼15% from Oort cloud-like objects (Fig. 4 and Table 3).
In scenario 2, we consider that the Moon accreted with between 1 and 25% of BSM H2O, with an initial δD value of −200‰ (refs 2, 3, 6, 7, 8), and reached a final BSM water content of 100 p.p.m. H2O (in agreement with the constraints imposed by modelling of Elkins-Tanton and Grove26) (equation (2)). Note that the lower the initial δD value is, the higher the maximum possible amount of H2O added from a cometary source.
Where; H2Oi refers to the initial mass of water present in the LMO (kg), which varied between 1% and 25% of the total amount of water (H2Ot), Di refers to the initial H-isotopic composition of the water in the LMO, H2OLA refers to the mass of water added during LAW (kg), fH2O is the proportion (fraction of amount added during LAW) of water accreted by a given type of impactor denoted with A or B, respectively, and all other parameters are the same as in equation (1).
Mixing of water from CI-type CCs with water from CO, CM, CV-type CCs and DDK comets in different proportions can yield final δD values <+100‰ (Table 3), satisfying isotopic constraints2,3,6,7,8. If deuterium-enriched comets were involved in the addition of water during LAW then up to ∼14% H2O could have originated from Oort cloud-type comets and ∼5% H2O from DEK comets if the remaining water is assumed to come from CI-type chondrites (Fig. 5a), proportions that increase to ∼7 and ∼19% H2O originating from DEK and Oort cloud comets, respectively, if the remaining water came from CM-type chondrites (Fig. 5b and Table 3). Significantly, in the case where CO-type CCs dominated the impactor population (Fig. 5c) up to ∼27% H2O could have been contributed by Oort cloud-like comets and up to ∼10% H2O from DEK comets (Table 3). In addition, we considered that the LMO underwent some degassing26,33 before prior to late addition (see Methods section) and found that the results limit the amount of initial water present before late addition (assuming degassed H-isotopic composition of ∼+800‰ (ref. 8)), which further restricts the amount of water contributed by D-rich comets (Table 4).
To a first order, a dominantly asteroidal source of water accreted during the LAW is similar to the dominant source inferred for the subsequent basin-forming epoch of the Moon, based on geochemical22,34 and mineralogical20 markers, implying that asteroids and not comets dominated the impactor population hitting the Moon during its first 500 million years of geological history35. Although the source of objects delivered to the Earth–Moon system may have largely remained the same, the dynamical delivery could have been much different, as the basin-forming epoch appears to have been characterized by asteroids with higher impact velocities compared with those during the early accretional epoch (LAW) (Bottke et al.19 and references therein). We note that the accretion of CC-type material during the LAW, that is required to account for the lunar H-isotope and water abundances in the lunar interior, is distinct from a late veneer dominated by EC- and OC-type objects that has been inferred from Os isotope studies of terrestrial rocks28,29. This dichotomy could imply that either the largest impactors that preferentially impacted the Earth due to stochastic accretion19 were OC- and EC-type objects, while the population of smaller impactors included abundant CC objects that preferentially impacted the Moon, or that while the accreting objects had OC- and EC-type Os isotope signatures, they had CC-type water abundances and H-isotope signatures, consistent with other evidence that support the presence of planetesimal-types that are not represented in modern meteorite falls36.
On the basis of combined N isotopes and C/N ratios, the lunar interior seems to be best matched with CO-type CCs37. In terms of N, we can estimate the amount of N delivered to the LMO based on our calculations of the mass of chondritic material added during LAW for each chondrite type (Table 2, restricted to consideration of material added to achieve a BSM with ∼100 p.p.m. H2O), and the average N contents in the different chondrite types (Table 1). Results yield N contents for the BSM of ∼0.95 to 2.21 p.p.m. for LMO depths of 400 and 1,000 km (Table 5), which is consistent with an average of ∼1 p.p.m. N in the lunar mantle as estimated by Füri et al.38 If minimal fractionation of 15N/14N occurred during basalt ascent and eruption38,39, then the N-isotopic composition measured in mare basalts of between ∼−2 and +27‰ (refs 37, 38) should reflect the isotopic composition of the lunar mantle source regions from which the basalts were derived. If all of the N in the LMO was added during LAW (scenario 1) then mixing of CI-, CM-, CO- and CV-type CCs provide suitable matches for the origin of lunar N (Methods section and Table 1) and restrict cometary contribution of N to as much as ∼5% (Table 5). This is comparable to scenario 2, where considering the lowest mare basalt δ15N value of −2 ‰ (ref. 37) to represent the N-isotopic signature of N accreted with Moon-forming materials before LAW, then up to ∼6% N could have been delivered by comets (Table 5). Still using scenario 2 but considering that the isotopic signature of N accreted when the Moon formed was consistent with that of the primordial Earth mantle, which is thought to be as low as ∼−40 ‰ (reviewed by Füri and Marty40), increases the possible cometary contribution to a maximum of ∼8% of total lunar interior N.
Therefore, it appears that in order to add the appropriate amounts of H2O and N to the lunar interior during LAW, while keeping the H- and N-isotopic compositions in line with sample measurements, requires the contemporaneous accretion of the various types of CCs (mostly CO, CI, CM and CV), together with <∼20% cometary input.
The very-low apparently indigenous lunar δD values of ∼−600‰ reported by Robinson et al.41 are quite anomalous. In fact, such low values have yet to be observed in terrestrial rocks, the lowest δD values being ∼−220‰ for Earth’s depleted upper mantle13. Therefore, such low values reported for an intrusive lunar rock41 could potentially indicate the presence of a proto-solar component (δD value of ∼−900‰ (ref. 42)) in the lunar mantle, in a similar way that has been used to explain Earth’s noble gas record43,44 and recent data from primitive terrestrial basalts13. If this source is indeed required to explain the extremely low H-isotopic compositions observed41, then this signature must have been added to the Moon from either material accreted from the proto-Earth or by the impactor involved in the formation of the Moon. Alternatively, such a low δD value could have arisen from contamination of a water-poor magma by solar wind-derived H present in the lunar regolith during magma ascent and emplacement.
The mantles of the early-Earth13,45 the Moon2,3,6,7,8 and Mars46 all seem to be characterized by water with remarkably similar H-isotopic compositions, suggesting that the same types of accreting objects as those modelled here for the Moon delivered a vast majority of the water to the rocky planets in the inner Solar System. Likewise, isotopic analyses of differentiated meteorites47 suggest that the interior of the eucrite parent body, probably the asteroid 4-Vesta, contained water characterized by H-isotopic signatures indistinguishable to that of the terrestrial planetary objects in the inner Solar System (Fig. 2).
Regarding the timing of volatile accretion to the Moon, major constraints are imposed by the time interval between lunar accretion and solidification of the LMO and overturn of mantle cumulates (∼10–200 Myr after Moon formation14) since there is water in the mantle source regions of the mare basalts4,6 and picritic glasses1,2,3,9. It is possible that some water with Earth-like H-isotopic characteristics may have been accreted with the Moon-forming material, to explain the H-isotopic composition of water in the source regions for the picritic glasses (that were likely deep enough in the lunar mantle to have escaped processes such as degassing of H2 (ref. 48)). To account for the H-isotopic composition of the primordial lunar mantle and the urKREEP reservoir together, it is necessary that any initial lunar water was complemented by the continuous accretion of water from CC-type objects to the LMO during LAW. Incidentally, late accretion of material to the Earth–Moon system did not only affect their volatile inventories49 but could also explain the recently identified difference between the bulk silicate Earth and BSM in terms of W isotopes50,51.
In summary, the lunar interior is characterized by an average bulk water content of the order of ∼100 p.p.m. based on samples studies4,9,10 and we propose that the majority of this water was delivered during the ca. 10–200 million year-long LMO phase of lunar differentiation, with the possibility that some of this water was inherited proto-Earth-derived materials. This late accretion was dominated by water-rich carbonaceous chondrite asteroids, on the basis of H and N data, likely comprising a mixture of CO-, CI-, CM- and possibly CV-type CCs, with a minor contribution of water from deuterium-rich Oort cloud or Kuiper belt comets (much less than 20% of the total water in the BSM). Our work, therefore, confirms that the inner Solar System received a similar flux of asteroidal and cometary material for much of its early history from 4.5 billion years (this work and Dauphas et al.52) to at least 3.9 Gyr (ref. 20) (LHB).
Mass balance calculations
For mass balance calculations, we considered a lunar radius of 1,737.4 km, a mantle density of 3,300 kg m−3, and varied the depth of the LMO from 400 to 1,000 km. We then calculated the mass of chondritic material added to the lunar interior that would result in BSM water contents of 10, 100 and 300 p.p.m. H2O, and considered both the addition of up to 0.02% lunar mass24,25 and up to 0.4% lunar mass4,9. Note that for a 400-km-deep LMO, a BSM with 10 p.p.m. H2O equates to 3.94 × 1017 kg H2O, 100 p.p.m. equates to 3.94 × 1018 kg H2O, and 300 p.p.m. equates to 1.18 × 1019 kg H2O (6.70 × 1017 kg H2O, 6.70 × 1018 kg H2O, and 2.01 × 1019 kg H2O, for BSM H2O contents of 10, 100 and 300 p.p.m., respectively, for a 1,000-km-deep LMO).
Data for the average water contents of different carbonaceous, ordinary and enstatite chondrites that were used in these calculations are shown in Table 1. For carbonaceous chondrites, CI-, CM-, CV-, CO- and CR- refer to the different groups, named according to one prominent meteorite of the group, respectively Ivuna, Mighei, Vigorano, Ornans and Renazzo. OC stands for ordinary chondrites. The results of these calculations, presented in Table 2, allowed us to identify suitable impactors for the modelling presented below. For example, to add 10 p.p.m. H2O to the BSM through the accretion of enstatite chondrites alone implies the addition of >1% lunar mass for both 400- and 1,000-km-deep LMO, and, therefore, ECs were not deemed important carriers of water to the lunar interior. The same was also true for equilibrated ordinary chondrites (Table 2). Un-equilibrated ordinary chondrites were also not considered further on the basis of them having prohibitively high H-isotopic compositions (Table 1) for the mixing calculations. Note that all of the scenarios in this work consider that water was efficiently delivered to the Moon regardless of if it originated in asteroids or comets, that is, we have not attempted to calculate the proportions of volatiles lost during impact into the crystallizing LMO.
In the mixing models, literature data for the water abundances and D/H ratios of bulk carbonaceous chondrite meteorites were utilized. Carbonaceous, ordinary and enstatite chondrite meteorites are characterized by δD values mostly between ∼−200 and +300‰ (refs 11, 53, 54, 55). There is also a possibility that the asteroid parent bodies to the CCs may have once contained ice11. If the asteroids that delivered water to the Moon contained ice, in addition to their bulk water contents hosted in phyllosilicates and organic matter compounds, then less amounts of chondritic material than those given in Table 2 would have been accreted to the Moon to account for the bulk inventory of water in the lunar interior. Until ice on asteroids is identified and characterized, it remains difficult to quantify how much ice, if any, could have contributed to the mixing of different H-reservoirs within the Moon. Comets 103P/Hartley 2 (ref. 56) and 45P/Honda-Mrkos-Pajdusakova57 contain water characterized by δD values between ∼+30 and +400‰, whereas other comets studied (for example, 8P/Tuttle and 67P/Churyumov-Gerasimenko) contain water distinctly enriched in D, with δD values >+900‰ as reviewed by Hartogh et al.56 In the modelling, the δD value for water in comet 67P/Churyumov-Gerasimenko (+2,400‰ (ref. 58)) was used to represent DEK comets, the δD value for water in Oort-cloud comets was set to ∼+930‰ (see Hartogh et al.56 for a review), and the water δD for DDK belt comets was given the same isotopic composition as comet 103P (+34‰ (ref. 56)).
The modelling for scenario 1 assumes no water in the lunar interior before lunar core-formation, and follows equation (1). We used the mid-range estimate for water content for the BSM of ∼100 p.p.m. H2O (refs 4, 9, 10) (equivalent to 3.94 × 1018 kg of H2O for a 400-km-deep LMO) and imposed an upper limit on the resultant δD values after mixing of <+100‰ to remain consistent with H-isotopic data obtained for lunar samples2,3,6,7,8. Only chondrites that were likely to have delivered significant amounts of water to the Moon without adding unrealistic amounts of chondritic material (Table 2) to the LMO were considered in the mixing calculations. Each chondrite-type (CI, CM, CO, CV) was considered in turn as being the most dominant chondrite in the impactor population (Table 3). CR-type CCs were not considered as being a significant part of the impactor population since the bulk δD value (Table 1) for such meteorites is significantly elevated with respect to the expected H-isotopic composition of the lunar mantle. In each case, the model extends from having 99% of the accreted water originating from CI-, CM-, CO- or CV-type CCs to having 99% of the water accreted from another type of chondrite or comet.
The two-component mixing models applicable to scenario 2 follow the same constraints as scenario 1 except that they assume that the Moon accreted with some water and late accretion added any remaining water needed to meet the BSM estimate of 100 p.p.m. H2O. These calculations follow equation (2) and consider an initial amount of water varying between 1% and 25% of the BSM estimate (that is, 1–25 p.p.m. for a bulk BSM water content of 100 p.p.m. H2O) characterized by a δD value of −200‰ (refs 2, 3, 6, 7, 8). The model is also adapted to consider each type of CC (CI, CO, CM or CV) as being the dominant source (Table 3).
We also considered that water initially dissolved in the molten LMO might have experienced degassing through loss of H2, in a similar way as has been proposed to explain for example the elevated H-isotopic compositions of water (present as OH) measured in apatite in some lunar basalts6. Such degassing would have resulted in any remaining water in the LMO before late addition having an elevated δD signature of ∼+800‰ (following the work of Tartèse et al.8). Therefore, we utilized equation (2) and allowed the model for scenario 2 to start with water having a δD signature of ∼+800‰ before the late addition of volatiles. All other parameters were kept the same (described above) and from this we were able to model the proportion of water that was initially in the LMO prior to late addition since the model is still constrained to a final H-isotopic composition of <+100‰ and to constrain the maximum contribution made by comets to the water inventory of the Moon assuming an initial H2O content of 1 p.p.m. prior to late addition (Table 4).
Bulk nitrogen content of the Moon and mixing calculations
Using the mass of each type of CC (CI, CO, CM and CV) accreted to an initially dry Moon during LAW (scenario 1) to end up with a BSM containing 100 p.p.m. H2O (Table 2), the average N contents for the various types of carbonaceous chondrites (Table 1, data sources53,59), and assuming LMO depths of 400 or 1,000 km, we have estimated the amounts of N added to an initially N-free LMO, that is, the bulk N content of the BSM (Table 5). Note that in some cases (e.g., CV CCs) the mass of chondritic material added, in order to accrete 100 p.p.m. H2O, exceeds the mass balance constraints imposed by HSEs. We then employed the mixing models described above for H2O to calculate the acceptable proportion of N contributed by comets (with δ15N values ∼+700‰ as a lower limit, see Füri and Marty40 and references therein) for scenario 1 (Table 5) and scenario 2 (Table 5), where it was assumed that the initial N-isotopic compositions of N were −2‰ (ref. 37) and −40‰ (Füri and Marty40 and references therein) in the same way as was done for H2O (Table 3). Notice that like for H2O, CR-type chondrites are unsuitable carriers of N due to their elevated δ15N values (Table 1).
All relevant data are available from the authors on request and/or are included with the manuscript (in the form of data tables or data within figures).
How to cite this article: Barnes, J. J. et al. An asteroidal origin for water in the Moon. Nat. Commun. 7:11684 doi: 10.1038/ncomms11684 (2016).
Hauri, E. H., Weinreich, T., Saal, A. E., Rutherford, M. C. & Van Orman, J. A. High pre-eruptive water contents preserved in lunar melt inclusions. Science 333, 213–215 (2011).
Saal, A. E., Hauri, E. H., Van Orman, J. A. & Rutherford, M. J. Hydrogen isotopes in lunar volcanic glasses and melt inclusions reveal a carbonaceous chondrite heritage. Science 340, 1317–1320 (2013).
Füri, E., Deloule, E., Gurenko, A. & Marty, B. New evidence for chondritic lunar water from combined D/H and noble gas analyses of single Apollo 17 volcanic glasses. Icarus 229, 109–120 (2014).
Chen, Y. et al. Water, fluorine, and sulfur concentrations in the lunar mantle. Earth Planet. Sci. Lett. 427, 37–46 (2015).
Greenwood, J. P. et al. Hydrogen isotope ratios in lunar rocks indicate delivery of cometary water to the Moon. Nat. Geosci. 4, 79–82 (2011).
Tartèse, R. et al. 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 (2013).
Barnes, J. J. et al. The origin of water in the primitive Moon as revealed by the lunar highlands samples. Earth Planet. Sci. Lett. 390, 244–252 (2014).
Tartèse, R. et al. Apatites in lunar KREEP basalts: the missing link to understanding the H isotope systematics of the Moon. Geology 42, 363–366 (2014).
Hauri, E. H., Saal, A. E., Rutherford, M. J. & Van Orman, J. A. Water in the Moon’s interior: Truth and consequences. Earth Planet. Sci. Lett. 409, 252–264 (2015).
McCubbin, F. M. et al. Magmatic volatiles (H,C, N, F, S, Cl) in the lunar mantle, crust, and regolith: abundances, distributions, processes, and reservoirs. Am. Mineral. 100, (2015).
Alexander, C. M. O. et al. The provenances of asteroids, and their contributions to the volatile inventories of the terrestrial planets. Science 337, 721–723 (2012).
Lécuyer, C. & Gillet, P. The hydrogen isotope composition of seawater and the global water cycle. Chem. Geol. 145, 249–261 (1998).
Hallis, L. J. et al. Evidence for primordial water in Earth’s deep mantle. Science 350, 6262 (2015).
Elkins-Tanton, L. T., Burgess, S. & Yin, Q. Z. The lunar magma ocean: Reconciling the solidification process with lunar petrology and geochronology. Earth Planet. Sci. Lett. 304, 326–336 (2011).
Grove, T. L. & Krawczynski, M. J. Lunar mare volcanism: Where did the magmas come from? Elements 5, 29–34 (2009).
Morbidelli, A., Marchi, S., Bottke, W. F. & Kring, D. A. A sawtooth-like timeline for the first billion years of lunar bombardment. Earth Planet. Sci. Lett. 355-356, 144–151 (2012).
Potter, R. W. K., Kring, D. A. & Collins, G. S. Scaling of basin-sized impacts and the influence of target temperature. Geol. Soc. America. Spec. Papers 518, SPE518–06 (2015).
Wieczorek, M. A. et al. The crust of the Moon as seen by GRAIL. Science 339, 671–675 (2013).
Bottke, W. F., Walker, R. J., Day, J. M. D., Nesvorny, D. & Elkins-Tanton, L. Stochastic late accretion to Earth, the Moon, and Mars. Science 330, 1527–1530 (2010).
Joy, K. H. et al. Direct detection of projectile relics from the end of the lunar basin-forming epoch. Science 336, 1426–1429 (2012).
Marchi, S., Bottke, W. F., Kring, D. A. & Morbidelli, A. The onset of the lunar cataclysm as recorded in its ancient crater populations. Earth Planet. Sci. Lett. 325-326, 27–38 (2012).
Kring, D. A. & Cohen, B. A. Cataclysmic bombardment throughout the inner Solar System 3.9–4.0 Ga. J. Geophys. Res. 107, (2002).
Nakajima, M. & Stevenson, D. J. in 45th Lunar and Planetary Science Conference, 2770, (2014).
Day, J. M. D., Pearson, D. G. & Taylor, L. A. Highly siderophile element constraints on accretion and differentiation of the Earth-Moon system. Science 315, 217–219 (2007).
Day, J. M. D. & Walker, R. J. Highly siderophile element depletion in the Moon. Earth Planet. Sci. Lett. 423, 114–124 (2015).
Elkins-Tanton, L. T. & Grove, T. L. Water (hydrogen) in the lunar mantle: Results from petrology and magma ocean modeling. Earth Planet. Sci. Lett. 307, 173–179 (2011).
Tartèse, R. & Anand, M. Late delivery of chondritic hydrogen into the lunar mantle: Insights from mare basalts. Earth Planet. Sci. Lett. 361, 480–486 (2013).
Meisel, T., Walker, R. J. & Morgan, J. W. The osmium isotopic composition of the Earth's primitive upper mantle. Nature 383, 517–520 (1996).
Walker, R. J. et al. In search of late-stage planetary building blocks. Chem. Geol. 411, 125–142 (2015).
Puchtel, I. S., Walker, R. J., James, O. B. & Kring, D. A. Osmium isotope and highly siderophile element systematics of lunar impact melt breccias: implications for the late accretion history of the Moon and Earth. Geochim. Cosmochim. Acta 72, 3022–3042 (2008).
Liu, J., Sharp, M., Ash, R. D., Kring, D. A. & Walker, R. J. (2015) Diverse impactors in Apollo 15 and 16 impact melt rocks: Evidence from osmium isotopes and highly siderophile elements. Geochim. Cosmochim. Acta 155, 122–153.
Fischer-Gödde, M. & Becker, H. Osmium isotope and highly siderophile element constraints on ages and nature of meteoritic components in ancient lunar impact rocks. Geochim. Cosmochim. Acta 77, 135–156 (2012).
Boyce, J. W. et al. The chlorine isotope fingerprint of the lunar magma ocean. Sci. Adv. 1, e1500380 (2015).
Steele, A. et al. Graphite in an Apollo 17 impact melt breccia. Science 329, 51 (2010).
Morbidelli, A. et al. Source regions and timescales for the delivery of water to the Earth. Meteorit. Planet. Sci. 35, 1309–1320 (2000).
Fischer-Gödde, M., Burkhardt, C., Kruijer, T. S. & Kleine, T. Ru isotope heterogeneity in the solar protoplanetary disk. Geochim. Cosmochim. Acta 168, 151–171 (2015).
Mortimer, J., Verchovsky, A. B., Anand, M., Gilmour, I. & Pillinger, C. T. Simultaneous analysis of abundance and isotopic composition of nitrogen, carbon, and noble gases in lunar basalts: Insights into interior and surface processes on the Moon. Icarus 255, 3–17 (2015).
Füri, E., Barry, P. H., Taylor, L. A. & Marty, B. Indigenous nitrogen in the Moon: constraints from coupled nitrogen–noble gas analyses of mare basalts. Earth Planet. Sci. Lett. 431, 195–205 (2015).
Cartigny, P., Jendrzejewski, N., Pineau, F., Petit, E. & Javoy, M. Volatile (C, N, Ar) variability in MORB and the respective roles of mantle source heterogeneity and degassing: the case of the Southwest Indian Ridge. Earth Planet. Sci. Lett. 194, 241–257 (2001).
Füri, E. & Marty, B. Nitrogen isotope variations in the Solar System. Nat. Geosci. 8, 515–522 (2015).
Robinson, K. L. et al. in 45th Lunar and Planetary Science Conference, 1607, (2014).
Geiss, J. & Gloecker, G. Abundance of deuterium and helium in the protosolar cloud. Sp. Sci. Rev 84, 239–250 (1998).
Halliday, A. N. The origins of volatiles in the terrestrial planets. Geochim. Cosmochim. Acta 105, 146–171 (2013).
Marty, B. The origins and concentrations of water, carbon, nitrogen and noble gases on Earth. Earth Planet. Sci. Lett. 313-314, 56–66 (2012).
Sharp, Z. D., McCubbin, F. M. & Shearer, C. K. A hydrogen-based oxidation mechanism relevant to planetary formation. Earth Planet. Sci. Lett. 380, 88–97 (2013).
Usui, T., Alexander, C. M. O. D., Wang, J., Simon, J. I. & Jones, J. H. Origin of water and mantle-crust interactions on Mars inferred from hydrogen isotopes and volatile element abundances of olivine-hosted melt inclusions of primitive shergottites. Earth Planet. Sci. Lett. 357-358, 119–129 (2012).
Sarafian, A. R., Nielsen, S. G., Marschall, H. R., McCubbin, F. M. & Monteleone, B. D. Early accretion of water in the inner solar system from a carbonaceous chondrite-like source. Science 346, 623–626 (2014).
Hirschmann, M. M., Withers, A. C., Ardia, P. & Foley, N. T. Solubility of molecular hydrogen in silicate melts and consequences for volatile evolution of terrestrual planets. Earth Planet. Sci. Lett. 345-348, 38–48 (2012).
Albarede, F. et al. Asteroidal impacts and the origin of terrestrial and lunar volatiles. Icarus 222, 44–52 (2013).
Touboul, M., Puchtel, I. S. & Walker, R. J. Tungsten isotopic evidence for disproportional late accretion to the Earth and Moon. Nature 520, 530–533 (2015).
Kruijer, T. S., Kleine, T., Fischer-Gödde, M. & Sprung, P. Lunar tungsten isotopic evidence for the late veneer. Nature 520, 534–537 (2015).
Dauphas, N., Robert, F. & Marty, B. The late asteroidal and cometary bombardment of Earth as recorded in water deuterium to protium ratio. Icarus 148, 508–512 (2000).
Kerridge, J. F. Carbon, hydrogen and nitrogen in carbonaceous chondrites: abundances and isotopic compositions in bulk samples. Geochim. Cosmochim. Acta 49, 1707–1714 (1985).
Robert, F. Water and organic matter D/H ratios in the solar system: a record of an early irradiation of the nebula? Planet. Space Sci. 50, 1227–1234 (2002).
Robert, F., Javoy, M., Halbout, J., Dimon, B. & Merlivat, L. Hydrogen isotope abundances in the solar system. Part II: Meteorites with terrestrial-like ratio. Geochim. Cosmochim. Acta 51, 1807–1822 (1987).
Hartogh, P. et al. Ocean-like water in the Jupiter-family comet 103P/Hartley 2. Nature 478, 218–220 (2011).
Lis, D. C. et al. A Herschel study of D/H in water in the Jupiter-Family comet 45P/Honda-Mrkos-Pajdušáková and prospects for D/H measurements with CCAT. Astrophys. J. 774, L3 (2013).
Altwegg, K. et al. 67P/Churyumov-Gerasimenko, a Jupiter family comet with a high D/H ratio. Science 347, 3–6 (2015).
Pearson, V. K., Sephton, M. A., Franchi, I. A., Gibson, J. M. & Gilmour, I. Carbon and nitrogen in carbonaceous chondrites: Elemental abundances and stable isotopic compositions. Meteorit. Planet. Sci. 1918, 1899–1918 (2006).
Balsiger, H., Altwegg, K. & Geiss, J. D/H and 18O/16O ratio in the hydronium ion and in neutral water from in situ ion measurements in comet Halley. J. Geophys. Res. 100, 5827–5834 (1995).
Bockelée-Morvan, D. et al. Deuterated water in comet C/1996 B2 (Hyakutake) and its implication for the origin of comets. Icarus 133, 147–162 (1998).
Meier, R. et al. A determination of the HDO/H2O ratio in comet C/1995 O1 (Hale-Bopp). Science 279, 842–844 (1998).
Hutsémekers, D., Manfroid, J., Jehin, E., Zucconi, J.-M. & Arpigny, C. The OH/18OH and OD/OH isotope ratios in comet C/2002 T7 (LINEAR). Astron. Astrophys. 490, L31–L43 (2008).
Bockelée-Morvan, D. et al. Herschel measurements of the D/H and 16O/18O ratios in water in the Oort-cloud comet C/2009 P1 (Garradd). Astron. Astrophys. 544, L15 (2012).
Villanueva, G. et al. A sensitive search for deuterated water in comet 8P/Tuttle. Astrophys. J. 690, L5–L9 (2009).
Biver, N. et al. Radio wavelength molecular observations of comets C/1999 T1 (Mc Naught-Hartley), C/2001 A2 (LINEAR) and 153/P Ikeya-Zhang. Astron. Astrophys. 449, 1255–1270 (2006).
Aléon, J., Engrand, C., Robert, F. & Chaussidon, M. Clues to the origin of interplanetary dust particles from the isotopic study of their hydrogen-bearing phases. Geochim. Cosmochim. Acta 65, 4399–4412 (2001).
Floss, C. et al. Identification of isotopically primitive interplanetary dust particles: a NanoSIMS isotopic imaging study. Geochim. Cosmochim. Acta 70, 2371–2399 (2006).
Waite, J. H. et al. Liquid water on Enceladus from observations of ammonia and 40Ar in the plume. Nature 460, 487–490 (2009).
Lellouch, E. et al. The deuterium abundance in Jupiter and Saturn from ISO-SWS observations. Astron. Astrophys. 670, 610–622 (2001).
Feuchtgruber, H. et al. Detection of HD in the atmospheres of Uranus and Neptune: a new determination of the D/H ratio. Astron. Astrophys. 341, L17–L21 (1999).
Lellouch, E. et al. First results of Herschel-PACS observations of Neptune. Astron. Astrophys. 518, L152 (2010).
Greenwood, J. P., Itoh, S., Sakamoto, N., Vicenzi, E. P. & Yurimoto, H. Hydrogen isotope evidence for loss of water from Mars through time. Geophys. Res. Lett. 35, 1–5 (2008).
Hu, S. et al. NanoSIMS analyses of apatite and melt inclusions in the GRV 020090 Martian meteorite: hydrogen isotope evidence for recent past underground hydrothermal activity on Mars. Geochim. Cosmochim. Acta 140, 321–333 (2014).
Boctor, N. Z., Alexander, C. M. O. D., Wang, J. & Hauri, E. The sources of water in Martian meteorites: Clues from hydrogen isotopes. Geochim. Cosmochim. Acta 67, 3971–3989 (2003).
Bjoraker, G. L., Mumma, M. J. & Larson, H. P. Isotopic Abundance Ratios for Hydrogen and Oxygen in the Martian Atmosphere. in 21st Annual DPS Meeting 991 (1989).
Barnes, J. J. et al. Accurate and precise measurements of the D/H ratio and hydroxyl content in lunar apatites using NanoSIMS. Chem. Geol. 337-338, 48–55 (2013).
McKeegan, K. D. & Leshin, L. A. Stable isotope variations in extraterrestrial materials. Rev. Mineral. Geochem. 43, 279–318 (2001).
STFC is thanked for a PhD studentship to J.J.B. and for research grants to M.A. (grant number ST/I001298/1 and ST/L000776/1). This research was supported in part by NASA Solar System Exploration Research Virtual Institute contract NNA14AB07A (David A. Kring, PI). We thank the reviewers for their insightful comments that have certainly enhanced the quality of this paper. Lunar and Planetary Institute contribution 1909.
The authors declare no competing financial interests.
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Barnes, J., Kring, D., Tartèse, R. et al. An asteroidal origin for water in the Moon. Nat Commun 7, 11684 (2016). https://doi.org/10.1038/ncomms11684
This article is cited by
A dry lunar mantle reservoir for young mare basalts of Chang’e-5
The origin and fate of volatile elements on Earth revisited in light of noble gas data obtained from comet 67P/Churyumov-Gerasimenko
Scientific Reports (2020)
An extremely heavy chlorine reservoir in the Moon: Insights from the apatite in lunar meteorites
Scientific Reports (2019)
Constraining the Evolutionary History of the Moon and the Inner Solar System: A Case for New Returned Lunar Samples
Space Science Reviews (2019)
Water on Earth
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.