As Fig. 1 shows, no source is a perfect fit for the LCROSS measurements. Volcanic sources and chondrites provide the right amount of sulfur, but do not provide sufficient hydrogen and nitrogen. Volcanic sources are also deficient in oxygen. Comets provide sufficient hydrogen, carbon, and nitrogen, but are depleted in sulfur – even when considering the most extreme value. Solar wind only contributes hydrogen and oxygen (see supplementary materials). We developed a model to determine if a mixture of sources can match the LCROSS observations, and found that no combination was able to match all four elemental ratios within the uncertainties of the LCROSS measurements – even when taking into account the uncertainties for the sources (see supplemental materials). The main limitation is fitting both the C/S and the N/C ratios observed by LCROSS. The two sources with sufficient sulfur to match C/S, volcanoes and chondrites, are too depleted in nitrogen and hydrogen for any cometary contribution to provide agreement with N/C and still match C/S. This is the case even using the maximum N/C and the minimum C/S for comets. The best fit is provided by 100% comets, which agrees with all ratios except for C/S.
To improve our constraints on the source, or mixture of sources, we consider processes that could fractionate elemental ratios between delivery of the source volatiles to the lunar surface and observation in the LCROSS plume, including volcanic atmospheric processes, impact processes, clathrate formation, and cycles of sublimation and recondensation. Because these processes are complex and difficult to accurately quantify, we determine whether the LCROSS observations represent upper or lower limits for the elemental ratios and summarize the results in Table 1.
4.1 Volcanic Atmosphere Fractionation
Volcanic sulfur is thought to be released as S2, which could rapidly be lost to the surface as solid elemental sulfur or aerosols before reaching a cold trap (26). This would result in a higher C/S ratio in the PSR compared to the source, so the observed C/S is an upper limit for volcanic C/S. This creates a challenge for explaining the LCROSS C/S as volcanic in origin, because volcanic C/S would need to be much lower than C/S in the LCROSS plume to provide sufficient sulfur to explain the observations.
The relative abundances of elements in volcanic gas can also be changed by the escape of molecules from the top of the atmosphere. Unfortunately, loss rates depend on a wide range of complex parameters that are not well constrained (27), making it difficult to quantify how much elemental ratios can fractionate as a result of escape. However, we can estimate upper and lower limits for LCROSS measurements compared to the sources based on the relative masses of the dominant species for each element. Escape from a volcanic atmosphere would be dominated by H and H2 (26, 27) that either originated in the volcanic gas as H2, or was produced by dissociation of water molecules. This would increase the C/H of the volatiles in the PSR, making the observations an upper limit for the source ratio. Atomic oxygen and OH produced by water dissociation could also be lost, making O/C in the PSR a lower limit compared to the source. Any nitrogen present would be in the form of either N2 or NH3, which are either the same mass as or lighter than volcanic carbon-bearing molecules CO and CO2. This means that the N/C in the PSR is a lower limit for N/C in a volcanic source when considering atmospheric escape. Because volcanic N/C is drastically lower than the LCROSS observations, escape does not provide a mechanism allowing for volcanic gas to be the source of nitrogen in the Cabeus PSR.
Although escape of hydrogen and oxygen leads to limits that provide worse agreement between a volcanic source and the LCROSS observations, water produced by solar wind surface chemistry would decrease C/H and increase O/C over time by adding water to the PSR (28), cancelling out escape fractionation. These ratios would allow for a combination of volcanic and solar wind sources. However, measured N/C and C/S ratios disagree with volcanic source composition, even accounting for processes that change elemental ratios in a volcanically-produced atmosphere, conclusively demonstrating that the volatiles sampled by LCROSS are not from a volcanic source.
4.2 Fractionation of Impact Material
Next, we consider fractionation of volatiles delivered by impacts of comets, asteroids, and micrometeoroids. The elemental ratios can be fractionated by impact loss and by escape during transport to cold traps. The total percentage of volatiles retained after impact depends on the impact velocity and angle (29). Volatiles lost to space escape rapidly as part of the outward flow of the impact plume. Fractionation is similar to hydrodynamic escape, with preferential loss of lighter species. However, light species flow outward rapidly enough to drag heavier species with them (e.g., 30). Additional loss to space could occur by escape during subsequent transport to cold traps over several Earth days (31). Fractionation can be estimated in the same way as with the volcanic atmosphere, assuming that lighter species are removed at a faster rate than heavier species. Hydrogen would primarily be in light molecules like H, H2, and water making the C/H in the PSR an upper limit compared to C/H of the source. Loss of oxygen and OH would make O/C in the PSR a lower limit compared to the source. According to simulations of impact chemistry of comets (32) and chondrites (33), nitrogen in an impact plume would primarily be in the form of N2 with some NH3 present, while carbon and sulfur are found in heavier molecules like CO, CO2, H2S, SO2 and OCS. As with the volcanically produced atmosphere, N/C in the LCROSS observations is a lower limit compared to the source. We also note that LCROSS and LAMP did not have the ability to detect N2, which is expected to be produced in impact plumes. The N/C in the LCROSS plume may have been higher than observed, arguing further for that the observation is a lower limit compared to the source. The masses for carbon-bearing species are generally lighter than sulfur-bearing species, suggesting that C/S in the LCROSS observations is a lower limit compared to the source. We applied our model again using these constraints (see Table 1) and found that only cometary ices, with some contribution from solar wind-produced water, can explain all four elemental ratios.
4.3 Clathrate Formation
During the cooling of an impact plume, clathrates can form with entrapped mixtures different from the coexisting gases. In this case, the entrapped mixture will be enriched in H2S and SO2, and depleted in CO compared to the initial mixture because H2S and SO2 have a higher propensity for trapping compared to CO at low pressure conditions (16). If insufficient water is available to trap all of the CO, H2S and SO2 present in the gas, C/S in the clathrates is lower than in the source. Ammonia is not trapped in clathrates, but would form ammonia hydrates at temperatures between 80 and 100 K, or condense as pure ammonia frost at temperatures below 80 K. If not all of the CO is trapped, but all of the NH3 ends up in the PSR, the N/C observed by LCROSS is an upper limit compared to the source. In this case either comets or chondrites could agree with the C/S and N/C. However, based on the water to CO ratio in clathrates the O/C ratio for volatiles trapped in clathrates must be lower than 6.75 if not enough water was available for all of the CO to be trapped (15). The LCROSS O/C disagrees with this limit, so additional water must be supplied by the solar wind. We again modeled a combination of sources assuming that the LCROSS C/S and O/C are lower limits based on clathrate formation processes and escape, that C/H is an upper limit based on escape, and ignoring N/C because of the competing influences of clathrate formation and escape. We found that a combination of cometary and solar wind sources fits these constraints, but that the modeled O/C is too high to support clathrate formation even accounting for a solar wind water source. Therefore, it is unlikely that ices that formed as clathrates can explain the LCROSS observations.
4.4 Sublimation and Recondensation
Finally, we consider how a cycle of sublimation and recondensation of volatiles could fractionate the elemental ratios. As volatiles are transported to the PSR, they could condense to the surface at night and sublimate during the day. A similar cycle could also take place within a PSR if diurnal temperatures vary enough to cause sublimation of some species depending on their volatility. The temperatures in the Cabeus PSR are very low and not likely to cause diurnal variations, but volatiles in this PSR could have been influenced by these processes before being trapped. Additionally, recondensation could occur within the Cabeus PSR when volatiles are released through impact gardening. This cycle would increase the abundance of water relative to other species observed in the LCROSS plume that have lower volatility temperatures (5). It would also increase the abundance of NH3, H2S, and SO2 relative to CO and N2. This means that C/S, N/C, and C/H in the PSR are lower limits compared to the source, while O/C is an upper limit. We modeled the source contributions with these constraints and found that a combination of comets, chondrites, and solar wind was possible. To narrow the possibilities further we add four more constraints shown in Table 1. Including these constraints limits the possible combination of source volatiles to 30–45% cometary and 55–70% chondrites with no solar wind contributions. Finally, we consider the combination of loss to space and a cycle of sublimation and recondensation. In this case the only reliable constraints are C/S, N/C, N/S and O/H. These constraints allow for any combination of comets and chondrites with no water provided by solar wind.
Table 1
Model constraints and results for determining the possible sources for the LCROSS plume based on understanding of fractionation processes.
| No fractionation | Volcanic atm. processes | Impact and Escape | Clathrate formation | Sublimation and recondensation | Escape, subl. & recondensation |
C/S | Within errors of LCROSS obs. & sources | Upper limit | Lower limit | Lower limit | Lower limit | Lower limit |
N/C | Lower limit | Lower limit | Un-constrained | Lower limit | Lower limit |
O/C | Lower limit | Lower limit | < 6.75 | Upper limit | Unconstrained |
C/H | Upper limit | Upper limit | > 0.07 | Lower limit | Unconstrained |
N/S | n/a | n/a | Lower limit | n/a | Lower limit | Lower limit |
S/O | Upper limit | Lower limit | Unconstrained |
S/H | Upper limit | Lower limit | Unconstrained |
O/H | Upper limit | Lower limit | Constrained by solar wind input |
Results | No good fit | No good fit | Comets & Solar Wind | No good fit | 30–45% Comets 55–70% Chondrites | Comets & Chondrites |