Evidence of water on the lunar surface from Chang'E-5 in-situ spectra and returned samples

A growing body of evidence indicates water (H2O/OH) exists on the lunar surface, but its spatial distribution, temporal variability, and sources remain debated. Earlier views of an anhydrous Moon were overturned by neutron and spectroscopic observations indicating hydrogen and hydration signatures, particularly near the poles and in permanently shadowed regions. Orbital instruments (M3, Deep Impact/HRIIR, Cassini/VIMS, OSIRIS-REx/OVIRS) and LCROSS plume analysis detected water/hydroxyl, while Apollo samples revealed water in volcanic glasses, melt inclusions, apatite, and anorthosite. Proposed sources include indigenous magmatic water, solar wind implantation, and meteorite/cometary delivery, with recent suggestions of contributions from Earth wind and impact-liberated exospheric dust. Many remote sensing studies favor solar wind implantation with possible diurnal variability and latitudinal trends, whereas other analyses (e.g., Bandfield et al.) argue for widespread, stable hydration without significant diurnal migration, potentially stored in lattice defects and not requiring indigenous water. The Chang'E-5 (CE-5) mission, which landed in a very young mare unit, provides a unique opportunity to address these questions via in-situ spectroscopy and laboratory analysis of returned samples to constrain the presence, variability, and sources of lunar surface water.

- Historical perspective: Early Apollo-era analyses suggested the Moon was extremely dry. Subsequent neutron spectroscopy indicated near-surface hydrogen globally, with polar enhancement. - Orbital spectroscopy: M3 (Chandrayaan-1), Deep Impact/EPOXI HRIIR, Cassini VIMS, and OSIRIS-REx OVIRS detected 3-µm absorption features attributed to OH/H2O. LCROSS confirmed water in permanently shadowed regions. - Sample evidence: Apollo materials showed water in volcanic glass beads, melt inclusions, apatite, and plagioclase, indicating indigenous reservoirs. - Proposed sources: Indigenous (magmatic) water; solar wind H+ implantation forming OH/H2O via interactions with surface oxygen and defects; cometary/meteoritic delivery; possible Earth wind contribution; impact-liberated exospheric water. - Ongoing debate: Remote sensing often supports solar wind-driven hydration with latitudinal dependence and possible diurnal variability. Detection of magmatic water at Bullialdus crater and in pyroclastic deposits supports indigenous contributions. Bandfield et al. argued hydration is widespread and stable across latitudes, local times, and terrains without strong diurnal migration, potentially not requiring indigenous sources. - Gaps: Orbital datasets require thermal/photometric corrections and can be ambiguous regarding sources; indigenous signals may be spatially heterogeneous and below detection thresholds in many areas. In-situ and returned sample analyses can help resolve source contributions.

- In-situ spectroscopy (LMS): The CE-5 lander’s Lunar Mineralogical Spectrometer (LMS) acquired hyperspectral data from 0.48–3.2 µm, including the 3-µm OH/H2O region. During scoop operations, 11 Full-Bands Observations (FBO) were collected on rock and soil targets; 8 were used after quality screening. Data processing included radiometric calibration (including in-situ calibration using gold and aluminum panels), reflectance (RADF) calculation, detector cross-calibration to produce continuous spectra, and geometric calibration. - Thermal correction: Because observations occurred near lunar local noon with high temperatures (≈335–360 K), thermal emission beyond 2 µm was removed using three models: Clark et al. (2011), Li & Milliken (2016), and a Planck fit per Groussin et al. (2007). Temperatures retrieved by the three models were consistent with Diviner-derived temperature–time trends. Analyses focused on ≤3100 nm due to reduced instrument response >3100 nm. - Photometric correction: Multispectral datasets spanning phase angles 21–94° were corrected to standard geometry (i=30°, e=0°, g=30°) using the Lommel-Seeliger model with polynomial phase functions derived from 146 quality-controlled targets. FBO spectra were corrected using interpolated coefficients. - OH feature analysis and content estimation: Continuum removal was applied; the 2.85 µm absorption (hydroxyl-dominated) was measured. Single Scattering Albedo was derived via Hapke modeling; ESPAT at 2850 nm was computed and converted to water content using the Li & Milliken (2017) empirical relation from stepwise heating experiments. The 2.95 and 3.05 µm bands were noted but not quantified due to weakness and complexity. - Laboratory analyses of returned samples: Three lunar soil powders and three polished rock-fragment sections were studied. Instruments included LMS Engineering Qualification Module (LMS-EQM) for lab spectra (900–3100 nm for analysis), Scanning Electron Microscopy (SEM) with Energy-Dispersive Spectroscopy (EDS) to identify phases, X-ray Diffraction (XRD) with Rietveld refinement to quantify mineral abundances, Raman spectroscopy to characterize apatite and assess F–Cl–OH variations in PO4 vibrations, and Electron Probe Micro-Analysis (EPMA) with stoichiometric calculations to estimate OH, F, and Cl in apatite (based on 13-anion stoichiometry). Sample prep ensured homogeneity and minimized preferred orientation; XRD patterns were measured and summed over 20 runs per sample. - Comparative context: In-situ LMS soil spectra were compared against M3 orbital data for similar temperatures, latitudes, and compositions; Diviner data provided temperature context; ARTEMIS P1 ion energy flux data established that CE-5 observations occurred while the Moon was inside Earth’s magnetosphere with reduced solar wind flux. Agglutinitic glass abundance in CE-5 soils was compared to Apollo 11 (10084) to infer solar wind contribution scaling.

- Detection of hydration: LMS in-situ spectra show clear hydroxyl-related absorption near 2.85 µm across multiple targets; laboratory spectra of returned soils also exhibit a 2.85 µm feature. - Quantified hydroxyl in soils and rock: Using Clark-model thermal correction and ESPAT calibration, estimated hydroxyl contents (ppm) for eight FBOs were: 0009: 11; 0010: 26; 0011 (rock): 152; 0012: 110; 0014: 15; 0015: 2; 0016: 7; 0017: 82. Soil spectra (excluding 0012) are very low, with a mean soil hydroxyl content of ~28.5 ppm, representing a weak hydration endmember at low latitudes. - Rock vs soil: The rock target shows deeper 2.85 µm absorption and higher apparent hydroxyl, but differences likely reflect surface optical properties (e.g., grain size, roughness, maturity) rather than composition; 1- and 2-µm band shapes/centers indicate similar mafic compositions for rock and soils. - Apatite presence and OH content: CE-5 soils contain apatite at 0.1 ± 0.1 wt%, 0.7 ± 0.1 wt%, and 1.4 ± 0.1 wt% across three samples. Raman peaks (≈959.3–961.7 cm−1) and EPMA stoichiometry show apatite is a Ca5(PO4)3(F,Cl,OH) solid solution with OH fraction 0–0.38 in analyzed grains. Estimated sample hydroxyl content attributable to apatite ranges from 0 to 179 ± 13 ppm. - Solar wind contribution assessed by agglutinitic glass: CE-5 soils have low agglutinitic glass (~16 wt%), about one-third of Apollo 11 (10084; ~59 wt%), which had ~70 ppm OH/H2O attributed to solar wind. Assuming linear scaling, CE-5 solar wind-generated hydroxyl would be ~1/3 of Apollo 11, consistent with the mean LMS soil estimate (~28.5 ppm), indicating a weak solar wind contribution during observations. - Environmental context for weak hydration: Observations took place near local noon at high temperatures (≈335–360 K), likely causing loss of molecular water; the Moon was inside Earth’s magnetosphere, reducing solar wind flux; the site is a young mare basalt unit, which globally exhibits weaker hydration than highlands. LMS soil 2.85 µm band depths (0–4%) are weaker than M3 data (3–7.5%) at similar temperatures and latitudes, placing CE-5 in the weak hydration regime. - Source inference: Excess hydroxyl seen in some soil spectra (e.g., 0012) beyond expected solar wind contribution likely reflects indigenous hydroxyl, with hydroxyapatite identified as a plausible specific source.

The combined in-situ and laboratory evidence demonstrates the presence of hydroxyl at the CE-5 landing site, but at relatively low abundances compared to many global contexts. The weak hydration signature is consistent with environmental and geological factors: high surface temperatures during acquisition, diminished solar wind flux inside Earth’s magnetosphere, and mare basaltic composition. Quantitative estimates align with remote sensing predictions for similar conditions. The detection and characterization of apatite, including its OH component and measured abundances, provide a credible indigenous source for hydroxyl, accounting for cases where hydroxyl exceeds levels expected from solar wind implantation alone. Comparison of rock and soil spectra suggests that apparent stronger rock hydration is more likely due to optical effects than higher intrinsic hydroxyl, supporting the use of soil spectra for broader comparisons. Overall, these results help reconcile competing views on lunar hydration by showing that both solar wind and indigenous sources contribute, with their relative importance varying with environmental conditions and lithology.

Chang’E-5 provided the first in-situ lunar surface spectra near 3 µm from a young mare region alongside returned samples for laboratory confirmation. The study quantified low hydroxyl contents in surface soils (mean ~28.5 ppm) and identified apatite containing OH as present and variable in abundance, capable of contributing up to ~179 ± 13 ppm hydroxyl depending on sample. Low agglutinitic glass contents and observation under Earth’s magnetospheric shielding indicate a weak solar wind contribution during the measurement period. Together, the in-situ and lab results indicate that hydroxyapatite likely supplies an indigenous component to the observed hydration at CE-5, helping to explain excess hydroxyl beyond solar wind implantation. Future work should pursue temporal monitoring across local times and solar wind conditions, broader sampling of terrains (including highlands and pyroclastics), improved instrument response beyond 3.1 µm, and integrated isotopic/volatile analyses to further disentangle sources and processes controlling lunar surface hydration.

- Observation conditions: LMS operated near local noon at high temperatures (~335–360 K), potentially depleting molecular water and biasing results toward weak hydration. - Instrumental response: Reduced LMS spectral response beyond 3100 nm precluded analysis >3.1 µm; downturns after thermal correction necessitated restricting interpretation to ≤3100 nm. - Spectral complexity: The 2.95 and 3.05 µm absorptions were too weak/complex to quantify, limiting constraint on molecular water vs hydroxyl partitioning. - Data screening: Three early FBO spectra were excluded due to scattered light artifacts (robotic arm), reducing the in-situ dataset used. - Thermal/photometric modeling assumptions: Emissivity approximations (e.g., ε≈1 in Planck fitting) and empirical thermal corrections introduce uncertainties (~few K). Hapke parameters and ESPAT-to-H2O calibration carry model-dependent errors. - Source attribution assumptions: Linear scaling between agglutinitic glass content and solar wind-generated hydroxyl is assumed; actual relationships may be more complex. Optical property differences complicate rock–soil hydration comparisons. - Environmental context: Measurements occurred while the Moon was within Earth’s magnetosphere, potentially not representative of typical solar wind conditions. - Sample representation: Returned materials are from a young mare basalt unit; results may not generalize to other lithologies (e.g., highlands). Limited number of apatite grains and sections analyzed constrain statistical robustness. - Laboratory artifacts: LMS-EQM lab spectra can be influenced by atmospheric water vapor along the optical path (∼1900 and 2600–2700 nm features), though not affecting the 2.85 µm conclusions.