Impact-processed nitrogen-bearing organics in Chang’e-5 and Chang’e-6 lunar regolith

Abstract

The Moon has preserved a unique record of organic matter delivered and reworked by asteroid and comet impacts. Here, we report diverse organic phases (particle-like, adhering-like, and inclusion-like) on the surfaces of lunar regolith grains returned by the Chang’e-5 and Chang’e-6 missions. They are predominantly amorphous carbon–like, containing N- and O-bearing functionalities and amide (─CONH─) linkages. The lunar organics show δD, δ¹³C, and δ¹⁵N values more negative than those of insoluble organic matter reported in carbonaceous chondrites and asteroids, consistent with impact-induced evaporation-condensation and surface reworking. The presence of solar wind implantation signatures in the organics supports long-term exposure on the lunar surface. Together, these findings suggest that the impacts both delivered and chemically processed organic matter on the lunar surface, generating N- and O-bearing functionalities.

Exogenous delivery, impacts, and solar wind implantation shape the evolution of lunar surface organic matter.

INTRODUCTION

Asteroids and comets continuously delivered organic matter and C-, N-, O-, P-, and S-based chemistry to the surfaces of habitable terrestrial planets in the early Solar System (1–4). On Earth, however, intense geological and biological activity has obscured the early records, complicating efforts to identify the sources and molecular forms of early carbon. The Moon, cogenetic with Earth but largely geologically inactive (5), preserves ancient exogenous organics as a primitive “time capsule.” This archive not only records delivery pathways (6, 7) but also enables evaluation of impact-driven reworking processes, making it a critical context for assessing the fate of extraterrestrial organics relevant to early Earth (4). Earlier Apollo-era analyses revealed bulk carbon and nitrogen contents in lunar regolith, but interpretations were clouded by terrestrial contamination concerns. Analyses of Apollo samples reported average total carbon of 124 ± 45 parts per million (ppm) and nitrogen of 81 ± 37 ppm in lunar regolith (8). These volatiles derive from a mixture of sources: solar wind implantation (9), exogenous input from micrometeoroids and comets (10), and endogenous magmatic contributions (11). Graphite has been reported in an Apollo 17 impact-melt breccia (12). Complex organic coatings intimately associated with Apollo 17 glass beads from the Shorty crater (74220) were identified; the coatings are kerogen-like, with ~1.2- to 1.6-nm aromatic domains composed of mixed amorphous and nanocrystalline structures (13).

Recently, the Chang’e-5 (CE5) mission collected regolith near Mons Rümker in the northwestern Oceanus Procellarum on the lunar nearside, sampling young basalts emplaced ~2.0 billion years ago (14); Chang’e-6 (CE6) achieved the first far-side return from the Apollo crater region within the South Pole–Aitken basin (15). Together, these newly returned samples have experienced only a short storage period, thereby minimizing the potential for terrestrial contamination during curation (16). The benzene polycarboxylic acid method (17) has been applied to CE5 lunar regolith and quantified benzene polycarboxylic acid–derived polycyclic aromatic carbon at 5.0 to 9.2 ppm. These studies indicate that meteoroid impacts continuously deliver carbon to the lunar surface and, through impact heating and recondensation, supply volatiles and organic precursors (13, 17, 18). However, these studies have primarily reported graphitized carbon or aromatic-rich components. The chemical speciation and bonding environments of nitrogen-bearing organics in lunar regolith remain poorly constrained and are not systematically documented—in marked contrast to the N-rich organics as amino acids and nucleobases reported in Ryugu (1, 19) and Bennu (2) asteroids and chondrites (20, 21). Because N-bearing organics (e.g., amines, amides, nitriles, and N-heteroaromatics) are central components of life’s molecular building blocks (22), identifying and characterizing them in lunar regolith are important for evaluating the fate of extraterrestrial organics during impact processing (4) and for assessing potential links to prebiotic chemistry.

In this study, organic matter was systematically investigated on the surfaces of CE5 and CE6 lunar regolith using a correlative multitechnique approach (fig. S1). Organic morphologies, host-phase associations, elemental compositions, organic functional groups, and stable-isotope signatures were measured. Our aim is to resolve how exogenous delivery and surface processes—impact, thermal alteration, and irradiation—jointly shape lunar organics. The occurrence and preservation potential of N-bearing organics under lunar surface conditions are also assessed, thereby constraining the evolutionary pathways of exogenous N-bearing organics relevant to early Earth and other planets.

RESULTS

Occurrence and elemental composition of organics

Scanning electron microscopy with energy-dispersive x-ray spectroscopy (SEM-EDS) mapping of CE5 and CE6 regolith grain surfaces reveals C- and N-rich hotspots (Figs. 1 and 2 and figs. S2 to S4). Twelve hotspots were identified in this study (table S1 and fig. S2). On the basis of morphology, surface organics are classified into three types: particle-like, adhering-like, and inclusion-like. Particle-like organics are observed in both CE5 and CE6 and range from ~0.5 to 20 μm in size. These organics commonly coaccumulate with inorganic debris on host grain surfaces, and fine mineral grains may overlie the organic particles (Fig. 1A). Adhering-like organics—more frequently observed in CE5—occur as irregular, often elongated films directly attached to mineral surfaces (Fig. 2, A to E). Inclusion-like organics appear as domains embedded within some regolith grains (Fig. 2, H to K), rather than as material resting on the grain surface. This interpretation is supported by the penetration depths of secondary electrons (SEs) and characteristic x-rays, indicating that these inclusion-like organics are not confined to the grain surfaces but instead extend to depths of tens to hundreds of nanometers.

Fig. 1. Particle-like organic matter on a lunar regolith grain (CE6-P2_r1).

(A) Secondary-electron (SE) image overlaid with the EDS C map, highlighting a carbon-rich, particle-like domain on the grain surface. (B and C) EDS map of C and N. The color scale indicates relative intensity. (D) Fluorescence (AutoFL channel) image of the same particle. (E) O-PTIR chemical images acquired at selected vibrational bands (2922 and 2852 cm⁻¹, CH₂ stretching; 1655 cm⁻¹, amide I, C═O; 1548 cm⁻¹, amide II, N─H). (F) Representative O-PTIR spectra collected from the organic particle (colored traces) and from an adjacent mineral surface lacking visible organic matter (gray trace) used as a local control. All scale bars, 20 μm.

Fig. 2. Adhering-like and inclusion-like organic matter on lunar regolith grains.

(A) SE image overlaid with the EDS C map, showing an adhering-like organic domain on the grain surface (CE5-P6). (B and C) EDS map of C and N for the same adhering-like domain. The color scale indicates relative intensity. (D) Fluorescence (AutoFL) image of the same particle. (E) O-PTIR chemical images of the adhering-like organic matter acquired at selected vibrational bands (2925 and 2853 cm⁻¹, CH₂ stretching; 1464 cm⁻¹, CH deformation). (F) Raman spectra from the adhering-like domain. (G) O-PTIR spectra from the adhering-like domain (colored traces) and from a neighboring augite surface used as a local mineral background (black trace). The shaded region marks the augite Si─O lattice absorptions. (H) SE image overlaid with the EDS C map for an inclusion-like organic domain (CE5-P7). (I and J) EDS map of C and N. (K) Fluorescence image of the same inclusion-like domain. (L) O-PTIR spectra collected on the inclusion-like domain (colored traces) and on an adjacent mineral surface (black trace) as a local background. Because of steep topography at the particle edge, O-PTIR mapping was not acquired for this target. All scale bars, 10 μm.

Two focused ion beam (FIB) sections (Fig. 3) were prepared from adhering-like organic matter in CE5-P7 and particle-like organic matter in CE5-P4. Bright-field STEM (BF-STEM) imaging shows that both types reside on the outermost surfaces of lunar minerals (Fig. 3). Particle-like organics are not uniformly in direct contact with the host but are agglutinated with other mineral fragments on the host surface (Fig. 3A), with a carbon-rich interfacial layer commonly present between dispersed organic particles and the host (Fig. 3B). Adhering-like organics directly cover the host surface (Fig. 3D). In both particle-like and adhering-like occurrences, fine mineral inclusions are common within the organic matrices, including pyroxene, ilmenite, and olivine (Fig. 3, A and D).

Fig. 3. STEM-EDS-EELS characterization of particle-like (CE5-P4) and adhering-like (CE5-P6) organic matter on lunar regolith grains.

(A) BF-STEM showing a particle-like organic domain at the grain surface (CE5-P4). (B and C) STEM-EDS maps for C and N. The carbon-rich boundary is marked. (D) BF-STEM image of an adhering-like organic coating on an augite surface (CE5-P6). (E and F) STEM-EDS maps for C and N. (G) C K-edge EELS spectra collected from the positions indicated in (B) and (E). (H) N K-edge EELS spectra from the same positions.

SEM-EDS data (data S1) indicate that the investigated organics are dominated by C, N, and O, which together account for >95 at % of measured atoms. Minor amounts of Na, Mg, Al, Si, S, Fe, Zn, Ca, and Cl are present (fig. S5). Given the ~0.6-μm x-ray escape depth under our analytical conditions, contributions from the underlying mineral substrate are expected to be limited. Mg, Al, Si, Fe, and Ca are therefore more plausibly attributed to fine mineral inclusions entrained within the organic matter (Fig. 3). Among moderately volatile elements (11, 23), Na and S are commonly detected, whereas Zn and Cl occur only in a subset of organic particles. The nitrogen atomic fraction is ~5 to 10 at %, corresponding to N/C ≳ 0.05; this N/C exceeds typical values reported for insoluble organic matter (IOM) in carbonaceous chondrites (Fig. 4E) (24).

Fig. 4. NanoSIMS depth profiles and isotopic and elemental compositions of lunar organic matter.

(A) H/C atomic ratios versus sputtering frame (depth) for adhering-like organics CE5-P6 and CE5-P8 and inclusion-like organics CE5-P7_r1 and CE5-P7_r2. (B) δD depth profiles for adhering-like organics and inclusion-like organics. Depth profiles were extracted from the ROIs defined in fig. S7. A rough estimate based on a typical sputter rate suggests that each frame may correspond to a few nanometers. (C) Comparison of δD for CE5 organic matter with literature values for IOM in meteorites (24) and for Ryugu IOM (15). (D) δ¹⁵N versus δ¹³C for CE5 organic matter alongside fields compiled for meteoritic IOM (24), interplanetary dust particles (40), comet 81P/Wild 2 (38), and returned samples from Ryugu (19, 29) and Bennu (2). (E) Atomic N/C versus O/C ratios of CE5 and CE6 organic matter measured by SEM-EDS (red symbols) and STEM-EDS (blue symbols) compared with the reported field for meteoritic IOM (purple shading) (24). All error bars represent 1σ uncertainties.

Structure and functional groups of organics

On CE5 and CE6 regolith grains, all three morphological types of organics—particle-like, adhering-like, and inclusion-like—exhibit pronounced fluorescence (Figs. 1 and 2), consistent with the presence of π-conjugated chromophores. Raman spectra from several particle-like occurrences (CE5-P1, CE6-P2_r2, and CE5-P3) show the characteristic D and G bands (figs. S3 and S4). The overall signatures are similar to amorphous carbon in IOM from aqueously altered carbonaceous chondrites (fig. S6) (25). However, the observed G band shows a lower peak position and a broader width than those of CM (Mighei-type) and CI (Ivuna-type) carbonaceous chondrites and a broader peak width, indicating higher disorder and smaller sp² cluster sizes (26–28). By contrast, adhering-like organics lack clear D and G bands in Raman but display strong aliphatic C─H stretching in the region of 2800 to 3000 cm⁻¹ and a band near 1630 cm⁻¹ attributable primarily to alkenyl or weakly aromatic C═C stretching (Fig. 2F). Raman spectra of inclusion-like organics could not be obtained because of strong fluorescence.

Optical photothermal infrared (O-PTIR) spectra reveal ─CH₂ and ─CH₃ stretching in all three types of organic matter, with ─CH₂ dominating. Although assignments in the fingerprint region are not unique, several features are diagnostic (table S2). In the particle-like sample CE6-P2 (Fig. 1), a pair of bands at ~1655 and ~1548 cm⁻¹ falls within the typical ranges of amide I [ν(C═O)] and amide II [δ(N─H) + ν(C─N)] (29, 30). This is consistent with the presence of nitrogen detected by EDS. The O-PTIR spectra of adhering-like organics (CE5-P6) and inclusion-like organics (CE5-P7) are similar (Fig. 2, G and L). Absorption near 1710 cm⁻¹ matches neutral carbonyl ν(C═O), potentially arising from ketones/aldehydes, partially conjugated esters, or lactones. Bands at 1670 to 1662 cm⁻¹ can be ascribed to amide I or α,β-unsaturated carbonyls. A broad feature at 1620 to 1610 cm⁻¹ is compatible with the asymmetric carboxylate stretch ν_asym(COO⁻), with possible contributions from aromatic C═C and H─O─H bending; the feature at ~1410 to 1380 cm⁻¹ may correspond to ν_sym(COO⁻), noting potential overlap with the ─CH₃ deformation (~1375 cm⁻¹).

C K-edge electron energy-loss spectroscopy (EELS) for the particle-like (CE5-P4) and adhering-like (CE5-P6) organic matter (Fig. 3G) both displays the features of amorphous carbon: a pronounced 1s → π* peak at 285.0 eV (C═C), followed above ~292 eV by a broad σ* continuum (20, 31). Weak heteroatom-related features are present in both specimens, but CE5-P6 exhibits more prominent shoulders at 286.1 and 288.1 eV, assignable to 1s → π* transitions of C═N and carbonyl/amide-type C═O (32), respectively, indicating a higher abundance of N- and O-functional groups in the adhering-like organics. This is consistent with the relatively higher N/C ratio for CE5-P6 (Fig. 4E). In the N K-edge EELS (Fig. 3H), both samples show a 1s → π* feature at 399 eV, indicative of sp²-bonded nitrogen (e.g., imine or pyridinic sites) (30, 33, 34), and a broad feature at 405 to 408 eV corresponding to C─N–related σ* transitions (20). These spectral characteristics resemble those commonly reported for IOM in carbonaceous chondrites (20) and Ryugu (19, 33) and Bennu (30) asteroid organics.

Together, the CE5 and CE6 organics represent a chemically heterogeneous, disordered carbonaceous assemblage comprising small sp² domains, aliphatic chains, and O- and N-bearing functionalities. The organic material may be oligomerized or exhibit a limited degree of cross-linking via amide (─CONH─) linkages (Fig. 1F). The paired carboxylate (ν_asym/ν_sym) signals (e.g., ~1610 and ~1410 to 1380 cm⁻¹) suggest the presence of carboxylate salts, potentially stabilized as ammonium salts (35), zwitterionic pairs (e.g., amino acid–like or amine-carboxylate ion pairs), or sodium salts.

Solar wind implantation and isotopes of organics

Solar wind implantation has been widely detected in the minerals of CE5 lunar regolith (36, 37). Here, we report the first identification of solar wind implantation signatures in lunar organics (Fig. 4). In adhering-like organic matter (CE5-P6 and CE5-P8), the H/C ratio decreases with depth (Fig. 4A), whereas δD increases (Fig. 4B): CE5-P6 rises from ~−190‰ at the surface to 0‰, and CE5-P8 rises from −170‰ to 180‰. For CE5-P6, the H/C decrease is primarily confined to the first approximately five frames, followed by an apparent plateau. Using the plateau segment of the depth profiles where H/C stabilizes, we estimate the original δD values to be −46.3 ± 40.8‰ (frames 10 to 19) for CE5-P6 and 109.6 ± 63.8‰ (frames 10 to 15) for CE5-P8.

By contrast, inclusion-like organic matter (CE5-P7) shows no clear depth-related changes in either H/C (Fig. 4A) or δD (Fig. 4B). The average δD values are 22.3 ± 33.6‰ for CE5-P7_r1 and 50.6 ± 25.6‰ for CE5-P7_r2. Overall, these δD values are lower than those of IOM in carbonaceous chondrites (24) and in samples from the asteroid Ryugu (Fig. 4C) (19). The inclusion-like organics appear to have been protected from direct solar wind implantation because of shielding by adjacent mineral topography.

For carbon and nitrogen isotopes, adhering-like organic matter (CE5-P6) yields δ¹³C = −40.4 ± 14.6 and δ¹⁵N = −99.8 ± 13.2‰; inclusion-like organic matter (CE5-P7_r2) yields δ¹³C = −76.8 ± 5.9‰ and δ¹⁵N = −56.6 ± 14.5‰. These δ¹³C and δ¹⁵N values are lower than those reported for carbonaceous chondrites (bulk) (24), the asteroids Ryugu (bulk) (1, 19) and Bennu (residue after water extraction) (2), comets (38, 39), and interplanetary dust particles (40). Given that the abundances of C and N in the solar wind are approximately four to five orders of magnitude lower than those of protons (H⁺) (41, 42), solar wind implantation is not expected to measurably affect C and N abundances or isotope compositions in organics. Assuming a scenario with H/C ≈ 0.5 and 50% of H attributed to the solar wind, the implied solar wind–derived N/C ratio is about 1 × 10⁻⁵, which is orders of magnitude below the measured N/C ratio of the organics (Fig. 4E). Therefore, any solar wind–driven shift in δ¹⁵N would be far below analytical precision, consistent with the lack of systematic depth-dependent trends in δ¹⁵N (fig. S8). H and N isotope measurements were conducted in separate NanoSIMS sessions (fig. S7); while partial sputtering before N analysis cannot be fully excluded, the negligible solar wind N contribution provides the primary explanation for the absence of a resolvable δ¹⁵N depth trend (fig. S8).

DISCUSSION

Assessment of the terrestrial contamination

The possibility of terrestrial contamination has historically complicated interpretations of lunar organic matter, particularly in Apollo regolith and breccias (16). Since then, studies of terrestrial contamination on blanks have shed more light on the challenges faced when identifying extraterrestrial organic material. In our CE-5/6 analyses, several strong lines of evidence disfavor a terrestrial origin for some of the organics reported here. Organics and minerals display hierarchical nesting and intergrowths: Organics encapsulate micrometer- to submicrometer-scale minerals (e.g., ilmenite) and occur in distinct modes on grain surfaces (Figs. 1 to 3), locally coexisting with moderately volatile elements (e.g., Na and, locally, Zn) (fig. S5 and data S1) (18, 23). These textural relations suggest in situ formation and modification under lunar conditions rather than simple adherence of terrestrial dust.

NanoSIMS depth profiles (Fig. 4, A and B) reveal solar wind implantation signatures within adhering-like organics, indicating long-term surface exposure that could not be replicated by laboratory contamination. Together, these lines of evidence indicate that while terrestrial contamination cannot be excluded in principle, the morphology, chemistry, and isotopic systematics of CE-5/6 organics are best explained by exogenous delivery coupled with impact-induced synthesis and surface reworking.

Formation pathways and surface processing of lunar organics

The amide-like functional signatures observed in our lunar organics (Fig. 1F) are similar to those reported for amide-bearing organic hotspots in returned asteroid samples. Amide-bearing organics with ¹⁵N-depleted compositions (δ¹⁵N = −213 ± 103‰, 2σ) in Ryugu C0052 are interpreted as reflecting nitrogen from distinct reservoirs or formation environments (29), potentially coupled with irradiation-driven synthesis, whereas negative δ¹⁵N (−28 ± 7‰, 2σ) values reported for some organic hotspots in Bennu have also been discussed in terms of low-temperature kinetic fractionation in cold environments (30). Unlike asteroid organics, which are commonly interpreted to record a continuum of processing from irradiation in the interstellar medium through accretion and subsequent parent-body aqueous alteration (21, 22, 43), the Moon is intrinsically volatile-poor (8). Therefore, lunar organics are more plausibly dominated by exogenous delivery, followed by impact-related modification and surface processing (4).

Although meteoritic and cometary delivery is an important source of lunar organics (13), lunar regolith commonly hosts organics as localized hotspots and nearly pure organic phases (Figs. 1 to 3). This occurrence contrasts with chondritic organic material, both in morphology (44) and in mineral settings, which are typically distributed within the fine-grained matrix (e.g., phyllosilicates) (45). It is more consistent with organic synthesis via impact-evaporation-condensation (4, 13). High-temperature/pressure shock decomposes, volatilizes, and redistributes preexisting organics (46, 47). Subsequent condensation and co-deposition of volatiles onto mineral surfaces yield concentrated organic phases tightly coupled to mineral interfaces (Fig. 3) (18, 48). Kinetic isotope effects during impact-evaporation-condensation may drive the δD, δ¹³C, and δ¹⁵N of the condensates to lower values relative to parent-body IOM (Fig. 4, C and D) (49, 50). In parallel, the preferential escape of light volatiles (e.g., CH₄, CO, and CO₂) (47, 51) leaves the condensates relatively nitrogen-enriched, yielding higher N/C ratios (Fig. 4E) (52).

The condensates overall resemble an amorphous carbon–like material while retaining or introducing N- and O-bearing functionalities. This study provides direct spectroscopic evidence for amide (─CONH─) linkages (Fig. 1F), indicating that limited cross-links (e.g., condensation-dehydration reactions between amino and carboxyl groups) (29) can be produced and preserved during the lunar impact-evaporation-condensation process (47, 53). Mechanistically, although transient shock heating can decompose many labile molecules (54), rapid quenching and recondensation can enable dehydration/condensation among amino- and carboxyl-bearing precursors, generating amide bonds and incorporating them into more cross-linked, aromatic/heterocyclic networks (52, 55).

After formation, these organics continue to be processed under lunar surface conditions, including ongoing solar wind implantation and irradiation (Fig. 4, A and B). Solar wind H⁺ implantation can introduce defects and radical sites, enhance structural disorder (28), increase H abundance (Fig. 4A), and shift δD toward more negative values (Fig. 4B) in lunar organics. The coupled depth-dependent variations in H/C and δD, followed by a plateau at depth, are therefore consistent with solar wind implantation overprinting a preexisting organic component.

The presence of complex, N-bearing organics in lunar regolith indicates that impacts on airless bodies both deliver and reprocess organic carbon in situ, concentrating it and enabling long-term preservation. These findings highlight the Moon as a natural laboratory for understanding how exogenous organics are modified and stabilized under impact-driven conditions, with broader relevance to other planetary bodies.

MATERIALS AND METHODS

Sample preparation

Nine lunar regolith grains (~100 to 500 μm) from CE5 (CE5C1000 YJFM00404) and eight grains (~100 to 800 μm) from CE6 (CE6C0200 YJFM001) were randomly selected and gently mounted by direct pressing onto an indium metal target without conductive coating (C, Au, or Pt). The samples were designated as CE5-P1 to CE5-P10 and CE6-P1 to CE6-P8 (table S1). This preparation method helps minimize potential contamination during handling and preserves the pristine state of the particle surfaces. These particles were subsequently analyzed sequentially by SEM-EDS, fluorescence microscopy, O-PTIR, Raman spectroscopy, NanoSIMS, and FIB-STEM-EDS-EELS (fig. S1). Sample storage and preparation were performed in a Class 1000 (ISO Class 6) cleanroom to minimize particulate and organic contamination.

SEM-EDS imaging

SEM-EDS analyses were performed using a Hitachi SU8600 field-emission scanning electron microscope equipped with an annular four-channel silicon drift detector (XFlash FlatQUAD, Bruker). To mitigate charging and enhance surface sensitivity, all measurements were acquired at a low accelerating voltage of 6 kV. The emission current was 20 μA, and the spot intensity was 80, corresponding to a probe current of ~300 pA (measured using a Faraday cup on the Hitachi SU8600). All SEM-EDS measurements were performed in a high vacuum. Minor charging was occasionally observed (fig. S2) but remained acceptable for imaging and EDS acquisition.

The large solid angle (>1.1 sr) of the FlatQUAD ensured sufficiently high count rates and reduced topography-induced shadowing even on rough surfaces. The working distance was ~11.5 mm. The EDS energy resolution was better than ~57 eV [full width at half maximum (FWHM)] at the C-K line (0.277 keV). For each particle, EDS mapping was first conducted over a wide field of view to locate carbon-rich hotspots; the field of view was then reduced for high-resolution elemental mapping and spot/area quantification of the targets. Quantification used the P/B-ZAF model in Bruker Esprit 2.5; for lines with energies <1 keV, the Φ(ρz) (phi-rho-z) approach was applied. Monte Carlo simulations at 6 kV indicated an x-ray generation depth and lateral radius of ~0.6 and 0.3 μm in organic hotspots, respectively. The results did not differ substantially among the analyzed organic domains. No beam-induced morphological changes or compositional drift was observed over repeated scans. EDS compositions (data S1) are region of interest (ROI)–averaged area analyses integrated over segmented organic hotspot regions from EDS maps.

To verify the qualitative/quantitative fidelity for light elements (C, N, and O), we analyzed organic reference materials containing C─N─O, including l-tryptophan (C₁₁H₁₂N₂O₂, 98% purity, Macklin) and cinchonine (C₁₉H₂₂N₂O, 99% purity, Macklin). Their nominal atomic ratios were used to assess the accuracy and potential bias of light-element quantification (table S1). The EDS-measured N/C ratio differed from the nominal value by <5% (relative error). O quantification showed large relative deviation at very low O abundance (cinchonine, O < 5 at %, up to ~70% deviation), whereas an O-richer standard (l-tryptophan, O = 18.75 at %) yields <20% deviation. Given the higher O abundances of lunar regolith organics (~15 to 30 at %), the expected O-quantification uncertainty is closer to the latter case.

As an independent verification, we requantified cinchonine by SEM-EDS at 3 and 5 kV using a CIQTEK SEM5000X equipped with a Bruker XFlash FlatQUAD detector, obtaining broadly consistent C─N─O results with those measured on the Hitachi SU8600 at 6 kV (data S1). We also assessed the carbon background on a clean Si wafer (~12.2 at %) by SEM5000X, which is comparable to the background level observed on a plagioclase fracture surface in our SU8600 measurements (~12.8 at %). We therefore emphasize that our EDS quantification focuses on C- and N-enriched organic hotspots, rather than the ubiquitous carbon background.

Fluorescence and O-PTIR measurements

O-PTIR analyses of the carbon-rich particles were performed using a mIRage-LS system (Photothermal Spectroscopy Corp.) equipped with a quantum cascade laser (QCL) mid-IR source and an integrated fluorescence imaging module for microscale infrared spectroscopy. The same regions identified by SEM-EDS were targeted via optical and fluorescence navigation. Fluorescence images were acquired in the AutoFL channel with 355- to 375-nm excitation and 440- to 880-nm emission collection; the signal is dominated by π-conjugated chromophores/fluorophores. The quantum cascade laser covered the wave number ranges of 2990 to 2700 and 1800 to 950 cm⁻¹. The 532-nm probe laser had an output power of 200 mW at the aperture and was attenuated to 0.09 to 1%, resulting in <1 mW at the sample surface. Depending on the signal response, the IR power was set to either 4.4 or 10%. An avalanche photodiode detector was used to detect the optical photothermal response. For the particles of interest, both O-PTIR point spectroscopy and imaging were performed. No sample damage was observed under the microscope. Band assignments were guided by reference spectra of organic functional groups (alkanes, alkenes, and carbonyl-containing compounds) (56–58).

Raman measurements

Raman spectroscopy was conducted on the carbon-rich regions identified by SEM-EDS, with optical navigation used to precisely target the same areas. Analyses were performed using a WITec alpha 300R equipped with a 532-nm laser and a 100×, 0.9–numerical aperture objective, providing a lateral spatial resolution of ~360 nm. The laser power was set to 1 mW, with an integration time of 5 s per acquisition, and the spectra were accumulated over five scans. The spectrometer was calibrated using the first-order peak of a single-crystal silicon at 520.7 cm⁻¹. In some cases, carbon-rich particles were too small to be confidently located under the Raman microscope, or strong fluorescence prevented spectrum acquisition. For the particles of interest, both Raman point spectra and imaging were performed. No laser-induced damage was observed.

In addition, using identical instrument settings, Raman spectra of a series of carbonaceous chondrites were acquired in situ on polished sections, rather than from isolated organic fractions. All Raman spectra were fit with two Lorentzian components for the D and G bands to obtain peak positions, FWHM, and intensity ratios (I_D/I_G). The Raman parameters of organic matter in CE5 and CE6 samples were compared with those of carbonaceous chondrites. As long as fitting procedures remain consistent across samples, the specific peak-shape model does not affect the comparative trends (27).

NanoSIMS analysis

The NanoSIMS analyses were carried out in two sessions (data S2), both performed in imaging mode. H isotopes were measured before C/N in a separate NanoSIMS session, and the ROIs were redefined for the C/N session (see dashed outlines in fig. S7).

H⁻, D⁻, ¹²C⁻, and ¹⁸O⁻ were measured using a 20-pA primary Cs⁺ beam. The mass resolving power was ~4000. The image field size ranged from 20 to 30 μm with a resolution of 512 by 512 pixels, and the dwell time was 1 ms per pixel. Each image frame required ~262 s, and a total of 15 to 25 frames were acquired depending on the signal decay. Because of the highly irregular topography of organic matter within the lunar regolith, the depth of sputter craters cannot be readily quantified by atomic force microscopy, The sputtering rate for silicate minerals such as glass or olivine is ~0.2 nm·μm²/pA·s (59, 60), corresponding to an average removal of ~2.6 nm per frame under the analytical conditions used here. In contrast, for biological materials, the sputtering rate is about 2.5 nm·μm²/pA·s (61), yielding ~33 nm per frame. On this basis, we infer that the actual sputtered depth per frame in our measurements likely ranges from several nanometers to the low tens of nanometers.

The hydrogen isotope results were expressed as δD with (D/H)_SMOW = 1.55 × 10⁻⁴. A coal working reference, with an average δD value of −147‰ (62), was used to correct for hydrogen isotope mass fractionation and H/C. The hydrogen background was monitored using the San Carlos olivine standard and was ~30 to 40 ppm. The vacuum in the analytical chamber was maintained at ~1 × 10⁻¹⁰ to 2 × 10⁻¹⁰ mbar, and a laboratory-developed cold trap was used to further reduce background hydrogen (63, 64).

Carbon and nitrogen isotopes (¹²C⁻, ¹³C⁻, ¹²C¹⁴N⁻, and ¹²C¹⁵N⁻) were analyzed at a mass resolving power of ~10,000 using a ~2-pA primary beam. The image field size was 20 to 25 μm with a resolution of 256 by 256 pixels, and the dwell time was 5 ms per pixel. Each image frame required ~327 s, and a total of 55 frames were collected. Under these analytical conditions, the sputtering depth is estimated to be ~0.33 nm per frame for silicate minerals, whereas it is ~4 nm per frame for biological materials. On this basis, we infer that the effective sputtered depth per frame for our organic-bearing regolith targets is likely intermediate, on the order of ~1 nm per frame.

The results were expressed as δ¹⁵N and δ¹³C with (¹⁵N/¹⁴N)_Air = 0.003673 and (¹³C/¹²C)_VPDB = 0.0112372. The IAEA-600 standard (caffeine, δ¹³C = −27.771 ± 0.043‰; δ¹⁵N = +1.0 ± 0.2‰) was used to correct for instrumental mass fractionation and quasi-simultaneous arrival (QSA) effects. QSA correction curves were established by varying the aperture slit size to obtain different k-values on the IAEA-600 (where k is the average number of secondary ions ejected per primary ion) (60, 62). Standard k-values ranged from 0.001 to 0.004, while sample k-values ranged from 0.001 to 0.003.

For both sessions, the results are reported as 1σ, which incorporates the internal precision of individual analyses and the uncertainty of the reference standard (including instrumental mass fractionation uncertainty after QSA correction). For depth profile analyses, the data were plotted using the mean value of each frame.

FIB-STEM-EDS-EELS analysis

TEM lamellae were prepared in a dual-beam scanning electron microscope/FIB (Thermo Fisher Scientific Helios 5CX) equipped with a 30-kV Ga⁺ ion column and an Omniprobe micromanipulator. Coarse trenching used a 7-nA ion current. Final thinning proceeded stepwise at 0.5, 0.3, and 0.1 nA down to 10 pA to obtain ~70-nm-thick lamellae. After thinning, low-kilovolt Ga⁺ cleaning was performed at 5 kV/48 pA and 2 kV/43 pA to minimize Ga implantation and the amorphous damage layer.

STEM-EDS was conducted on an aberration-corrected scanning/transmission electron microscope (Thermo Fisher Scientific Spectra Ultra) with a probe-side spherical aberration corrector (S-CORR), operated at 300 kV. EDS imaging used an Ultra-X six-detector silicon drift detector array. EDS mapping was acquired with a pixel dwell time of 5.00 μs. EELS measurements were acquired on a spherical aberration–corrected JEOL ARM 200F operated at 200 kV. Dual EELS was performed with a Gatan Continuum ER/1065 spectrometer to record the zero-loss and core-loss spectra. The energy resolution, estimated from the FWHM of the zero-loss peak, was better than 0.6 eV. Processing included zero-loss peak alignment (energy-axis calibration), removal of plural scattering by Fourier-ratio deconvolution, and spectral normalization, all performed in Gatan DigitalMicrograph (version 3.62).

Supplementary Materials

The PDF file includes:

Figs. S1 to S8

Tables S1 and S2

Legends for data S1 and S2

Other Supplementary Material for this manuscript includes the following:

Data S1 and S2