Origin of water on Earth

Where does Earth’s water come from, and could it have been delivered by comets?

Comparing the deuterium-to-hydrogen (D/H) ratio in cometary water with that of terrestrial water provides a way to test this hypothesis. As remnants of planet formation, comets preserve primitive material that can help us investigate both the origin of Earth’s water and the formation of our Solar System and other planetary systems. Their water carries an isotopic record of the chemical and physical conditions under which it formed and was subsequently processed. The terrestrial reference is Vienna Standard Mean Ocean Water (VSMOW), with D/H = 1.5576 × 10⁻⁴. Measurements consistent with this value support the possibility of a cometary contribution to Earth’s oceans, although an isotopic match alone cannot establish how much water comets delivered. Despite its scientific importance, this comparison remains limited by observational challenges. The current sample comprises just 13 Solar System comets, including two studied in situ by ESA’s Giotto and Rosetta spacecraft. Because deuterium-bearing species produce extremely faint spectroscopic signatures, successful remote measurements have largely been restricted to exceptionally favourable targets, many of them visible to the naked eye. Even with the most powerful instruments available, substantial uncertainties and methodological differences complicate the interpretation of the results.

[/caption]Figure 1 summarises the progress made in measuring the deuterium content of cometary water. These investigations began with ESA’s Giotto encounter with 1P/Halley in 1986. Two instruments, the Ion Mass Spectrometer (IMS) and the Neutral Mass Spectrometer (NMS), provided independent estimates of the D/H ratio. The original analyses yielded mutually consistent values of approximately 3 × 10⁻⁴, about twice the terrestrial ocean value (Balsiger et al. 1995; Eberhardt et al. 1995). A subsequent reanalysis of the NMS data by Brown et al. (2012) revised the result to (2.1 ± 0.3) × 10⁻⁴, closer to VSMOW. The exceptionally bright appearances of C/1996 B2 (Hyakutake) and C/1995 O1 (Hale-Bopp) then enabled the first ground-based detections of cometary HDO, a water molecule in which one hydrogen atom is replaced by deuterium. Observations of its 464.925 GHz transition with the Caltech Submillimeter Observatory (CSO) and the James Clerk Maxwell Telescope (JCMT), respectively, yielded D/H ratios broadly consistent with the original Giotto results (Bockelée-Morvan et al. 1998; Meier et al. 1998). However, these measurements required independent estimates of water production from observations of its dissociation products, introducing differences in observing time and the regions of the coma sampled.

Two other comets, C/2002 T7 (LINEAR) and C/2001 Q4 (NEAT), were investigated through the products released when sunlight dissociates water molecules. Ordinary water produces OH and H, whereas HDO can produce either OD and H or OH and D. Using the Ultraviolet and Visual Echelle Spectrograph (UVES) at the Very Large Telescope (VLT), Hutsemékers et al. (2008) observed the ultraviolet OH and OD bands near 310 nm in C/2002 T7. No individual OD line was detected, but combining the strongest expected lines yielded an approximately 3σ detection. The measured OD/OH ratio of (3.5 ± 1.0) × 10⁻⁴ was converted, accounting for the photodissociation branching ratios, into a water D/H ratio of (2.5 ± 0.7) × 10⁻⁴. In C/2001 Q4, observations obtained in 2004 with the Hubble Space Telescope’s Space Telescope Imaging Spectrograph (STIS) instead targeted the Lyman-α emission of atomic hydrogen and deuterium. Weaver et al. (2008) derived D/H = (4.6 ± 1.4) × 10⁻⁴, substantially above VSMOW. Another elevated value, (4.09 ± 1.45) × 10⁻⁴, was reported for the Halley-type comet 8P/Tuttle from observations obtained in 2008 using the VLT’s Cryogenic Infrared Echelle Spectrograph (CRIRES), targeting HDO transitions near 3.7 μm (Villanueva et al. 2009). These results extended the observed range of cometary D/H, although their uncertainties remained considerable.

The first D/H determination in a Jupiter-family comet provided a striking contrast. Observations of 103P/Hartley 2 with the Heterodyne Instrument for the Far-Infrared (HIFI) aboard ESA’s Herschel Space Observatory yielded D/H = (1.61 ± 0.24) × 10⁻⁴, consistent with terrestrial ocean water (Hartogh et al. 2011). The measurement used the HDO transition at 509.292 GHz. A similar approach applied to the Oort-cloud comet C/2009 P1 (Garradd) gave D/H = (2.06 ± 0.22) × 10⁻⁴, intermediate between VSMOW and the approximately twofold enrichment found in the earliest measurements (Bockelée-Morvan et al. 2012). These findings initially suggested a possible isotopic distinction between Jupiter-family comets, supplied from the trans-Neptunian region, and comets from the Oort Cloud. Subsequent results, however, showed that no simple division between these populations could explain the available data. Herschel substantially improved measurement consistency by observing water isotopologues nearly simultaneously and with closely matched telescope beams. Nevertheless, determining the reference water production rate still required either modelling an optically thick H₂O line or observing the rarer H₂¹⁸O isotopologue and adopting an oxygen isotopic ratio to infer the total water abundance.

Following the end of Herschel’s operations in 2013, measurements depended on the remaining facilities and favourable cometary appearances. Comet C/2014 Q2 (Lovejoy) illustrated the difficulty of comparing results obtained with different techniques. Paganini et al. (2017) measured HDO and H₂O simultaneously at infrared wavelengths using the 10-m Keck II telescope, obtaining D/H = (3.02 ± 0.87) × 10⁻⁴, or 1.94 ± 0.56 times VSMOW. Biver et al. (2016) instead combined observations of the 241.561 GHz HDO line from the IRAM 30-m telescope with later observations of H₂O and H₂¹⁸O at 556.936 and 547.676 GHz from the Odin satellite. Their result, approximately (1.4 ± 0.4) × 10⁻⁴, was consistent with terrestrial water. The two estimates differed by roughly a factor of two and agreed only at approximately the 2σ level. In the same study, Biver et al. reported D/H = (6.5 ± 1.6) × 10⁻⁴ for C/2012 F6 (Lemmon), based on an approximately 4σ HDO detection. This value, more than four times VSMOW, contrasted sharply with their result for Lovejoy despite both being Oort-cloud comets studied with similar methods. Such comparisons may reflect real diversity, but they also expose important observational limitations: measurements taken on different dates and with different beam sizes can sample different stages of rotational variability and different contributions from sublimating icy grains. The discrepancy for Lovejoy therefore cannot, by itself, establish a temporal change in the isotopic composition of its nucleus.

Rosetta’s extended investigation of 67P/Churyumov–Gerasimenko provided an unprecedented opportunity to measure cometary isotopic abundances in situ. Using the ROSINA Double Focusing Mass Spectrometer (DFMS), Müller et al. (2022) reported a water D/H ratio of (5.01 ± 0.41) × 10⁻⁴, approximately 3.2 times VSMOW, with little variability in the measurements they analysed. Rosetta also enabled investigations of deuterium enrichment in other molecules, including methane, ethane, propane, butane, methanol, hydrogen sulphide and ammonia. However, a reanalysis of more than 4,000 water D/H measurements by Mandt et al. (2024) challenged the interpretation of the water results. The authors proposed that preferential adsorption of HDO onto dust grains, followed by the release of deuterium-enriched water from their icy coatings, could raise the D/H ratio measured locally near the nucleus. Using post-perihelion measurements obtained more than 120 km away, they derived D/H = (2.59 ± 0.36) × 10⁻⁴, approximately 1.7 times VSMOW, which they interpreted as more representative of the nucleus. This interpretation remains debated: a subsequent analysis by Müller et al. (2026), accounting for hydronium contributions and differences in mass calibration, recovered a higher mission-average value of (4.67 ± 0.59) × 10⁻⁴. The disagreement highlights the importance of distinguishing the bulk isotopic composition of nucleus ice from local coma effects and uncertainties in the analysis of overlapping mass-spectral signals.

The close approach of the Jupiter-family comet 46P/Wirtanen in 2018 provided another opportunity to compare cometary water with Earth’s oceans. Observations with the Stratospheric Observatory for Infrared Astronomy (SOFIA) targeted the H₂¹⁸O and HDO transitions near 547 and 509 GHz. The resulting D/H ratio of (1.61 ± 0.65) × 10⁻⁴ was consistent with VSMOW, although its derivation required an assumed oxygen isotopic ratio, as in the corresponding Herschel measurements. Lis et al. (2019) used this result to propose a possible relationship between D/H and the fraction of a comet’s surface required to explain its water production, potentially linked to the contribution of sublimating icy grains. A further terrestrial-like value was reported by Cordiner et al. (2025) for the Halley-type comet 12P/Pons–Brooks: ALMA observations obtained in 2024 yielded D/H = (1.71 ± 0.44) × 10⁻⁴. Although ALMA mapped both HDO and H₂O, the two species were observed on different dates. The analysis therefore used simultaneous methanol observations to connect the datasets, supported by independent checks of the methanol-to-water ratio. Observations of the 183 GHz water line also required careful correction for absorption by Earth’s atmosphere. Other searches, including those targeting 153P/Ikeya–Zhang, C/2007 N3 (Lulin), C/2012 S1 (ISON) and 45P/Honda–Mrkos–Pajdušáková, yielded upper limits rather than detections.

The contribution of comets to Earth’s water therefore remains unresolved. Measurements consistent with VSMOW show that some comets contain water isotopically compatible with our oceans, but the small sample and the diversity of observational methods prevent a robust assessment of the comet population as a whole. Low signal-to-noise ratios, non-simultaneous observations, differences in spatial coverage, optical-depth effects and assumptions about reference isotopic abundances further complicate comparisons. Resolving these limitations requires a much larger sample of homogeneous, internally consistent measurements with high signal-to-noise ratios. A dedicated instrument capable of measuring hydrogen and deuterium simultaneously with high senstivity, over the same region of the coma, would provide a decisive advance. Beyond testing the cometary contribution to Earth’s oceans, such a dataset would constrain the chemistry of water in the early Solar System and the formation and redistribution of the planetesimals from which planetary systems emerge.