A glass of water seems an unlikely souvenir from an age when collisions melted rock. Yet a hot young planet became Earth with its ocean. Did comets bring all the water after its birth?
Earth assembled from material in the young Solar System about 4.5 billion years ago. Hydrogen could arrive both in its building material and in delivered bodies; chemical reactions could also produce water. There is no established single source for all terrestrial water or exact division of contributions. We need to explain both the origin of the ingredients and how they reached the ocean. NASA: Earth’s formation, research on the hydrogen budget and its model limitations.
Earth grew rather than appearing as a finished sphere
In the disk around the young Sun, solid particles provided material for larger bodies. Their further growth by capturing material and colliding is called accretion. When large bodies struck one another, energy of motion became heat. Radioactive decay added heat too. Large parts of the young planet could melt.
Within a molten body, metal separated from silicate material: the denser metallic phase sank and built the core. Earth’s material therefore did not simply mix once and remain unchanged. Elements partitioned between metal, mantle and atmosphere; some gas was subsequently lost to space. Young, Shahar and Schlichting: heating, melt and exchange between reservoirs.
For water, this matters: having no visible liquid ocean does not mean lacking its chemical ingredients. Hydrogen can be chemically bound, dissolved in melt or present in gas. A hot surface and a steam atmosphere represent a different state of water, not evidence of its complete absence.
Delivering hydrogen and delivering ready-made water are different routes
Meteorites are accessible samples of material related to the early Solar System. Candidates include carbonaceous chondrites, some rich in water-bearing substances. Such bodies could add water and other volatile components to the growing Earth. “Delivery” need not mean one late shower after formation was complete.
Another research direction concerns enstatite chondrites. Several isotope signatures resemble those of terrestrial material. Their minerals were long taken to indicate nearly waterless building blocks. Yet these meteorites contain hydrogen, including hydrogen associated with sulfur. Investigating this bond reveals a possible supply of a water ingredient, rather than ancient ocean drops inside every stone. Barrett and colleagues: sample, measurements and terrestrial contamination checks.
There is also a chemical route: in models, hydrogen in a primordial atmosphere reacts with oxygen from the magma ocean’s material. Water can form without its entire volume being delivered as ready-made ice. But the model requires an atmosphere, its retention and sufficient exchange with the melt. Successfully reproducing several properties of Earth does not prove that only this route actually occurred. The original atmosphere–magma ocean model.
Three questions about the origin of water
Hydrogen in building material
- Evidence to test
- Meteorite hydrogen and isotopes; terrestrial contamination of samples.
- What the evidence does not establish
- How much survived the entire accretion history.
Delivery by water-bearing bodies
- Evidence to test
- Composition, isotopes and possible delivery trajectories.
- What the evidence does not establish
- A unique source from a D/H match alone.
Atmosphere–melt reactions
- Evidence to test
- Hydrogen retention, chemical equilibrium and properties of the products.
- What the evidence does not establish
- That a successful model is the only real history.
Isotopes narrow the search without naming the supplier
An ordinary hydrogen nucleus has one proton. Deuterium also has a neutron. The ratio of deuterium to ordinary hydrogen, D/H, can be compared in water from different objects. It preserves information about formation conditions and the material’s subsequent history.
Imagine two bottles with the same salinity: similarity does not prove they were filled in the same place. Likewise, a close D/H makes a source compatible with a particular history but does not identify a unique address. Different sources can have similar compositions; mixing, chemical exchange and gas loss alter the original picture. Even measuring cometary water requires separating signals from gas, dust and the instrument. Mandt and colleagues: isotopes and the difficulties of coma analysis.
Researchers therefore also examine other isotope systems, meteorite compositions and whether a formation model allows the required delivery and retention. We cannot declare comets the source of the entire ocean, or rule out every comet on the basis of one object.
Branch · mixing modelOptional · you can skip this
Why is one ratio insufficient?
Take imaginary sources with ratios R_(A) = 1 and R_(B) = 3 in identical arbitrary units. With no losses or isotope redistribution:
R_(mix) = fR_(A) + (1 − f)R_(B).
Here f is the fraction of ordinary hydrogen atoms supplied by A, not a fraction of rock mass. All ratios are dimensionless. For f = 0.75, we obtain 0.75 × 1 + 0.25 × 3 = 1.5. At f = 0 we have only B; at f = 1, only A.
When both endpoint compositions are known and the conditions hold, the fraction can be found. But a third unknown source or a later change in D/H adds unknowns. One measured number no longer determines the whole history. These numbers are educational: they represent no particular comet or meteorite and estimate no contribution to Earth.
An ocean results from retention and redistribution
As the surface cooled and pressure and temperature allowed, water vapour could condense. Outgassing from the interior transferred water to surface reservoirs. These processes explain an ocean forming from ingredients already present; a volcano does not by itself explain where its hydrogen originated.
Counting only today’s ocean is insufficient too. Some hydrogen is in the interior, and estimates of its amount in the core depend on experiments and models of partitioning between metal and silicates. A “water equivalent” for hydrogen in metal does not mean an underground liquid ocean. A 2025 study accounting for the joint partitioning of hydrogen and carbon required a carbonaceous-chondrite contribution in its models; the authors explicitly state the dependence on equilibrium assumptions. Experiments and limitations of core-formation models.
Earth thus formed through growth, heating and separation of material. Its water connects ingredient delivery, chemistry, losses and cooling. Evidence lets us test individual parts of that history, but does not yet yield one proven recipe for all the water. To distinguish samples and possible suppliers more clearly, return to asteroids, comets and meteorites.