A giant collision dumps enormous energy into an icy moon, but new simulations show that shattered fragments shed the heat rapidly—so smaller moons can lose existing oceans or fail to form oceans that would otherwise have developed.

A collision powerful enough to break an icy moon into a cloud of debris appears, at first glance, to be an efficient way to make an ocean. The impact converts motion into heat. Gravity pulls the pieces together. The rebuilt moon should begin its second life with warmer ice and perhaps a buried sea.

New simulations show why that simple accounting can point in the wrong direction. The energy is enormous, but much of it arrives while the moon is no longer a moon. Its fragments have far more exposed surface area than the intact body, so they shed heat rapidly before and during reaccretion. For smaller moons, the collision also reorganises the interior in a way that removes a useful layer of insulation.

The finding comes from a peer-reviewed paper in Nature Astronomy led by Marc Neveu of the University of Maryland and NASA’s Goddard Space Flight Center. This is one study, not settled consensus. It examines a deliberately narrow set of severe, head-on impacts rather than reconstructing the history of a particular named moon. Its value lies in following the consequences across two radically different timescales: the violent first day and the next 4.5 billion years.

The ocean question begins with broken moons

Several icy moons around the outer planets may be the descendants of earlier bodies that were destroyed. The compact systems around Saturn and Uranus have inspired scenarios in which generations of moons collided, dispersed and assembled again. The same possibility is visible in another part of the outer Solar System. Space Daily recently reported that two small inner moons of Neptune may contain clay-rich material exposed from shattered larger predecessors.

This matters because some small and mid-sized moons are also candidate ocean worlds. Mimas, Enceladus and Dione at Saturn, and Miranda and Ariel at Uranus, have all entered discussions about present or past water beneath ice. The collision history and the ocean history therefore cannot be treated as separate stories.

Mimas is an especially useful warning against judging an interior by its surface. The moon looks heavily cratered and geologically quiet, yet a 2024 analysis placed a global ocean under its shell and suggested it formed only 5 million to 15 million years ago. Our earlier account of Mimas’s unexpectedly young ocean described a liquid layer that appeared long after the moon itself formed. An impact model must consequently ask not only whether water melts immediately, but whether the interior can reach melting hundreds of millions or billions of years later.

Four collisions connect a day of violence to an age of cooling

Neveu and colleagues Raluca Rufu, Alyssa Rhoden, Kevin Walsh and Yuval Steinberg connected two numerical tools. A smoothed-particle hydrodynamics model followed rock and ice through collision, fragmentation and reaccretion. A thermal-structural evolution model then followed the rebuilt body’s temperature, layering, melting and freezing over geological time. Each impacted result was compared with a no-collision control.

The collision calculation used about 500,000 particles and ran for 24 or 48 hours. The smaller target had a radius of 500 kilometres and was struck by a 250-kilometre-radius impactor at one, two and three times the pair’s mutual escape velocity. A second target, 1,000 kilometres in radius, was hit by a 500-kilometre-radius body at three times escape velocity. The moons orbited in the gravitational setting of a Uranus-like planet at about 10 Uranian radii.

These were head-on impacts, selected to maximise heating and disruption. In the team’s definition, a disruptive collision left a largest surviving fragment with less than half the target’s original mass. The setup therefore asks a useful extreme question. If a moon cannot gain a durable ocean after one of the hottest plausible disruptions, merely noting that a collision released a great deal of energy is not enough.

It is equally important not to reverse that logic too far. Four idealised collisions do not sample every impact angle, speed, composition or orbital environment. They establish a mechanism and show that it can dominate under the tested conditions. They do not provide a complete map of outcomes for every satellite system.

Fragments turn an impact’s heat into a short-lived pulse

The counterintuitive result begins with geometry. Break one cold sphere into many pieces and the combined surface area rises sharply relative to the material’s volume. Hot fragments can radiate energy into space much faster than the original moon. By the time gravity gathers them into a new sphere, part of the energy that seemed available for melting has already gone.

The accepted manuscript available through NASA’s Technical Reports Server quantifies how brief the bonus can be. Heat added during reaccretion was comparable to less than 10 million years of radiogenic heating in the roughly 500-kilometre moon, and about 30 million years in the 1,000-kilometre moon. Those intervals are small beside the billion-year histories over which buried oceans appear and disappear.

Falling surface debris did not supply a large delayed reservoir either. Depending on the run, reaccreted material formed layers only about 0.1 to 8 kilometres thick. A sensitivity calculation raised the initial post-impact temperature as high as 570 kelvin, well above water’s melting point under ordinary pressure, yet the smaller rebuilt moon still failed to keep a long-lived ocean. Extra melt strengthened convection, which carried heat outward and promoted faster freezing.

The potato comparison used in Southwest Research Institute’s account of the study is unusually apt. A whole baked potato keeps its heat; thin chips cool much more quickly. A shattered moon receives a fierce thermal pulse, but for an ocean the duration and location of heat matter as much as the peak temperature.

Small moons lose the disorder that kept them warm

Cooling fragments are only half of the smaller moon’s problem. Before impact, the 500-kilometre control retained a mixed outer region containing both ice and rock. That jumbled layer was structurally untidy but thermally helpful. Rock distributed through the ice reduced heat transport and acted as an insulating blanket over the warmer deep interior.

In the no-collision calculation, the deepest ice reached its melting point after roughly 250 million to 300 million years. The resulting subsurface ocean persisted for around a billion years before refreezing. It was not created by a sudden event. It emerged because long-term internal heat remained trapped long enough for melting to begin.

The disruptive impacts briefly melted enough ice for the materials to separate. Dense rock sank toward the centre while ice rose above it. The three smaller-moon impacts produced reaccreted bodies with radii of about 527, 421 and 364 kilometres. Each ended with a more distinct rocky core and cleaner ice mantle, but the newly organised ice shell let heat escape more efficiently than the mixed layer it replaced.

Under those model conditions, an ocean that existed at impact could be lost, and a future ocean that developed in the unstruck control could fail to form. The qualification is important: the paper’s overall conclusion is that impacts changed ocean presence for only part of a moon’s history. The collision can reset or shorten the liquid interval without permanently dictating every later state.

A thousand-kilometre moon tells a different story

The large target retained heat more effectively. Size changes the balance because volume grows faster than surface area, while stronger gravity aids reaccretion. In the uncollided 1,000-kilometre control, differentiation began after about 300 million years. An ocean appeared around 500 million years, grew to roughly 200 kilometres thick and then gradually diminished.

When the team applied its fastest impact to that target at about the time the control ocean appeared, the rebuilt moon had a radius near 902 kilometres. Its pre-existing ocean survived and became thicker for roughly two billion years. In a test with negligible later radiogenic heating, almost the entire ice layer melted and stayed that way for at least 800 million years.

That does not mean a giant impact reliably creates an ocean in a large icy moon. Across the simulations, no post-impact ocean appeared in a body that would otherwise have remained frozen. The pronounced enhancement required a late disruptive collision involving a 1,000-kilometre-class target, an event the authors consider unlikely in recent Solar System history.

Size, timing and prior thermal state therefore matter more than the word “giant” suggests. The same basic event can erase a small moon’s insulating structure while thickening liquid water in a larger body that already possessed it.

Changing the orbit did not provide a second heat source

A collision does more than heat and reshape a moon. It can alter the reaccreted body’s orbit, potentially exposing it to stronger tidal flexing. In the simulations, semimajor axes shifted by about 500 to 10,000 kilometres and eccentricities rose by roughly 0.002 to 0.04. The team also tested an eccentricity of 0.05 at about 250,000 kilometres from the Uranus-like planet.

Those changes did not generate enough solid-body tidal dissipation to melt the moons. The result is specific to the modelled distance and material response. A moon placed much closer to its planet could behave differently.

Tides remain one of the central explanations for ocean worlds. Space Daily’s earlier examination of tidally heated exomoons beyond the classical habitable zone describes the underlying principle: a slightly eccentric orbit lets a giant planet repeatedly flex a moon and replace heat that escapes. The new calculation says that the orbital disturbances generated by these four impacts did not create that sustained engine.

The simulations are informative because their boundaries are visible

The hydrodynamic collision model omitted material strength. Including strength could leave larger surviving pieces, produce less disruption and yield cooler reaccreted moons. After the first one or two days, the complicated three-dimensional debris state was converted into a spherically averaged, one-dimensional moon for the long thermal calculation. That makes billion-year evolution tractable, but it cannot retain every local hot region or compositional irregularity.

The post-impact ice was treated as water ice without ammonia or another antifreeze. Spin and obliquity were not explored, and debris that remained bound to the planetary system was not followed through every possible later encounter. The authors provide supporting outputs through a public Zenodo archive, allowing other researchers to inspect and extend the calculations.

There is also a gap between an ocean and a habitable ocean. These models track the presence and thickness of liquid water, not its salinity, chemical energy, exchange with rock or capacity to support life. An impact could influence all of those conditions without changing the simple yes-or-no answer about whether water remains liquid.

An ocean is a heat budget, not an impact scar

The tempting picture is a frozen moon struck like a bell, glowing internally long after the collision. The model offers a less cinematic sequence. The body shatters, hot fragments cool, debris falls back, rock sinks, ice rises and the rebuilt shell begins leaking energy according to its new structure. The final thermal state cannot be inferred from the violence of the opening event.

For small moons, that reorganisation can be more consequential than the heat deposited. For larger moons, the same event can preserve and deepen a sea that was already there. In neither size class did the tested impacts manufacture a durable ocean that the uncollided moon could never have formed.

This makes the study useful not because it supplies one rule for every icy satellite, but because it replaces a misleading shortcut. Energy delivered is not energy retained. To understand whether a moon can keep an ocean, its history must include how it was broken, how it reassembled and how efficiently the new body could lose heat afterward.