
Collisions between planets, planetesimals and other objects are common in the galaxies and essential for planet formation. Researchers are focusing on these collisions for clues about which exoplanets have greater or lesser potential habitability. (NASA)
What can get the imagination into super-drive more quickly than the crashing of really huge objects?
Like when a Mars-sized planet did a head-on into the Earth and, the scientific consensus says, created the moon. Or when a potentially dinosaur-exterminating asteroid heads towards Earth, or when what are now called “near-Earth objects” seems to be on a collision course. (There actually aren’t any now, as far as I can tell from reports.)
But for scientists, collisions across the galaxies are not so much a doomsday waiting to happen, but rather an essential commonplace and a significant and growing field of study.
The planet-forming centrality of collisions — those every-day crashes of objects from grain-sized to planet-sized within protoplanetary disks — has been understood for some time; that’s how rocky planets come to be. In today’s era of exoplanets, however, they have taken on new importance: as an avenue into understanding other solar systems, to understanding the composition and atmospheres of exoplanets, and to get some insight into their potential habitability.
And collision models, it now seems likely, can play a not insignificant role in future decision-making about which planetary systems will get a long look from the high-demand, high-cost space telescopes that will launch and begin observing in the years ahead.
“We’re learning that these impacts have a lot of implications for habitability,” said Elisa Quintana, a NASA Ames Research Center and SETI Institute research scientist who has been modeling space collisions. Her paper was published in 2016 in the Astrophysical Journal, and took the modeling into new realms.
“When you think of what we know about impacts in general, we know they can effect a planet’s spin rate and rotation and consequently its weather, they can bring water and gases to a planet or they can destroy an atmosphere and let the volatiles escape. They effect the relationship between the planet’s core and mantle, and they determine the compositions of the planets. These are all factors in increasing or decreasing a planet’s potential for habitability.”

An artist rendering of a protoplanetary disk around a newly-formed star. Tiny grains of dust grow over millions of years into planetesimals and planets through collisions and the accretion of matter. (NASA)
To better understand the logic of impacts, Quintana ran two-billion year simulations of the protoplantary disk-to-mature solar system process, and produced a unique look at how those systems and their planets form. What’s more, she did it 280 times, which is many more simulations than has been done in the past, and came up with the creation of 164 Earth-like worlds.
Starting about 10 million years after the solar system formed, she recreated dynamics from our own solar system to make the simulations, but tweaked the starting points to make them applicable to other extra solar systems.
In the past, impact simulations like these generally used a “perfect accretion” model, which meant that all the material from one planet or moon-sized body would stick to the larger one it hit. But we know that is actually not what always happens — that both the impacted and impactor can fragment and eject rock into the sky. But this scenario, as well as a less dramatic “hit and run” impacts, is hard to model.
Yet the new Quintana et al model does indeed add these kinds of break-ups into the equation. The result is a model that is governed by the known physical and protoplanetary rules, but with a large dose of chaos.

Elisa Quintana is a research scientist at the SETI Institute and at the NASA Ames Research Center. (SETI Institute)
As Quintana explained: “One simulation could produce 3 final terrestrial planets. If you run the same simulation again, but move one rock in the disk by 1 meter (keeping everything else exactly the same) the simulation could produce 5 planets. The butterfly effect!”
What this means is that the architecture of any solar system is but one of many that could have been produced by the same protoplanetary disk.
What she reported in that paper and from subsequent work is that:
- The models generally produce three to four rocky inner planets, as in our solar system.
- In a system with giant planets like Jupiter and Saturn, the process of increasingly large bodies colliding is roughly 200 million years long. At that point, the rocky inner planets of the system would have been formed, and the material to add significantly to the planets (or the system) would be largely depleted.
- But in a system without a Jupiter or Saturn, the process of accreting material onto planets takes much longer and moon-sized objects and smaller planetesimals remain prevalent at 500 million years. Indeed, she said, these substantial but not planetary orbiting objects would probably be present in an inner solar system even 4.5 billion years later (the age of our solar system.) In other words, the inner solar systems would be filled with objects like the crowded Oort Cloud of our system — the regions some 100,000 times further from the sun than Earth.
- While impacts are common in the first model they generally end quite abruptly — and with a very big bang. Since the inner solar system has been largely cleared of smaller objects by this time, what’s left is large ones that exert increasingly great gravitational pulls. The result is the kind of giant impact that formed our moon. Every simulation run with a Jupiter or Saturn in place delivered at least one giant impact to each inner rocky planet.
Histogram of the total number of giant impacts received by the 164 Earth-like worlds produced in the authors’ fragmentation-inclusive simulations. [Quintana et al. 2016]
The models were run on the Pleaides supercomputer at NASA Ames Research Center, which allowed for additional factors (the fragmenting of planets, those hit-and-runs) to be included. And then its enormous capacity allowed for so many more models to be run, and run quickly.
Below are animations of the first scenario (with a Jupiter and Saturn already in place, as it would be at 10 million years after disk formation) and the second without the giant planets.
The green lines are orbits of moon-sized planetesimals, the blue lines Mars-sized planet embryos, and the red is “fragmented material” the size of half a moon kicked during impacts. (More on this later.) The Jupiter is purple and the Saturn is yellow. (Pop-up button on upper right allows for full screen view.)
As you’ll see, the planet formation process ends quite early where a Jupiter and Saturn are in the system — they dominate the gravitational dynamics and clear out smaller objects quickly. But with a sun only and no larger planets, the smaller objects remain for billions of years.
Animations by Chris Henze, NASA Ames

This artist concept illustrates how a massive collision of objects, at 1,200 miles in diametert perhaps as large as the dwarf planet Pluto, smashed together to create the dust ring around the nearby star Vega. New observations from NASA’s Spitzer Space Telescope indicate the collision took place within the last one million years. Astronomers think that embryonic planets smashed together, shattered into pieces, and repeatedly crashed into other fragments to create ever finer debris. (NASA)