Back to Index Page

© 2005 by Donald F. Robertson.



This article may be distributed at will, but only if it is not changed in any way, and only if the author's name, the copyright notice, the name of the journal it first appeared in, and this notice remain attached. In addition, this article may not be sold for money, or published for sale in any way, without the author's prior written permission.

This article originally appeared in Astronomy Now.


Most moons in the Solar System are very small compared to their planets. Ganymede is bigger than the planet Mercury but it must be a grave disappointment to its parent, Jupiter. Even Titan, in spite of its name and for all its methane streams and rain, is positively tiny compared to Saturn.

There are two exceptions. Earth's moon weighs in at about 1.2 percent of Earth. Pluto's Charon is ten to fifteen percent of Pluto's mass. According to Dr. Robin M. Canup of the Southwest Research Institute in Boulder, Colorado, all other satellite-to-planet mass ratios in our star system are less than about 0.0002 to one.

In the past, the moon and Charon were considered overweight flukes. As more has been learned about Earth's moon and how it may have formed, Dr. Canup told the American Geophysical Union meeting in San Francisco, giant moons have begun to look less of a surprise.

It is widely believed that our moon formed when a Mars-sized proto-planet collided with the early Earth. Large impacts between proto-planets may not have been rare events early in the history of the Solar System. If these occasionally resulted in a large moon, giant satellites may not have been particularly rare either. Computer models of worlds in collision support this idea.

The catastrophic theory for the moon’s formation seemed rather, well, catastrophic, to many scientists when it was first proposed in the mid-1970s and it was not an easy sell. However, according to Dr. Canup, attempts to explain the moon's birth non-catastrophically failed to describe key features of the Earth-moon family. Moreover, the collisional hypothesis provided a ready explanation for some of the moon's oddest features.

First, given its large size, the moon does not seem to weigh enough. Its density is far less than Earth's. Though it does appear to have a very small iron core, overall the moon is depleted of iron. It is also very dry, with few volatile elements. Both facts are hard to explain.

Several lines of evidence suggest that Earth's moon was born of material originating approximately the same distance from the sun as Earth is today. If so, the primordial gas and dust that condensed to form the Earth, and the materials that made up the moon, should have been very similar. So, why is the density and volatile content of these two worlds so different today?

Second, and even more curious, is the total angular momentum of the Earth-moon system. That is a measure of the total rotational motion of the Earth and moon together, with contributions from the Earth’s spin and the Moon’s orbital motion around the Earth. This total rotational motion is believed to have been roughly conserved, or retained, over the long history of the Earth-moon system.

The Earth-Moon angular momentum appears to be very high. The other terrestrial planets -- those located in the inner Solar System and broadly similar to Earth in mass and composition -- rotate at about the same rate as Earth (Mars) or more slowly (Mercury and Venus). Yet, if all of the angular momentum in the moon’s orbit around the Earth were added to that of the Earth’s rotation, then our planet would have about a four-hour day, Dr. Canup told the conference. Where did all that rotational momentum first come from?

A collision with the proto-Earth with another proto-planet could answer both of these questions. In crude terms, if Earth's moon were made of material splashed up from a giant impact, most of that material would come from the relatively light mantle and crustal rocks of both the impactor and Earth. It would not easily come from either world’s deeply buried iron core. As the splashed-up material cooled and fell back together to form the moon, the end result would be a world depleted in iron. High temperatures generated in the impact would tend to drive off the water and other volatiles. If the impact was not head-on but happened at an angle, it could spin up the early Earth. That would explain the excess angular momentum we see today.

Dr. Canup has been running computer simulations that try to model the moon's formation in an impact. These programs track large numbers of overlapping model "particles." Each mathematical particle represents a three-dimensional collection, or averaged distribution, of the actual material involved in the impact. (No computer could track every particle of matter created when a Mars-sized object hits Earth!) In this "smooth particle hydrodynamic" model, the simulated particles are attracted by gravity and pushed apart by pressure, much as the individual particles would be in the real explosion created by a real collision between two large proto-planets. The simulation also tracks the heating experienced by the modeled particles as they are compressed early in the collision, and cooling as they expand later on. The evolving trajectory and other properties of each particle is tracked as the modeled impact plays out.

The impact simulation is run many times, each starting with slightly different proto-planets striking each other at different angles and velocities. Dr. Canup looks for end results that resemble the conditions needed to produce the real Earth-moon system.

It does appear that the conditions "most successful" at generating a moon like the one we see in our own sky are created when an object ten to fifteen percent of Earth's mass grazes the Earth at a rather sharp angle. The angle would have been about forty-five to fifty degrees, Dr. Canup told Astronomy Now.

Ten to fifteen percent seems pretty small, but this would have been a truly titanic event. One model described at the conference involved a Mars-sized body striking the proto-Earth at an angle, disrupting but not completely destroying the impactor. Surrounded by rocky debris, the shattered iron core of the smaller world recoalesces as it pulls away from the Earth. Hours later, before completing one highly elliptical orbit, it falls back to the Earth's surface and begins to sink toward Earth’s center, removing most iron from the material that will form the moon. Another clump of lighter, mostly rocky debris also falls back, but just misses Earth's surface. Tidal forces pull it apart into long streams of asteroid-like debris. Within a day, the proto-Earth is surrounded by beautiful and structurally-complex rings, much like those around Saturn today, but made of high temperature molten rocks and vapor instead of shattered ice. This material cools and recoalesces to form the moon, probably within a hundred years.

A more oblique impact leaves more of the incoming world intact, since only one side of the impactor strikes the proto-Earth. The rest misses, but is slowed enough to end up in orbit. Intact parts of the impactor could have created several small moons which later combine in their own titanic collisions. Since this particular model would be less efficient at concentrating iron in the Earth, it may be less like the real collision that created the real, iron-depleted moon.

If one can model the creation of the Earth-moon system, why not the superficially similar Pluto-Charon system? Dr. Canup did just that, using tens to hundreds of thousands of modeled "particles" evolving over one to four days. Very little is known about this distant pair of worlds, so different runs assumed different, but reasonable, compositions for both Pluto and Charon. These ranged from uniform frozen bodies assembled from unsorted boulders of rocks and ices -- essentially unchanged since the beginnings of the Solar System -- to highly evolved differentiated worlds with rocky cores and icy crusts.

As in the Earth-moon family, the name of the computer game was to come up with initial conditions that would result in the observed size, orbit, and the total angular momentum of the real Pluto-Charon system.

The models suggest that Charon could have formed from the intact impactor itself -- more like the second model for our moon -- rather than recoalescing from rings or a disk of debris, like the first lunar model. In these models, only the outer surface of one side of the impactor grazes Pluto, at a more shallow angle than any proposed for proto-planets striking the Earth. The rest of the impactor flies by Pluto remaining more-or-less intact, but slowed enough to enter an elliptical orbit. Tidal evolution gradually makes Charon's orbit more circular.

A low impact speed seems to be needed to form the Pluto-Charon system from an impact,” said Dr. Canup. That suggests “the orbits of Pluto and the impactor around the sun might not have been very elliptical at the time. Since most of the incoming world's body missed Pluto, an intact grazing collision may not have experienced much heating. Pluto and Charon's densities indicate that both are still about half water ice; clearly, all of this water was not vaporized and lost in the impact.

One result stood out across the models. Undifferentiated impactors striking Pluto tend to produce intact moons, while differentiated impactors do not. The first condition appears to match expectations best, since thermal modeling of an impactor a third to half of Pluto's size suggests that it never would have become warm enough to let the rocks settle to its core. Pluto, because of its larger size -- and also because of heating during the Charon-forming impact -- probably was differentiated into a rocky core and a mantle of ices. These results can be tested in 2016 or 2017 when the New Horizons probe flies by Pluto on its way further out into the Kuiper Belt of icy debris beyond Neptune.

Dr. Canup finds the history of the "giant impact" theory for the origin of Earth's moon ironic. It was first proposed, she told Astronomy Now, to help account for the apparent "uniqueness of the moon in the inner Solar System. It was thought that such an outcome might be a rare collisional event. But the actual modeling shows that moons -- of various sizes and properties -- were probably being formed frequently. They were byproducts of the very types of collisions between early proto-planets that yield Earth-sized planets. It seems reasonable instead to believe that most of the solid planets in the Solar System probably had impact-produced satellites at some point in their early history."

Where are these moons now? Why don't Mercury, Venus, or Mars have large satellites?

They could have formed only to be subsequently shattered by later impacts or shoved into unstable orbits by tides. Even Earth's moon is not an unchanging fixture. Tidal friction between Earth's oceans and their floors under the influence of the moon’s gravity acts to slow the Earth's spin. Conservation of angular momentum causes the moon’s orbit to drift farther away. Likewise, if the Earth had been rotating slowly after the lunar-forming impact, instead of rapidly as needed to account for our current twenty-four hour day, Dr. Canup suggests, "the moon would have tidally evolved inward and have been lost as it crashed into the Earth."

Like all the giant Jovian planets, Saturn and its family of satellites is analogous to a miniature star system. Saturn serves as the “star,” and Saturn's moons as the “planets.” Herschel Crater on Mimus, and Odysseus, the "eye of Tethys”, provide dramatic evidence that these worlds, too, experienced extraordinary impacts.

If large satellites orbit Earth or Pluto, why don't Mimus or Tethys, or any of the big satellites of Jupiter or Neptune's Triton, have large satellites? Dr. Canup told Astronomy Now that such systems "might have fleetingly existed in the course of early Solar System formation." The more rapid tidal evolution in a giant planet's miniature system would make the lifetimes of any satellites of satellites "extraordinarily short compared to Solar System timescales." Long ago, they would have crashed into their parent moons or escaped into a free orbit around the planet, getting swept up by one of the moons that did survive.

Ultimately, the best way to determine if the impact model of moon formation makes sense is to look at other real star systems. NASA is designing future space telescopes powerful enough to detect terrestrial-type worlds around nearby stars. Someday, we may be able to determine if many of those worlds have big moons, like Pluto and Earth, and it would then be clear that large-moon formation is a frequent event.

Earth's moon would not be a fluke, but wholly normal. Since oceanic tides driven by the moon are believed to have had an important role in the formation and rapid evolution of life on Earth, that would have implications for the frequency of complex life elsewhere in the galaxy.

The next time you see a beautiful summer's moon rising over the hills, consider how its very uniqueness may be telling us how common such sights may really be.

Donald F. Robertson is a freelance space industry journalist based in San Francisco.

Back to Index Page