The Origin of the Moon

Page design by Gregg Herres and William K. Hartmann

The Giant Impact, as pictured in a painting by William K. Hartmann on the cover of Natural History Magazine in 1981. Copyright William K. Hartmann

The idea in a nutshell: At the time Earth formed 4.5 billion years ago, other smaller planetary bodies were also growing. One of these hit earth late in Earth's growth process, blowing out rocky debris. A fraction of that debris went into orbit around the Earth and aggregated into the moon.

Half an Hour After the Giant Impact, based on computer modeling by A. Cameron, W. Benz, J. Melosh, and others. Copyright William K. Hartmann

Why this is a good hypothesis:

• The Earth has a large iron core, but the moon does not. This is because Earth's iron had already drained into the core by the time the giant impact happened. Therefore, the debris blown out of both Earth and the impactor came from their iron-depleted, rocky mantles. The iron core of the impactor melted on impact and merged with the iron core of Earth, according to computer models.
• Earth has a mean density of 5.5 grams/cubic centimeter, but the moon has a density of only 3.3 g/cc. The reason is the same, that the moon lacks iron.
• The moon has exactly the same oxygen isotope composition as the Earth, whereas Mars rocks and meteorites from other parts of the solar system have different oxygen isotope compositions. This shows that the moon formed form material formed in Earth's neighborhood.
• If a theory about lunar origin calls for an evolutionary process, it has a hard time explaining why other planets do not have similar moons. (Only Pluto has a moon that is an appreciable fraction of its own size.) Our giant impact hypothesis had the advantage of invoking a stochastic catastrophic event that might happen only to one or two planets out of nine.

What were some earlier ideas?

1. One early theory was that the moon is a sister world that formed in orbit around Earth as the Earth formed. This theory failed because it could not explain why the moon lacks iron.
2. A second early idea was that the moon formed somewhere else in the solar system where there was little iron, and then was captured into orbit around Earth. This failed when lunar rocks showed the same isotope composition as the Earth.
3. A third early idea was that early Earth spun so fast that it spun off the moon. This idea would produce a moon similar to Earth's mantle, but it failed when analysis of the total angular momentum and energy involved indicated that the present Earth-moon system could not form in this way.

Where did the theory come from?

Hartmann and Davis were familiar with the work done in the Soviet Union in the 1960's, on the aggregation of planets out of countless asteroid-like bodies called planetesimals. Much of this work was pioneered by a Russian astrophysicist named V. S. Safronov.

Picking up on Safronov's general ideas, Hartmann and Davis ran calculations of the rate of growth of the 2nd-largest, 3rd largest, etc., bodies in the general vicinity of Earth, as the Earth itself was growing. Just as the asteroid belt today has a largest asteroid (Ceres) at a 1000 km diameter, and several smaller bodies in the 300-500 km diameter range, the region of Earth's orbit would have had several bodies up to about half the size of the growing Earth. Our idea was that in the case of Earth (but not the other planets) the impact happened late enough, and in such a direction relative to Earth's rotation, that abundant enough middle material was thrown out to make a moon.

How did the theory develop?

After we first presented the theory in 1974 at a conference on satellites, Harvard researcher A. G. W. Cameron rose to say that he and William Ward were also working on the same idea, but coming at it from a different motivation -- the study of angular momentum in the system -- and that they had concluded the impacting body had to be roughly Mars size (a third or half the size of Earth). Our paper was published in 1975 (Hartmann and Davis, Icarus, 24, 504-505) Cameron and Ward published an abstract on this idea at the Lunar Science conference in 1976, two years after the PSI paper.

Five Hours After Impact, based on computer modeling by A. Cameron, W. Benz, J. Melosh, and others. Copyright William K. Hartmann

Some work was done by Thompson and Stevenson in 1983 about the formation of moonlets in the disk of debris that formed around Earth after the impact. However, in general the theory languished until 1984 when an international meeting was organized in Kona, Hawaii, about the origin of the moon. At that meeting, the giant impact hypothesis emerged as the leading hypothesis and has remained in that role ever since. Dr. Michael Drake, director of the University of Arizona's Planetary Science Department, recently described that meeting as perhaps the most successful in the history of planetary science.

A collection of papers from that meeting was published by the Lunar and Planetary Institute (Houston) in the 1986 book, Origin of the Moon, edited by PSI scientist William Hartmann, together with Geoffry Taylor and Roger Phillips. This book remains the prime reference on this subject. In the meantime, researchers such as Willy Benz, Jay Melosh, A. G. W. Cameron, and others have attempted computer models of the giant impact, to determine how much material would go into orbit. Some of these results have been used by Hartmann to make the paintings on this web page, attempting to show how the impact would have looked to a human observer (if humans had been around -- they didn't come along until 4.5 billion years later!)

In the 1990's, Dr. Robin Canup wrote a Ph.D. dissertation on the moon's origin and the giant impact hypothesis, which produced new modeling of the aggregation of the debris into moonlets, and eventually, into the moon itself. Dr. Canup is continuing the modeling of the lunar accretion process.

Current status: In 1997, Dr. Canup's work received a great deal of publicity by media news sources, some of whom mistakenly thought that the giant impact was a brand new idea. Canup's early work, presented in July 1997, suggested the debris from an impact might not make a moon, but only a swarm of moonlets. Her later work (fall 1997) led to more "success" in aggregating the debris into a single moon.

Moon Forming Out of Rings Copyright William K. Hartmann

Studies of lunar rocks show that the moon originally had a molten surface. As this so-called magma ocean cooled, intense volcanism continued for about 900 million years. An early volcanic eruption is shown here. Copyright William K. Hartmann

Thus, the giant impact hypothesis continues to be the leading hypothesis on how the moon formed. Is it right? Can it be disproven by more careful research? Only time will tell, but so far it has stood up to 25 years of scrutiny. At PSI we have worked with several leading researchers to propose new work or the accretion mechanics using a variant of the PSI planet building model. But this work has not been funded.

The Moon Today. This late afternoon scene on the moon typifies the moon as it has been for about 3 billion years. Volcanism has ended. Meteorite impacts are rare. The quiet landscape awaits the return of human explorers. Copyright William K. Hartmann

For more info:

Hartmann, W. K. and D. R. Davis 1975 Icarus, 24, 505.

Hartmann, W. K. 1997. A Brief History of the Moon. The Planetary Report. 17, 4-11.

Hartmann, W. K. and Ron Miller 1991. The History of Earth, (New York: Workman Publishing Co.)

Hartmann, W. K., R.J. Phillips, and G.J. Taylor, eds. 1986. Origin of the Moon. (Houston: Lunar and Planetary Institute.)

Back to PSI main page