Imagine a system of molten silicate material, where low-density minerals float and higher density minerals sink. Minerals rich in iron and magnesium (such as olivine and pyroxene) will settle toward the bottom of the magma body while those rich in the elements aluminum and calcium (such as plagioclase feldspar) will float. Just such a scenario – on a global basis – is thought to have created the crust of the Moon.
Before Apollo, many believed that the Moon was a primitive, undifferentiated lump of cosmic debris. By studying the samples returned by Apollo 11, scientists identified small fragments of white, plagioclase-rich rocks (anorthosite). There are no known magma compositions corresponding to this rock type – anorthosite is created by removing low-density plagioclase from a crystallizing system and concentrating it by floatation. From the evidence of fragments in the lunar soil, large amounts of anorthosite were inferred to be present in the nearby highlands of the Moon. As the highlands make up more than 85% of the surface of the Moon, it was postulated that the crust of the Moon formed early in its history by global melting, an episode termed the “magma ocean.”
Expecting only minor volcanic activity and perhaps a local igneous intrusion, the concept of a global ocean of magma was surprising to most scientists. Given its small size and consequent paucity of radioactive heat-producing elements, the idea that most of the Moon might have melted and differentiated was astounding. The existence of an early magma ocean, which implied high-energy processes, provided us with clues to lunar origin. Once it was recognized that the Moon had a crust, it was important to gain an understanding of its composition and physical nature.
On subsequent missions, Apollo astronauts were tasked with laying out a series of seismic stations across the near side. These stations allowed us to measure “moonquakes” – both natural events as well as those created artificially by slamming spent rocket stages and satellites into the Moon. Seismic recording allowed us to infer the speed at which seismic waves traveled through the lunar interior. These estimated speeds indicated densities that implied composition, allowing us to deduce the probable chemical and mineral composition of the lunar interior.
The Apollo seismic network indicated that the crust of the Moon was about 50-60 km thick in the central near side, a surprisingly large value, especially compared to the thickness of the crust of the Earth (which varies from as thin as 5-10 km under the ocean basins to over 30 km in continental areas). Such a thick crust for the Moon led to the postulation of a global magma ocean, as so much anorthosite could only be produced under the conditions of near global melting. Subsequent studies incorporating gravity data from Lunar Orbiter and other missions suggested that the lunar crust is variable in thickness, with values exceeding 100 km in some regions of the far side highlands.
Re-analysis of the Apollo seismic data gave the first indication that those values might be overestimated. Using modern techniques on these old data, new analysis revealed that the crust might be thinner than we had originally thought, on the order of 40-50 km thick. This lower value of crustal thickness had some implications for estimating the bulk chemical composition of the Moon, but because it was considered to be a relatively minor adjustment, it caused no major difficulties for the rest of lunar science.
However, the recent GRAIL mission to the Moon (using high precision gravity mapping) ascertained the thickness of its crust to be 34-43 km. Why should this new value worry some scientists? Because we are now entering realms in which the new estimates of crustal thickness create consistency problems for other aspects of lunar science. A crust as thin as 35 km on the near side of the Moon implies that the largest impacts – the multi-ring basins – should have excavated considerable amounts of material from the layer below the crust, the mantle of the Moon. One might object that, as this region of the interior is inaccessible, we don’t know what the mantle would look like. But in fact, the density constraints imposed by the seismic and gravity data dictate that it must be a rock type rich in iron and magnesium, made up mostly of the minerals olivine and pyroxene. Such rocks are not unknown in the lunar collections, but they possess chemical and mineralogical characteristics indicating their origins at much shallower (crustal) depths. In other words, there does not appear to be any material from the lunar mantle in the Apollo collections. Given our obviously incomplete sampling of the Moon, should this be a problem?
Several Apollo landing sites (e.g., Apollo 14 and 15) were specifically chosen to maximize the chances of sampling ejecta from the enormous 1100 km diameter Imbrium basin (one of the biggest impact features on the Moon). Virtually any reconstruction of the dimensions of the excavation cavity of this basin indicates that it should have dug up material from tens of kilometers depth, much deeper than the new value of crustal thickness implied by the GRAIL data. So where is this debris from the mantle of the Moon? True enough, it is possible that it may have been missed during the limited exploration time available to the Apollo crews, but the astronauts were trained to recognize such rocks and none were found. Additionally, because we can map rock types by remote sensing (both from spacecraft and from Earth), we have an understanding of the regional distribution of rocks around these large impact features. Despite a 30-year, exhaustively detailed search of the Imbrium impact basin (an area larger than Texas), we have found no convincing evidence for mantle material on the surface of the Moon.
So where does this leave us? In science, new data can solve some problems but at the same time, it may also create new ones. Modern analyses of the old seismic data and new information on the Moon’s gravity field both suggest a relatively thin crust, with mantle material being very close to the surface (a few km) in some areas. On the other hand, none of the ubiquitous impact basins and large craters of the Moon show evidence for mantle material in their ejecta, either in the Apollo collections or in remote sensing data. Could our understanding of impact mechanics be completely wrong? Or are we misunderstanding the new gravity data? How could an event that formed an impact crater thousands of kilometers across excavate only a few kilometers deep?