The "Water on the Moon" Hoopla, Part 2: The murkier part of the story

In Part 1 of my writeup of yesterday's "Water on the Moon" press briefing I explained how scientists using the Moon Mineralogy Mapper (M³ or "M-cubed") on India's first lunar orbiter, Chandrayaan-1, found the first evidence for widespread water on the surface of the Moon, and corroborated their discovery with data from spectrometers on Cassini and Deep Impact. For the first time, we know there's water on the Moon. That's pretty conclusive. But how much water is there, and is it in a form that human explorers could use? This part of the story has many more questions and many fewer definite conclusions.

First, let's back up a bit, all the way to the days of Apollo and Luna. The Moon is the only body in the solar system from which we have returned substantial samples under somewhat controlled conditions. And there's a lot of lunar samples, almost 400 kilograms' worth. These samples have been studied to death, and one thing that we have learned about the Moon from the samples is that they are surpassingly dry, lacking any abundance of water above what might arise from their exposure to Earth. Even a vacuum chamber on Earth contains a few water molecules, and water molecules are very, very sticky, so it's very difficult to determine what's Earth contamination and what's original water from the Moon. Jeff Taylor has a very nice summary of the analysis of water in lunar rocks here (which also deals with an interesting recent paper that hints that there may actually be more water in the lunar mantle than previously thought -- but that's a story for another day).

At yesterday's press briefing, the director of NASA's planetary science division, Jim Green, caused the gathered photographers to collectively twitch their trigger fingers when he held up a tablespoon, saying that all the water in all the lunar samples ever brought to Earth wouldn't fill that tablespoon. So: the Moon -- or at least the surface, where astronauts can grab samples and spectrometers can see spectral features -- is dry.

That conclusion has not changed since yesterday. The Moon is still dry, drier than the driest desert on Earth. What has changed is that we now have evidence to suggest that the Moon is not completely devoid of water. Absorption features at a wavelength between 2.5 and 3 microns clearly indicate that some small amount of water is present. How much?

Before we can figure out how much, we have to figure out what form it takes, and I'm going to spoil the punch line by telling you that no one knows yet what form the water takes. As M³ principal investigator Carlé Pieters said in yesterday's briefing: "There is a variety of hypotheses for what physical form this water could take on the surface. What we're seeing occurs in the uppermost surface of the lunar soil, the upper two millimeters. It could occur as one monolayer a few molecules thick; it could be mixed into the surface; it could be altered minerals; there could be various gradients."

A monolayer would be reminiscent of the adsorbed water that the Phoenix team found at their landing site on Mars. Altered minerals would be quite another thing -- it would be hydroxyl groups bonded within the chemical structure of the minerals that make up the rocks. M³ team member Roger Clark said, "The hydroxyl is an OH bond that creates a chemical reaction with other minerals in the surface, which creates hydroxyl-bearing minerals. We haven't identified the minerals yet, but clay minerals are an example."

Molecular water and the hydroxyl ion do have slightly different spectral signatures. If the spectral feature were only due to hydroxyl, it would bottom out at 2.8 microns. The spectral signature from molecular water would bottom out at a longer wavelength. Of course, nothing in nature is ever pure, and the complicated shape of the absorption feature seen in the spectral data from Chandrayaan-1, Cassini, and Deep Impact all suggest that there is both molecular water and hydroxyl ions present. And, Carlé said, "One thing we do not know is the relative proportion of water and hydroxyl."

But you can make assumptions and do back-of-the-envelope calculations. All three of the papers presented yesterday contained such calculations, and it seems safe to say that the data are consistent with water abundances ranging from 10 to 1000 parts per million. This is not a lot. The very highest number corresponds to one liter of water in one ton of lunar rock. And it may be water that is bonded chemically to the minerals in that lunar rock, which might be tough to liberate.

And I'm sad to say the story gets even more complicated than that: the amounts of water and hydroxyl observed on the Moon may even vary with time. This wrinkle arose from the Deep Impact data, presented yesterday by Jessica Sunshine. She explained that Deep Impact just happened to take data on the Moon twice in June of this year -- once on June 2, once on June 9. That one-week separation between the two observations corresponds to one quarter of a lunar day. Parts of the lunar surface that were just going into morning sunlight on June 2 were at solar noon on June 9. And areas that were under high noon illumination on June 2 were seeing the sun set on June 9. Jessica found that the strength of the water absorption feature was strongly dependent upon the time of solar day:

If this observation is real, it implies that there is more water in the lunar surface at sunrise and sunset than at noon -- that water moves around the surface on very short timescales, the time scale of the lunar day. It may even be destroyed (photodissociated) under the noonday Sun, and created anew at dawn and dusk.

However, there was disagreement among the scientists on yesterday's panel about whether this observation of a diurnal variation was real. Roger Clark -- the one presenting the Cassini VIMS data -- wrote the following in his paper: "Both VIMS and Deep Impact data indicate stronger absorption near the lunar terminator and [the Deep Impact team] attribute this as evidence for movement of water with the diurnal cycle. However, viewing geometry might account for some or all of this apparent variability."

The effects of viewing geometry can be a difficult thing to disentangle from the observations. The problem is that the noontime data comes (by definition) from parts of the Moon lit from directly overhead by the Sun; terminator data comes (by definition) from parts of the Moon lit very obliquely by the Sun. Clark is saying that the change in the strength of the absorption band can be explained away by the change in illumination angle. I asked Jessica Sunshine about this in an email and she said: "It is the nature of science for there to be disagreements. However, we did address this in our paper. It is possible that some, but not most, of the change in absorption strength could be due to geometric effects. However, there are very significant changes in the shape (width) of absorptions with temperature that simply can not be attributed to geometry or photometric effects."

The key thing, Jessica is saying, is the shape of those absorptions. Look at the graph above again. The red plot, the noon data, has a sharp minimum at around 2.8 to 2.9 microns, then a "shoulder" going up to 3.6 microns. The blue plot, the evening data, lacks that shoulder. I'm no spectroscopist, but I could be convinced that both plots contain a 2.8 micron signal from the hydroxyl absorption (which would come from water bonded within minerals in rocks, and would therefore not be expected to be mobile on lunar-day time scales), while the evening plot contains a much stronger absorption feature around 3 to 3.2 microns (which you might argue came from the water). Jessica said this change in shape "shows there must be two species." For a final reality check, I tossed this at yet another lunar spectroscopist whose name wasn't on any of these three papers, and was told that this was a very reasonable argument, especially given the change in shape that Jessica showed, but it was not completely convincing. More data is needed. (Heh. More data is always needed.)

So let's assume, for the moment, that the observation of diurnal variation is real. What does that mean? It means that there must be some process that causes the lunar surface to lose water during the day, and another one that replenishes it at night. It's not hard to imagine how the Moon loses water during the day; it's hot, and there's no atmosphere. Any water in the soil that was not forcefully chemically bonded would naturally go off as a vapor. Once vaporized, it might condense elsewhere, or it might just as easily be photodissociated from the hard light of solar radiation, turning it into protons and hydroxyl ions that go their merry separate ways, reacting with lunar minerals and weathering them. But how would you get that water back in the evening?

There's two principal ways in which the Moon can get more water. One is easy to explain: comets bring it in. Comet impacts aren't very common but the Moon is old, and as Roger Clark points out in his paper, there's probably been enough water delivered to the Moon from comets over the last 2 billion years to cover the whole globe to a depth of half a millimeter. But this is not a process that would replenish lunar water over the course of a lunar day.

Jessica advanced a different explanation for how you get water on the Moon: she suggested that water might be continuously generated on the lunar surface when solar wind protons bombard mineral grains. Lunar minerals contain lots of oxygen; it's one of the most abundant elements on the Moon. Solar wind protons can make water by reacting with those minerals. This is not at all a new idea; it's been around for a very long time.

Which isn't to say it's a process that's well understood. Carlé said: "We have to understand the physics of this silicate surface and the vacuum around it, which is awash in solar wind particles and micrometeorites. The physics is just in its infancy." And Jessica said: "There's a lot of unknowns that we need to work out." And finally Rob Green, who's the project instrument scientist for M³, said "There are many more questions today than we had six months ago." Which I think is an amusing encapsulation of the paradoxical nature of scientific "advancement" -- every time we learn something trying to answer one question, what we learn makes us ask ten more questions!

But if there's a process operating on the airless Moon, producing detectable amounts of water every lunar day, that's important, because, as Jessica said, "We should see the same effects on any oxygen-rich body with no atmosphere. This includes Mercury as well as asteroids."

Well, cooooool.

So there could be water on airless bodies all over the solar system. Finding water in space is desperately important for human exploration, because water is so danged heavy and we need so much of it that launching enough for long-lasting human habitation of any place, whether it be the Moon or an asteroid or what have you, quickly becomes prohibitively expensive. If you're talking about setting up permanent human occupation anywhere off of Earth, you have to have a way to create or extract water in space. So does this amount of water discovered on the Moon represent enough to support permanent human habitation?

Maybe. I wish I could be more definite, after all this discussion. But it's not very much water, and it would not be easy to extract. On the other hand, it's there, and that's something that we didn't know about before yesterday. Water is there in the lunar rocks. More importantly, the data from VIMS and Deep Impact seems to suggest that it's not just at the poles, it's also at lower latitudes. So if it becomes a priority for us of Earth to establish a permanent human colony on the Moon, we at least know now that there is water there to be extracted, if we choose to learn how to do so, and that we wouldn't necessarily have to build our lunar base next to a permanently shadowed region at one of the poles. That's a big step.

Speaking of permanently shadowed regions near the poles, that's going to be the next piece of this story. In less than two weeks, the LCROSS mission is due to crash into a permanently shadowed region of crater Cabeus A, attempting to answer once and for all whether those permanently shadowed areas have acted as permanent traps for water that has moved around the lunar surface. Like Deep Impact, LCROSS is sending in a heavy impactor and watching from behind, broadcasting its measurements all the way down. Unlike Deep Impact, LCROSS is going to follow the same fate as its impactor. But the double impact is going to be watched by another spacecraft, Lunar Reconnaissance Orbiter, and also by numerous space- and Earth-based observatories. Will they see water? Who knows? I hope they see lots of it. If they don't, we won't know if it's because those permanently shadowed regions aren't water traps after all, or if the chosen target just happened to be a dry spot for some reason.

Fortunately, Lunar Reconnaissance Orbiter carries a wholly different kind of spectrometer that studies the Moon in ultraviolet rather than infrared wavelengths. This spectrometer relies on ultraviolet light sources all over the galaxy for illuminating the Moon, so can "see" into those "permanently shadowed" regions to look for the Lyman-alpha absorption feature characteristic of hydrogen in frost on the lunar surface. If water is created on the lunar surface during the day, and is mobilized by the Sun, some of it should wind up coming to ground in a cold polar crater and staying there, where the Lyman-Alpha Mapping Project will see it. We're still looking for more of that lunar water; you haven't heard the end of this story.