The US space agency Nasa sells dreams. There’s nothing wrong with that – after all, space and its exploration have always been a source of reverie, from Johannes Kepler’s youthful space-travel fantasy simply called The Dream to visions of the ‘final frontier’. The problem with dreams is that sooner or later you must wake up.

To judge from an article on lunar bases on Nasa’s web site, it’s reluctant to do that. “When multiple spacecraft all found unequivocal evidence for water on the moon it was a boon to possible future lunar bases, acting as a potential source of drinking water and fuel,” the article says. It explains that the atomic components of water – hydrogen and oxygen – on the lunar surface move towards the poles, “where [water] accumulates in the cold traps of the permanently shadowed regions.” Since it was first proposed several years ago, this idea that the polar craters, particularly the so-called Shackleton crater at the south pole, are lined with ancient ice has inspired many hyperbolic newspaper stories about colonising the Moon. But it’s looking ever less likely that it is true.

A new paper in the journal Geophysical Research Letters drives another nail into the coffin of lunar living. It suggests that what was at first taken to be bright, reflective ice in the Shackleton crater is in fact more likely to be white rock.

When the Apollo missions reached the Moon at the end of the 1960s, they brought back a sobering message: it seemed to be a dry, barren dustball. But the modern dream of “water on the Moon” began in earnest in 1994, when Nasa’s spacecraft Clementine orbited the Moon and studied the mineral composition of its surface. The reflections of radio waves beamed into the shadowed polar craters suggested that they might contain ice. But follow-up studies using radio telescopes on Earth failed to find any such evidence.

Then in 1998 another Nasa Moon mission, the Lunar Prospector spacecraft, used a special instrument to search for hydrogen atoms – a possible signature of water molecules – on the Moon’s surface. It detected the hydrogen signals from polar craters, but when at the end of its mission the spacecraft was purposely crashed into a south polar crater in the hope that it might send up a plume of water detectable from Earth, nothing of the sort was observed.

No Moon river

Each alleged sighting of lunar ice provoked new headlines forecasting future moon bases, feeding an apparent public thirst for space colonization. But for scientists, the debate has remained unresolved. In 2009 NASA launched the Lunar Reconnaissance Orbiter, designed to map the Moon’s surface in even more detail and carrying several instruments that might be able to detect ice. Last year a team of planetary scientists reported that the south polar Shackleton crater has a bright floor and even brighter inside walls, suggesting that some material has gradually slipped down the slopes onto the bottom of the crater. The researchers suspected that this stuff could be simply lunar “soil”, called regolith because it is really just mineral dust, with no organic matter. Lunar regolith is bright and reflective when freshly exposed – the bombardment from cosmic rays, solar wind and meteorites gradually darkens it, but on the crater’s walls it is particularly well sheltered from such disturbances. But the team also offered the tentative possibility that the bright material could be a very thin layer of rock dust mixed with 20% ice.

Now along comes a team of Japanese space scientists to squash that enticing idea. Led by Junichi Haruyama of the Japan Aerospace Space Exploration Agency (Jaxa), based in Kanagawa, they have analysed data from Jaxa’s lunar orbiter Selene – better known in Japan as Kaguya after a legendary moon princess. Last year the team reported that Selene (which operated from 2007 to 2009) had found rocks made of the mineral anorthosite all over the Moon. This stuff is thought to be formed when meteorites hit the Moon and melt its surface, and the researchers suggested that the Moon might have a thick layer of it several kilometres beneath its surface, created by a massive impact soon after it was formed.

Lunar anorthosite is very pure and bright white, as shown by the lumps of it brought back by the Apollo missions. But here’s the clincher: unlike ice, anorthosite absorbs infrared radiation strongly at a wavelength of 1.25 micrometres, providing a distinctive signature of this mineral. And that absorption was just what was seen by Selene on the inner wall of the Shackleton crater. So it looks as though it isn’t ice.

Haruyama and colleagues don’t rule out the existence of water elsewhere on the Moon, for example hidden away in sub-surface caverns. But they suspect that the amounts might be small. That may still be scientifically interesting, raising questions about how it got there and how it might move around on the surface. Yet without a significant amount of water on the Moon, it is hard to see how any substantial space colony could be established there – the cost of sending up regular water supplies (which would be used not just for drinking but for making hydrogen as fuel) just doesn’t look viable.

That would be a shame, because there’s surely useful science that could be done from a moon base, not least in terms of finding out how this ball of rock formed in the first place (we still don’t really know). It might also literally bring us back down to Earth, forcing us to accept that the universe is a truly inhospitable place, so that we’d better take care to keep our house in order.

Given the history so far, it would be unwise to imagine that this is the last we will hear of water on the Moon. But it would be unwiser still to start planning to build a colony up there.