原文链接

The Milky Way should be surrounded by mini galaxies. That it isn’t spells trouble for gravity as we know it, says Marcus Chown

LIKE moths about a flame, thousands of tiny satellite galaxies flutter about our Milky Way. For astronomers this is a dream scenario, fitting perfectly with the established models of how our galaxy's cosmic neighbourhood should be. Unfortunately, it's a dream in more ways than one and the reality could hardly be more different.

As far as we can tell, barely 25 straggly satellites loiter forlornly around the outskirts of the Milky Way. "We see only about 1 per cent of the predicted number of satellite galaxies," says Pavel Kroupa of the University of Bonn in Germany. "It is the cleanest case in which we can see there is something badly wrong with our standard picture of the origin of galaxies."

It isn't just the apparent dearth of galaxies that is causing consternation. At a conference earlier this year in the German town of Bad Honnef, Kroupa and his colleagues presented an analysis of the location and motion of the known satellite galaxies. They reported that most of those galaxies orbit the Milky Way in an unexpected manner and that, taken together, their results are at odds with mainstream cosmology. There is "only one way" to explain the results, says Kroupa: "Gravity has to be stronger than predicted by Newton."

Challenging Newton's description of gravity is controversial. But regardless of where the truth lies, the Milky Way's satellite galaxies have become the latest battleground between the proponents of dark matter and theories of modified gravity.

Our standard picture of the universe comes from many decades of observations. It asserts that visible matter - the kind of stuff that you, me, the planets and stars are made of - is outweighed by a factor of 6 or 7 by invisible, cold dark matter. No one knows what dark matter is made of, but its existence has been postulated to explain how the stars in spiral galaxies can orbit at such breakneck speeds without being flung off into the void. There isn't enough ordinary matter out there to hold on to everything, so the extra gravitational grip provided by large amounts of dark matter stops these speeding stars flying off into space.

Dark matter is also thought to have played a key role in shaping the early universe. In the aftermath of the big bang, it was the dark stuff that first began to clump together under the force of gravity because its lack of interaction with light meant it was not blasted apart by the big-bang fireball. Later on, normal gaseous matter fell into these clumps - dubbed dark matter haloes - where it congealed into stars to make visible galaxies.

A key feature of this dark matter scenario is that dark matter haloes of all sizes form. According to the standard model of cosmology, a halo as large as the one thought to have seeded the Milky Way should be surrounded by thousands of mini haloes, which themselves should have seeded small satellite galaxies.

So why don't we see them? It could simply be because most of the satellite galaxies contain only a few thousand stars and their faintness makes them extremely hard to spot (see New Scientist, 15 August, p 10).

Another problem is that it is not obvious to the human eye that an apparent group of stars in the sky is a bound collection rather than a chance alignment of stars at wildly different distances. Proving their connectedness requires computerised search techniques and detailed analyses of the colours of the stars to give their relative distances and types - a painstaking and expensive business.

Tidal dwarfs

Nevertheless, the rate of discovery of satellite galaxies has been boosted in the past five years by a detailed search by the Sloan Digital Sky Survey. Whereas only nine satellites were discovered in the 30 years before SDSS, another 15 have been found since. The biggest are about 1000 light years across - less than 1 per cent of the diameter of the Milky Way's disc - and the smallest about 150 light years across. Despite this progress, the total number of satellites known falls far short of that predicted by the cold dark matter paradigm.

The missing-satellites problem is not the only puzzle. Kroupa and his Bonn colleague Manuel Metz, together with Gerhard Hensler at the University of Vienna, Austria, and Helmut Jerjen of Mount Stromlo Observatory near Canberra, Australia, have studied the location and motion of the small number of known satellite galaxies. They found that a high proportion of the galaxies appear to be confined to a plane perpendicular to the disc of the Milky Way. What's more, most of the galaxies orbit the Milky Way in the same direction. "This is completely incompatible with the dark matter model of the Milky Way's formation," says Kroupa. He points out that the satellites should be more like a swarm of bees, moving on random orbits and distributed in a spherical shell around our galaxy.

If the origin of the Milky Way's satellite galaxies cannot be explained by the dark matter model, how did they originate? Kroupa says a clue can be found in a long trail of gaseous material and stars called the Magellanic Stream, which was torn free of the Large Magellanic Cloud through the effects of the Milky Way's gravity (The Astrophysical Journal, vol 697, p 269).

Such tidal effects were much more common 10 to 12 billion years ago when the Milky Way was born, because galaxies in the rapidly expanding universe were a lot closer together than they are today. Kroupa and his colleagues argue that the young Milky Way's gravity tore gas from a passing galaxy to form ancient "tidal dwarfs" which ended up as satellite galaxies. "Like the Magellanic Stream, such galaxies would naturally form a planar stream and share the same motion," says Kroupa.

It seems a neat solution. But the idea of the satellites being ancient tidal dwarfs raises another issue. Measurements of the velocities of the stars within the galaxies show that they are orbiting their galaxies very fast - so fast that by rights they should be flung off into intergalactic space.

This is precisely the problem that astronomers find in spiral galaxies and which they introduced dark matter to fix. "The problem is that the dark matter fix cannot be used in the case of tidal dwarf galaxies," says Kroupa. The reason is to do with the different way that ordinary matter and dark matter behave when galaxies interact or collide.

These differences are most apparent in a celestial object called the Bullet cluster, which formed when two galaxy clusters collided. Images taken in space by the Chandra X-ray Observatory reveal that when the clusters collided, the two vast clouds of gas slammed into each other and slowed down. But maps of the mass distribution suggest that the two clusters of dark matter sailed right through each other unaffected, leaving the ordinary matter languishing behind.

Kroupa reckons the dark matter and ordinary matter would have become separated in a similar way when the tidal dwarfs formed. This presents a conundrum: evidence from the breakneck speed of the stars in the satellite galaxies "screams dark matter", says Kroupa, "but all the other evidence says these galaxies cannot possibly contain dark matter".

“ All the evidence screams that the Milky Way’s satellite galaxies cannot possibly contain dark matter”

So how is it possible to explain the anomalously fast speeds of the stars within tidal dwarf galaxies? The only answer, says Kroupa, is to modify gravity. He favours an alternative to dark matter known as modified Newtonian dynamics, or MOND, devised in the early 1980s by Mordehai Milgrom, now at the Weizmann Institute in Rehovot, Israel. MOND has it that below a critical acceleration, gravity is stronger than Newton's law dictates. So because the stars sweeping along the outer edges of spiral galaxies experience lower acceleration than those of the inner galaxy, they are gripped a little more strongly than we would expect under Newton. With a straightforward formula, Milgrom can explain the motion of stars in every spiral galaxy for which we have velocity measurements.

MOND is a logical alternative to dark matter. However, it is difficult to find circumstances in which the two scenarios predict different outcomes. Now all that could change. Milgrom thinks that the failure of the dark matter model to predict the numbers, location and velocities of the Milky Way's satellite galaxies is a significant observation. "It is the cleanest situation where MOND succeeds and dark matter fails," he says.

James Binney at the University of Oxford begs to differ, however. In stark contrast with Milgrom, he claims that the satellite-galaxy problem bolsters the dark matter scenario. "This is actually the cleanest situation where dark matter succeeds," he says.

Dark galaxies

How can the proponents of MOND and dark matter have such diametrically opposite interpretations of the same observations?

According to Binney, you need to look at the details of the dark matter scenario for galaxy formation. In the aftermath of the big bang, quantum fluctuations in space-time led to some regions of the universe gaining lots of matter and other, void-like regions very little. The voids expanded faster than the dense regions, whose expansion was restrained by the gravity of the matter they contained.

As the voids spread out and connected with each other, they squeezed dark matter and ordinary matter into sheets and streams. "We see this in the distribution of galaxies," says Binney. The universe looks like "Swiss cheese" with concentrations of galaxies separated by enormous voids.

He sees this process of matter squeezing into sheets and streams as acting on the scale of the Milky Way too: dark matter would have streamed into the Milky Way along certain paths (New Scientist, 18 July, p 34). So Binney sees it as quite natural that we see satellite galaxies largely confined to a single plane and with their velocities correlated. "Their properties are perfectly explicable within the dark matter scenario," he says.

But if the dark matter model does tally with the locations and motion of the satellite galaxies, why do we see only about 1 per cent of the number we would expect? Binney sees no problem here either. He says the missing galaxies are simply too faint for us to have detected them yet. Or "they may be exclusively composed of dark matter" with not enough gas to light up stars, he adds.

Binney points to a recent study by a team led by Sergey Koposov of the Max Planck Institute for Astronomy in Heidelberg, Germany, which concluded that the satellite galaxies we see are just the tip of the iceberg. From the properties of the observed satellite galaxies, Koposov predicts that the number of ultra-faint galaxies yet to be discovered should run into the thousands (The Astrophysical Journal, vol 696, p 2179).

But it is not clear how galaxies with their vast concentrations of gas and dark matter can be starless. Suppressing star formation involves complex mechanisms that are poorly understood - this much everyone agrees. "It's the Achilles' heel of the dark matter model," Binney admits. "But that just means we've still got much to do to flesh out the model."

Milgrom and Kroupa are not persuaded. They maintain that the mechanism preventing the existence of stars is the fatal flaw in the dark matter model. They face an uphill struggle convincing others, however: the majority of astronomers are wedded to dark matter and will not throw more than 30 years of work out of the window lightly. The truth is, says Binney, that both dark matter and MOND are deficient in their own ways.

So what will it take for one side to give ground? The answer may lie in mapping the gravitational landscape on the outskirts of the Milky Way. By making ever more detailed maps of the motion of all the visible satellite galaxies and globular clusters, it should be possible to deduce the presence of all the satellite galaxies that are too faint to see. If it turns out that there are indeed thousands of ultra-faint satellites, as the dark matter model predicts, then the proponents of dark matter will have backed the right horse. If not, then dark matter may yet stumble before the finish.

Without such a gravity map, both sides are slugging it out with balloons on sticks rather than boxing gloves. For now, the Milky Way's environs remain a distant battleground between two great world views.■

Marcus Chown is the author of Quantum Theory Cannot Hurt You (Faber, 2008)

The next big fin East meets west Rise of the robogeeks Why do people die that way? Clever fools: Why a high IQ doesn't mean you're smart