Wednesday, February 14, 2007
Dark matter: what we know
I've been thinking a lot recently about the cosmic dark matter problem. It really is vexing.
First, what do the data say:
It was discovered that we can measure how galaxies rotate, by looking at the Doppler shift of atomic emission lines. It's the same basic principle as pointing a Doppler radar at a tornado or hurricane, although what we see is the y-z plane instead of the x-y plane. Great precision is possible because an atomic emission line covers a very narrow frequency range, and we know how to precisely separate different frequencies of light. So we made all these cute hook-echo-like pictures of galaxies, showing how fast the various parts of the galaxy are orbiting its center. (For each pixel we actually get a distribution of velocities, because we see through the whole thickness of the galaxy. The peak/valley of the distribution is from the stuff we see moving head-on and is the rotation speed for that distance from the center.)
And since the stars in a galaxy orbit according to gravity, and gravity depends on the mass enclosed by the orbit, this tells us how a galaxy's mass is distributed.
We can also tell how mass is distributed by the amount of light that shines out. We have to make guesses about how bright the average star is, how far away the galaxy is, and how much dust there is blocking the light, but the predictions turn out to be constrained to a pretty narrow range as cosmological things go. And even if we get it wrong for a particular galaxy, many of the errors will average out to zero when we measure a large number of galaxies.Both types of measurement were made. They turned out shockingly different. It varies for different galaxies, but on average the rotations show about four times the mass that the brightnesses show. Nor can it be an evenly-distributed error, because although many galaxies have been observed with faster rotation than predicted by brightness, not a single one has ever been observed with slower rotation.
Because the missing gravity source does not glow, it was dubbed dark matter. (However it need not be ordinary everyday matter, so don't take the word "matter" too literally.)
The Doppler rotations were confirmed by radio emission lines from hydrogen gas clouds, so it isn't some sort of problem with UV/vis spectrometry.
In fact, an entire galaxy has been discovered   that has only the radio "glow" of a handful of hydrogen gas, no apparent visible stars at all, and the usual amount of dark matter expected for a typical galaxy. The dark matter to normal matter ratio is at least 100:1. If this discovery pans out, the deficit of ordinary matter will be impossible to explain away as a measurement error. Even if the entire galaxy were hidden by dust, the dust would be visible by dint of covering up distant background galaxies.
Other evidence for dark matter comes from the hydrogen clouds in large galaxy clusters. We can measure their temperature from their x-ray emissions, and temperature tells how fast the atoms are moving. Because the clouds have not evaporated, there must be enough gravity to confine them given the atoms' thermal motion. And once again, the visible matter in the clusters does not have enough gravity to explain the confinement.
The latest, and most interesting, evidence of dark matter is from the Bullet cluster of galaxies, a pair of galaxy clusters that have collided. The stars are teeny tiny compared to the empty space between them, and so they mostly just pass by each other. Conversely, the gas/dust clouds are fluffy and diffuse, and so they collide strongly as the galaxies pass through, practically plowing to a stop and forming a new combined cloud (which glows x-ray hot from all the kinetic energy that has been thermalized).
The cool thing about the Bullet cluster is that the non-cloud parts can be directly weighed by optical means. They form a gravitational lens that bends light, distorting the images of the distant background galaxies. The lensing measurements show the usual amount of dark matter, located right there with the visible stars. As usual, the lens mass is much greater than the visible mass of the stars. Rather than plowing to a halt alongside the gas clouds, one set of dark matter passed right through the other set.
However, other measurements have looked for gravitational lensing inside our own galaxy. If there are dense, star-sized clumps of matter in our galaxy, they will occasionally pass between us and a distant star, causing lensing for a few hours or days. Individually such alignments are rare, but with billions of objects zipping around, the odds are actually good we will see dozens a year. These measurements have been made and do not show nearly enough "microlensing" to account for the missing mass. So dark matter is not organized as clumps.
What else do we know about dark matter:
- It doesn't participate in the electromagnetic force. Otherwise light would bounce off and space would be pretty foggy.
- It doesn't intereact via the weak force much, if at all. Otherwise radioactive decays would be much faster or much slower.
- It doesn't interact via the strong nuclear force much, if at all. Otherwise it would do obvious things to ordinary baryonic (made of protons and neutrons) matter, and spectacular things to neutron stars.
- The relationship between its inertial and gravitational masses are not clearly shown by the data.