Parsing particle experiments

The massive Via Lactea simulation may provide clues to a puzzle that’s plagued particle and theoretical physicists for years.

It has to do with DAMA, an experiment buried beneath an Italian mountain that’s designed to detect dark matter. Instruments there watch for signs indicating dark matter particles – believed to be WIMPs, for weakly interacting massive particles – have collided with the large atomic nuclei in thallium-doped sodium iodide crystals. Physicists predict these rare collisions – and the detectable “nuclear recoils” they produce – should occur more frequently when it’s summer in the northern hemisphere.

“The sun itself is moving through dark matter in the galactic halo. When Earth is moving in the same direction as the sun” – summer in the northern hemisphere – “you would get faster particles than if the opposite,” says Michael Kuhlen, a postdoctoral researcher at the University of California, Berkeley, and a member of the Via Lactea group. That should produce a seasonal modulation that rises above background noise.

That’s exactly what DAMA researchers assert. “They have been saying for years: ‘We have detected dark matter. We have seen the presence of dark matter particle scattering.’  The problem is other experiments that have a different approach have not been able to confirm this. People don’t really understand what’s going on.”

At least a couple of theories attempt to explain the inconsistency. One postulates that dark matter is inelastic: When it collides with atomic nuclei, it produces a new particle with more mass and energy. Another predicts that dark matter particles are less massive than expected.

In either case, Kuhlen says, predictions for how many recoils an experiment finds should be sensitive to the velocity of dark matter particles, because a higher minimum velocity is required to produce a given recoil energy.

So he and his Via Lactea colleagues looked at what the model and another simulation, the University of Zurich’s GHALO, had to say about velocity distribution. Physicist Neal Weiner of New York University, one of those who put forward the inelastic dark matter theory, also collaborated.

The paper, in the Feb. 23, 2010, Journal of Cosmology and Astroparticle Physics, considered the Maxwell-Boltzmann (MB) distribution, a statistical representation of the amount of energy apportioned among an identical set of particles. An MB distribution is smooth, with fewer particles at the high and low ends of the velocity scale.

But the Via Lactea II model found dark matter velocity distribution is anything but smooth. Just as the model predicts dark matter lumps and clumps spread around a galaxy, it also predicts matter velocity spikes and lumps, with a “noticeable excess” of high-velocity particles, the paper says.

They also found that the velocity substructures associated with subhalos and tidal streams can affect the expected recoil event rates, the energy at which such events occur and the direction of scattering dark matter particles.

Most dark matter scattering predictions rely on smooth, MB velocity distributions. The researchers’ data indicate “that may be a bad assumption to make if you are considering an inelastic or light dark matter model or if your detector is particularly sensitive to high-velocity dark matter,” Kuhlen says.

Researchers “put limits on the properties of dark matter. In many cases those limits are too stringent because they did not account for velocity structure and it could affect their answer,” Kuhlen explains. “If they ever do detect an event with significance it’s possible they would be misled as to the properties of the particle that led to that event if you do not account for this.”

The authors have posted their velocity distribution data on the Internet for others to consider in their calculations.

At bottom, the group’s findings show “it’s not as simple as you thought,” Kuhlen says. “The universe is more messy, and this is an additional source of uncertainty you need to take into account” – as if there weren’t enough already.