For the last 16 years, researchers at the DAMA/LIBRA experiment in Italy have seen a controversial annual oscillation in the signal from their dark-matter detector. This type of variation would be seen if the Milky Way galaxy was wreathed in a "halo" of dark matter. But apart from the CoGENT dark-matter experiment in the US, no other dark-matter searches have seen a similar effect. Now, a physicist at Durham University in the UK has proposed an alternative source for the modulation in the form of neutrons, which are knocked out of atoms by muons and neutrinos scattering in the rock or shielding material around DAMA/LIBRA.

The most recent cosmic microwave background (CMB) data from the Planck mission reveal a universe that is composed of 26.8% dark matter and 68.3% dark energy, with less than 5% of "normal" visible matter, such as galaxies and gas clouds. Dark matter is thought to interact weakly with ordinary matter, making dark-matter particles extremely difficult to detect. The DAMA detector – located deep underground at the Gran Sasso National Laboratory – reported the first signs of dark matter in 1998, and further data over the years has cemented the result at a statistical significance at 9.3σ – well beyond the 5σ that usually signifies a discovery in particle physics.

Summer high

The modulation, which peaks in the month of May, is thought to come about as the solar system sweeps through a dark-matter halo enveloping the Milky Way. The peak occurs in the northern-hemisphere summer because the tangential velocity of the Earth as it orbits the Sun is in the same direction as the motion of the solar system at that time of the year. This means that the number of collisions detected by DAMA should be at a maximum in summer and then drop in the winter.

Now, however, Jonathan Davis of Durham University has developed a new model to explain the DAMA/LIBRA signal without invoking dark matter. Rather, he shows that neutrons scattering in the detector could easily produce an annual signal. The neutrons are released when solar neutrinos and atmospheric muons scatter in the shielding material or the rock that envelops the DAMA/LIBRA set-up.

Scattered signal

The muons come from cosmic rays decaying in the atmosphere and their rate varies across the year, peaking around 21 June. The solar neutrinos' rate also varies annually, but instead they peak around 4 January. "When combined, this means that the neutrons from both of these sources also have a rate that varies annually but peaks somewhere in between the two and can match the DAMA phase that is in late May," explains Davis. "There is an annual peak because the interference between the muons and neutrinos is not perfect. So, they don't exactly cancel...the idea is that because both of the constituent signals peak at different times, when they add up there is some cancellation but this is not total," he says, further explaining that it is the remnant signal after the cancellation that peaks around late May, just like the DAMA data.

Muon mimic

While the idea of muons mimicking the DAMA signal is not new, the timing of muons in isolation does not match the DAMA data and the idea was dismissed. But Davis's model solves this problem by adding the effect of solar neutrinos. Davis told that it is currently unclear how important a role the lead shielding surrounding the experiment plays in neutron production; however, it is likely to be significant. Lead is particularly good at producing neutrons from neutrinos and muons. "Indeed, the cross-section – which gives you the interaction rate – for neutron production from neutrinos and muons is particularly high. Also the lead shielding is very close to the DAMA detector," he says. He also points out that neutrons produced in this way have a spectrum that tends to peak at low energy, similar to what one would expect from dark matter, meaning that "the signals can be easily confused".

Davis acknowledges the fact that neutrinos – often referred to as the ghosts of matter – are known particularly for their ability to not interact with matter as they pass through it. But he says that as the DAMA detector is particularly sensitive to low-energy recoils, it will pick up the neutrons produced by these neutrinos. "Also, most other experiments, such as LUX, have more advanced shielding than DAMA, so they would be able to stop the neutrons before they get to the detector," he says. He also states that other neutrino experiments do see the modulation he considers – he points to papers from the Borexino and SuperKamiokande experiments, which measure the modulation caused by neutrinos precisely. "However, these are directly down to neutrino scattering, not neutrons from neutrinos. The phase should be the same though," he says.

Other experiments?

When it comes to the CoGeNT experiment, which also sees the same type of early modulation, Davis is intrigued. "In principle, the model would be the same, however, since CoGeNT is in a different lab to DAMA, the phase of the signal would be different. CoGeNT has had a lot of trouble recently with surface event backgrounds, so we will have to wait and see as it is not clear what it is seeing," he cautions. Other dark-matter experiments, he says, have not seen the signal, probably thanks to a combination of shielding and thresholds. Because most of the more recent experiments employ more effective neutron shields than DAMA, the neutrons, which make up the DAMA signal, would not reach detectors such as CDMS, LUX or XENON100. Also, DAMA is more sensitive to low-energy recoils than most experiments, and so might be more susceptible to the muon/neutrino signal than other experiments.

To check for the accuracy of Davis's model, the DAMA/LIBRA collaboration could study in more detail how the phase of its signal changes with the energy of the events. According to Davis, this has been studied before, and DAMA found that the phase does change with energy – something that you would not expect from standard dark matter but that is explained by his model. Also, with the increasing number of data that DAMA will collect in the coming years, the researchers will be able to look for "an additional mode with a period of 11 years, which would be expected if the signal is due to muons (it comes from solar activity), but not for dark matter" Davis says.

Davis is also keen to emphasize the importance of future dark-matter experiments – such as DM-Ice, KIMS, and ANAIS – which are looking to replicate DAMA. "My model gives them something they can test as a comparison with dark matter," he says. "This is especially interesting for DM-Ice because it will be in Antarctica, so the muons will have the opposite phase."

A preprint of the research is available on the arXiv server.

UPDATE: The paper has now been published in Physical Review Letters.