Skip to main content
Gravity

Gravity

A light in the dark?

01 Aug 2008

For the last 10 years physicists in Italy have been claiming to have directly detected dark matter, which is believed to make up 23% of the universe. Edwin Cartlidge finds out why their results continue to create controversy

A light in the dark?

If you can bear to hear the truth you’ve spoken
Twisted by knaves to make a trap for fools,
…you’ll be a Man my son!

It is rare that poetry enters modern scientific discourse, which is why the above quote from the poem “If…” by Rudyard Kipling catches the attention of the visitor to the homepage of DAMA, a group of dark-matter hunters in Italy. Although the physicists in the group do not spell out explicitly what the poem refers to, its meaning seems fairly clear. The quote is an eloquent retort to their critics (the “knaves”), who for the last 10 years have been unconvinced of the researchers’ claim to have made the first-ever direct observation of dark-matter particles (“the truth”). The poem illustrates the unusual intensity of the debate surrounding their claim.

DAMA is a collaboration of physicists from the University of Rome “Tor Vergata”, the University of Rome “La Sapienza”, the Italian National Institute for Nuclear Physics and the Institute of High Energy Physics in Beijing, China. Led by Rita Bernabei of Tor Vergata, the group operates a sodium-iodide detector buried 1400 m beneath the Gran Sasso mountain in central Italy. It says that it has observed a seasonal variation in the signal from the detector caused by dark-matter particles interacting with the sodium iodide. If that is true, the researchers will have hit the cosmology and particle-physics jackpot.

The DAMA signal is a sinusoidal modulation that peaks in early June and goes through a minimum in early December — a signature that the group interprets as being caused by the Earth’s motion through the “halo” of dark matter that surrounds the Milky Way. The rationale is that the Earth travels through the halo (which is assumed to be static) as the Sun orbits around the galactic centre with a velocity of about 230 km s–1. Since our planet travels “with” the Sun in June and “against” the Sun in December, its own velocity relative to the dark-matter halo will be at a maximum and minimum at these respective times. This variation should then manifest itself as an annual variation in the rate of dark-matter particles passing through the DAMA detector — a dark-matter wind. At least, this is how the DAMA researchers interpret the results.

The group originally made its claim in 1998, with an experiment called DAMA/NaI that used 100 kg of detecting material. It repeated its claim in 2000 and 2003 after having collected more data. Then at a conference in Venice in April this year, the group announced four-years worth of results from its upgraded DAMA/LIBRA (Large Sodium Iodide Bulk for Rare Processes) experiment, which added an extra 150 kg of sodium iodide to the apparatus. The group maintains that, using the combined data from the two experiments, “the presence of dark-matter particles in the galactic halo is supported at a confidence level of 8.2 standard deviations”, which equates to a chance of less than one in 4 ×  1015 that the result is a statistical fluke (arXiv:0804.2741; accepted for publication in European Physical Journal C).

That the DAMA group has seen an annual modulation in its signal is now widely acknowledged by other researchers in the field, who had previously doubted that the group’s original experiment had collected enough data to show this. However, the DAMA researchers remain at odds with others as to what is causing the modulation.

That disagreement is best articulated by Juan Collar from the University of Chicago, who is a spokesman for the dark-matter experiment COUPP (Chicagoland Observatory for Underground Particle Physics) at Fermilab in the US. Blogging for Cosmic Variance, he wrote that “There is evidence for a modulation in the data at 8.2 sigma, stop. Compatible with what would be expected from some dark-matter particles in some galactic halo models, full stop. Anything beyond this is wanting to believe, and it smears on the rest of us in the field. Of course…there is no other observed process in nature that peaks in the summer and goes through a low in the winter, so this must be dark-matter, right? (Occam is turning in his grave, rusty razor still in hand…).”

In response to such criticisms, Bernabei remains adamant of her group’s claim. “I don’t believe logically our results can mean anything other than the dark-matter signature,” she says. “We’ve been looking at this for over a decade, and in that time no-one has come up with an alternative explanation.”

Dark-matter hunting

In the current cosmological paradigm, the existence of dark matter is inferred indirectly from its gravitational effects. The particles do not emit electromagnetic radiation (hence, “dark”), and, for all we know, they might not have any other means of interaction. However, some theories outside the Standard Model of particle physics predict a number of possible candidates for dark matter, such as axions, sterile neutrinos or WIMPs (weakly interacting massive particles), that can interact via the weak nuclear force. Most ongoing dark-matter experiments are in fact tuned to find direct evidence for WIMPs, as these are the most popular candidates among theorists.

According to such theories, a standard dark-matter detector will never observe more than a handful of WIMP collisions every year, even though trillions of dark-matter particles are believed to stream through the Earth every second. To improve their odds, physicists must therefore design dark-matter detectors with as large a detecting mass as possible. They also need to shield it from interference as best they can, so as not to swamp the tiny signals that are collected. The detector material itself, the surrounding casing and the rest of the lab must all have extremely low levels of radioactivity. Furthermore, the lab should be located deep underground so that interference from cosmic rays is kept to an absolute minimum (a kilometre of overlying rock reduces the cosmic-ray flux by about a factor of a million), and the detectors must be placed inside a shield made of lead or other suitable materials (see “Inside DAMA/LIBRA”).

In building these experiments, however, there is a trade off in the amount of data that can be collected and the degree to which background radiation can be filtered out. Most dark-matter experiments concentrate on enforcing the latter, and these include CDMS in the Soudan mine in Minnesota, EDELWEISS in the Modane Underground Laboratory under the Franco- Italian Alps, ZEPLIN in the Boulby mine in the north-east of England, and CRESST, also at Gran Sasso.

The crystals used in these detectors allow the experiments to distinguish between two distinct kinds of event: a dark-matter particle colliding with nuclei in the detecting material; and background signals — such as gamma rays or electrons — colliding with the crystal’s electrons. CDMS, for example, consists of stacks of silicon and germanium discs cooled to just a fraction of a degree above absolute zero, and it measures the vibrational energy deposited by a collision in the crystal lattice as well as the amount of ionization generated in neighbouring atoms by a recoiling particle. For a given energy deposition, the degree of ionization in germanium or silicon differs significantly depending on whether this is caused by electron or nuclear recoil, which allows the CDMS researchers to distinguish between these two types of event.

DAMA takes a different, and unique, approach. It measures the light given off when putative dark-matter particles collide with the nuclei inside the sodium iodide and excited neighbouring electrons drop back down to a stable energy level. On its own, this process of “scintillation” cannot be used to distinguish between electron and nuclear recoils, making it harder to clearly identify dark-matter signatures.

DAMA’s big advantage with respect to other experiments is that sodium iodide can be grown to make very large crystals. The DAMA/LIBRA detector’s mass of 250 kg dwarfs the 4 kg CDMS detector. So whereas CDMS filters out all gamma rays, electrons and neutrons, and has recently racked up 100-days worth of data without recording a single event, DAMA/LIBRA has collected over 800,000 (mostly background) events in four years. This far greater volume of data allows the researchers to look for an annual modulation signature, which is a variation of just a few per cent in the overall signal.

Separating the dark matter from the chaff

To demonstrate that the modulation is not caused by a background source, the DAMA researchers have performed a number of analyses. They have shown that the modulation occurs at energies between 2 and 6 keV, while there is no modulation between 6 and 14 keV, a signature that would also be expected if the cause was background radiation. Another important piece of evidence is that the modulation is only seen in “single-hit” events (those in which one of the 25 individual parts of the detector generates a flash of light in a photomultiplier tube) and not in multiple-hit events. Dark-matter particles could not generate the latter because the chances of them interacting with any given particle in the detector are so low.

The researchers also found no significant annual variation in a number of specific experimental parameters that could potentially mimic the dark-matter signature, including the temperature and levels of radon gas inside the experimental apparatus. In addition, using data on the cosmic-ray flux reaching the lab obtained by the neighbouring MACRO experiment (a flux that is known to vary annually with the temperature of the atmosphere), the researchers also showed that the variation in muon flux is far too small to account for the modulation in their dark-matter data.

But other researchers in the field are not convinced. Richard Gaitskell from Brown University in Rhode Island in the US, who is involved in several dark-matter searches, points out that there could be plenty of other background sources that the DAMA researchers have not yet thought of that have a period of a year, a maximum in June and a minimum in December. “Even if you ruled out 50 mundane sources, who is to say that you’ve ruled them all out. There could be 100. The trouble is that there are so many things that are annually modulated,” he says. Although there are no known background sources that produce this effect, Gaitskell suggests that the only way to remove this doubt is to provide greater detail of the signals in the different parts of the detector, which should be very similar if the signal is dark matter but different if the culprit is instead a mundane background source.

Gaitskell’s concern over backgrounds and possible systematics is shared by Collar, who points out that the lowest energy data — between 1 and 2 keV — may be contaminated by noise from the photomultiplier tubes. He maintains that if this noise is modulated, then it would to some extent “invade” higher energies and therefore replicate the dark-matter signal. As such, he would like to see the data plots (see “A sign of dark matter?”) extended down to 1&2dash;2 keV. “People have been asking for this data analysis for years,” he says. “The fact [the DAMA researchers] haven’t responded to this and other requests for specific data analyses stinks to high heaven.”

Collar is also surprised that the DAMA group has yet to publish data on any possible “diurnal” variation in its data, in other words the tiny daily variation in dark-matter particles reaching the detector due to the Earth’s rotation. Although a very subtle effect, he believes that the DAMA researchers have now collected enough data to look for it. “Maybe they have seen something that confirms this and are waiting for more data,” he adds. “Or maybe they haven’t seen anything and are hoping it will show up.”

Indeed, Collar claims that the DAMA group’s reluctance to reveal such data is indicative of a general lack of openness on the part of the Italy-based researchers. Bernabei maintains that all serious comments and suggestions that have been put forward by other dark-matter researchers have been dealt with, but Collar does not agree. “If you claim something like they have, you will be under the microscope. Five years ago people started asking questions but [the DAMA researchers] clammed up. That secrecy has not changed,” he says.

Data conflicts

At the root of the scepticism towards the DAMA group’s claim is the fact that the results appear to conflict with those of other experiments. Not only have other experiments so far not detected dark matter, the levels of sensitivity that they have obtained suggest that DAMA ought not to have done so either. The CDMS data collected to date show that the “WIMP–nucleon cross section” — a measure of how readily a WIMP interacts with a nucleon (neutron or proton) — for WIMPs with masses in the region of about 60 GeV/c2 must be less than 6 ×  10–8 picobarns. (A barn is the standard unit to denote cross sections in high-energy physics, which approximately corresponds to the cross-sectional area of a uranium nucleus, i.e. of the order of 10–28 m.) However, the analysis the DAMA group published in 2003 was consistent with WIMPs having a cross section of 7 ×  10–6 picobarns: in other words, a value at least 100 times higher than the CDMS limit. In addition, the XENON collaboration at Gran Sasso and the CoGENT collaboration, led by Collar, claim their experiments prove that the DAMA experiment cannot be observing WIMPs with lower masses. This was reinforced by a report last month from Collar’s group (see “Limiting factors”).

But the fact that the DAMA results are in apparent conflict with null results from other dark-matter experiments is a delicate one and the “apparent” in “apparent conflict” boils down to the fact that the DAMA experiment’s design is fundamentally different from other experiments, which makes it difficult to make comparisons. This view is defended by Bernabei, who believes her collaboration’s claim is not refuted by other experiments. In particular, she points out that the DAMA set-up follows a “model-independent approach” since it is not constrained, unlike the other experiments, to look for a specific kind of recoil event. She claims that her group’s results are consistent with a variety of hypothetical candidates, like axions or “light” dark matter, and not just WIMPs. “The [other researchers] should be more cautious and more honest in claiming the exclusions that they do,” she adds.

Indeed, Petr Vogel, a nuclear theorist at the California Institute of Technology in the US, believes that DAMA could be detecting something else other than WIMPs, since the experiment is sensitive to any type of recoil, including those involving electrons. “That would be rather unexpected, but not totally out of the question,” he adds. “So, unless somebody finds some prosaic explanation to the annual variation observed by DAMA, this exotic possibility needs to be explored.”

Axions, very light particles predicted to exist by extensions to the Standard Model, are the most obvious candidates to cause an electron-recoil signal. Bernard Sadoulet, a physicist at the University of California Berkeley and a spokesman for CDMS, says that his collaboration will shortly start analysing electron-like recoil events obtained by its experiment, which up to now have been filtered out but which are nevertheless present in the data. He says that a peak in the energy spectrum of these kind of events at about 3 keV would suggest DAMA is detecting axions.

This peak at 3 keV is also currently being investigated by Collar’s group using the germanium CoGENT detector, and Collar believes that the outcome will have a decisive effect on the DAMA results. “If we see something, then people will go crazy trying to reproduce DAMA’s results. But if we see nothing then people will probably lose interest in the claim.”

One experiment cannot by itself irrevocably confirm the direct observation of dark matter — the DAMA result needs to be replicated by an independent experiment such as CoGENT, CDMS or another sodiumiodide detector. But, whatever happens, it is not going to be easy for the DAMA researchers and other groups in the field to reconcile their differences. Like many others, Peter Cooper of Fermilab, who works with Collar on COUPP, believes that the DAMA group should be more open in its dealings with the rest of the community. But he also says that other physicists must be prepared to accept the DAMA claims, should they be vindicated. He hopes that people on both sides “calm down a bit and return to the science”, adding that scientific debate should be solved by rigorous experimentation and analysis, and not by whoever can shout the loudest. “People can say whatever they want,” he says. “It’s what nature says that matters.”

Copyright © 2024 by IOP Publishing Ltd and individual contributors