A long-standing and controversial claim by physicists in Italy to have detected dark matter might be the result of the unanticipated contamination of their photomultiplier tubes (PMTs). So argue researchers in the US, who reckon that a pattern of signal pulses recorded by the DAMA detector at Italy’s Gran Sasso National Laboratory could simply have been generated by small amounts of helium leaking into the experiment – a hypothesis that they say could be easily put to the test.
DAMA, run by Rita Bernabei of the University of Rome “Tor Vergata” and colleagues in Italy and China, consists of 25 cylindrical sodium iodide scintillators, each weighing 10 kg and capped at either end by a PMT. The idea is that a tiny fraction of any dark matter particles streaming through the detector will collide with nuclei, creating tiny flashes of light. When this light reaches the photocathode in each PMT, electrons are emitted via the photoelectric effect. These electrons are then “multiplied” in a high-voltage cascade through a series of dynodes, which produce a measurable electrical signal.
Bernabei and colleagues look for a roughly 1% sinusoidal variation in the rate of dark-matter collisions throughout the year – with a peak in the summer and a trough in the winter. This would correspond to small changes in the speed with which the Earth ploughs through the halo of dark matter believed to be enveloping our galaxy. To ensure that these signals are not smothered by flashes from cosmic rays and radioactivity, the detector is located 1400 m below Gran Sasso mountain and is shielded in successive layers of copper, lead, paraffin and rock, all of which have extremely low levels of radioactivity.
No plausible alternative
DAMA’s claim to have detected dark matter dates back to 1998 but has generated much controversy because similarly sensitive experiments have not confirmed DAMA’s findings. As a result, some physicists argue that a more prosaic (but hitherto unidentified) process could account for the annual variation. Bernabei and colleagues nevertheless remain defiant, reporting last May a very large statistical confidence in their oscillation of 9.5σ and pointing out that no plausible alternative process “has been found or suggested by anyone” in more than two decades.
Now, however, Daniel Ferenc at the University of California Davis and colleagues claim to have identified just such a process. The four-strong group, made up of Ferenc, his wife and two sons – all of whom are scientists – has for some time been developing light sensors that they reckon could replace PMTs in many applications. While doing so, they realized that there was one substance – helium – that they could not prevent from entering their devices and would therefore, says Ferenc, also be a problem for DAMA.
To establish which light flashes recorded in their experiment could be due to dark matter, Bernabei and co-workers apply various selection criteria. To remove processes taking place outside the crystals they only accept signals that are simultaneously registered by the two PMTs of any given scintillator. It is also important to rule out PMT signals that are “dark noise” caused by, among other things, thermally-excited electrons being spontaneously emitted from the photocathodes. This is done by only accepting signals that are bunched together within a timeframe of up to 600 ns. This is a little more than the “decay time” of the scintillator — the typical time it takes to produce light from a dark-matter collision.
But Ferenc and colleagues reckon that helium penetrating the DAMA shielding and ending up in the PMT vacuum tubes could make dark noise look like dark matter. Some dark-noise electrons could ionize helium atoms while en route to the first (positively-charged) dynode inside the device. The positively-charged ions would then be pulled back towards the cathode where they would free up other electrons.
Using a computer simulation of the kind of PMTs used in DAMA, Ferenc and colleagues found that these later “after-pulse” electrons would typically follow the initial thermal electrons with a delay nearly equal to the scintillator decay time. Such a signal could therefore be misidentified as coming from dark matter. This is unique for helium, because ions of heavier atoms and molecules would accelerate too slowly to reach the cathode within 600 ns.
Even at fairly small concentrations of helium, the team calculates two such after-pulses could be produced in both PMTs of a single scintillator at the same time at a high enough rate to explain the dark-matter signals claimed by Bernabei and colleagues.
Ferenc acknowledges that he and his colleagues do not know whether helium present in Gran Sasso fluctuates enough to cause a 1% annual variation within DAMA’s PMTs. But he says it is at least plausible, given that radon, which produces helium via alpha decay, is known to vary. In fact, he says, geological processes should produce far more helium than they do radon.
A light in the dark?
To test his group’s hypothesis, Ferenc says that the DAMA researchers simply need to unplug half of the 50 cables from the PMTs and then plug them back in so that they monitor coincidences between PMTs attached to different scintillators. If the annual modulation is indeed due to dark matter then it should vanish following the rewiring.
Richard Gaitskell of Brown University in the US, co-spokesperson of the LUX dark matter experiment in South Dakota, says that afterpulsing could account for at least some of the coincident signals recorded by DAMA. But he reckons that there is a “vanishingly small” chance of any significant amounts of helium reaching the PMTs, given that DAMA uses nitrogen to “purge” the surrounding environment of any unwanted gases. What would “resolve matters”, he says, are results from another suitably sensitive sodium-iodide experiment – several of which are in fact being developed.
The research is described in a preprint on arXiv.