An unexpected post this time around, but I just noticed the Berkeley Labs’ press release from earlier this week announcing the start of construction of LZ – the next generation U.S. detector designed for the direct detection of dark matter. LZ (short for “LUX-ZEPLIN”) is the follow-up experiment to the LUX and ZEPLIN-III, both of which used Xenon as a detector medium to search for nuclear recoils from collisions with dark matter particles in our galaxy. Just last year LUX set the most stringent results to date on the spin-independent cross section of dark matter particles under the WIMP hypothesis (I explained the WIMP miracle briefly in my last post, in the context of the DAMA experiment).

To understand what LZ is setting out to do, it’s helpful to first look at what LUX accomplished. Because we know the density of dark matter particles in the local part of the galaxy, we can write down the number of expected signals from dark matter scatters in terms of two quantities: the dark matter particle’s mass and it’s interaction cross section (the cross section is related to the strength of the dark matter particle’s interaction with nuclei). Then, based on the number of recoils seen (or not seen) in a detector, we can set a limit in this parameter space. When we do this, we’re essentially saying that if the dark matter particle had mass X (say, 10 GeV/c^2), we know that it’s cross section has to be less than Y, or else we would have seen more signals. Since the excluded cross section depends on the mass, we can draw the excluded “region” in the mass-cross section plane, and get plots like this:

Spin-Independent limits for dark matter detection in the mass-cross section plane, as of 2013. Courtesy http://newscenter.lbl.gov.

In this plot, the upper right regions above the curves are excluded by different experiments. The shaded regions are those that are excluded, while the dashed lines are the expected exclusion regions based on future experiments (assuming they don’t see anything!) If you look carefully, you can see that for a wide range of mass values, the limits from LUX in 2013 were the most stringent to date at the time. Their updated limits that were just released last month are essentially at the dashed line labelled “LUX 300-day”¹.

(footnote: unfortunately, I wasn’t able to find an updated plot with the latest limits from all experiments, but the expected curves shown here are essentially the same as the latest results.)

On this plot, you can also see the expected results from a full run of LZ – as the article above notes, LZ’s biggest competitors are the XENON1T experiment in Italy and PandaX-II in Japan. All three experiments have similar construction schedules and should have relatively comparable sensitivities, so the pressure is on to see who can put out the best results first. This is a great example of competition driving innovations in experimental physics – all three detectors are based on the same principle and use liquid xenon as their detector material, but they will each have a variety of strategies to try and make the most of their results. You can read a bit more about the competition in Symmetry Magazine’s article and in the Berkeley Lab press release above.

One last thing that I can’t help but make note of while I’m on the subject – if you looked at the plot above you probably couldn’t help but notice the shaded yellow region in the bottom left marked off by the thick dashed orange line. These regions aren’t excluded, but they denote the parameter space where we expect to see what’s called coherent neutrino scattering. When the dark matter detectors reach this level of sensitivity, they’ll be sensitive to neutrino recoils off the nuclei in the detector, which have a very similar signal to dark matter recoils. These coherent neutrino scatters have never been observed, but once we reach this region, it will be tough to continue setting new limits as we might not be able to tell whether a signal was due to dark matter or an ordinary neutrino. For this reason, this region is often called the “neutrino floor”. If you look carefully, you can see that there is a small space where the expected LZ curve enters this region – it might not be long before we hit this floor! Some research is already underway into techniques that could be used to discriminate between neutrino events and dark matter events – I’ll have a lot more to say on these later.

¹Unfortunately, I wasn’t able to find an updated plot with the latest limits from all experiments, but the expected curves shown here are essentially the same as the latest results.

As the new year begins with no new tantalizing signals from the LHC, perhaps it’s a good time to look back at an older experimental anomaly that has somehow survived with no satisfying explanation for almost two decades: the annual fluctuations in the DAMA experiment. As one might expect for an anomaly that’s been around this long, there’s a lot of material (and some controversy) to unpack here. In this posting I’m going to try and explain what DAMA set out to do and what they observed. Next time around, I’ll try to dig into some of the controversies — both scientific and political — that have prevented a consensus interpretation of DAMA’s results.

Direct Detection of Dark Matter

DAMA is one of many so-called ‘direct detection’ experiments searching for dark matter in the local part of our galaxy. The basic premise is as follows: we have good cosmological reasons to believe that there is some type of massive particle permeating our galaxy, and in particular our solar system. These particles are massive, but their mass could be essentially anything, they must be electrically neutral (or else we’d have seen them), but they could in principle interact with ordinary matter through some much weaker force. Thus, these particles would whiz around in our solar system, constantly passing through us and everything around us similar to the way most neutrinos do. While we don’t know exactly what type of particle could constitute the dark matter, we know that it can’t be described by anything in the standard model, and while there are countless theoretical models that propose various candidates, none have any experimental evidence thus far.

But for now, let’s suppose that the dark matter particles do have some very weak interaction with the electrons and protons we’re accustomed to (the so-called “weakly interacting massive particle”, or “WIMP” hypothesis). Then, if we take a big tank of ordinary matter and watch carefully, eventually a dark matter particle in the solar system will collide with some of the matter in our tank, causing the nucleus of the particle we’re watching to recoil with some energy that we can in principle measure. This technique is what’s called direct detection (to distinguish it from “indirect detection”, which searches for two dark matter particles colliding and annihilating to produce standard model particles — I’ll talk more about this in a post to come!)

DAMA’s twist: Annual modulations

One of the most difficult things about the direct detection methods we’ve talked so far is how to eliminate background. Since the detection material is made of ordinary matter, it interacts ordinarily through radioactive sources and cosmic rays and it’s not unexpected to see some signal in the detector even in the absence of dark matter. Most experiments deal with this by carefully shielding against and quantifying their backgrounds and/or discriminating them by some other means. The DAMA experiment takes a different approach.

The underlying idea behind DAMA is that the dark matter is distributed roughly randomly in a spherical shape throughout the galaxy, and thus since the ordinary matter in our galaxy is rotating, there is a sort of “wind” of dark matter particles passing through our solar system in one direction. Since the earth rotates around the sun roughly in this same plane, during parts of the year the earth’s movement will be with this wind — decreasing the flux of dark matter particles passing through the earth, while the other part of the year the earth’s movement will be against the wind, correspondingly increasing the flux. In contrast, the usual radioactive and cosmic backgrounds in a detector have no reason to oscillate in such a way. Thus, a detector can bypass the need to distinguish between dark matter particles and backgrounds if they can see some sort of annual oscillation in their data: if the oscillation has a one year period and the right phase, it must be the dark matter!

The Results:

DAMA set out to observe exactly this annual modulation, and to many in the community’s surprise, it wasn’t long before there was a clear oscillation in their data. The plot below shows the number of detection events per day as a function of time, with the constant part of the signal subtracted out. The solid line overlaid shows the fit to an ordinary sine-wave with a period of one year, and the phase fixed to peak when the earth is moving maximally against the rotation of the galaxy. You can read the rest of the paper for more details and some of the other check’s DAMA has done, but the data is quite clear: DAMA is observing something interacting with their detector that has exactly the annual fluctuations we expect from dark matter.

Oscillation observed by DAMA with recoil energies between 2-6 keV, adapted from doi:10.1140/epjc/s10052-010-1303-9.

If you believe everything I’ve said so far, it’s hard not to look at the plot above and believe that the dark matter question must be settled. It’s not hard to take the amplitude of the graph above, combined with the energy of the nuclear recoils in the detector and the known density of dark matter in our galaxy that we can obtain from cosmology and extract information about the mass and interaction strength of the dark matter particle. In terms of the statistical confidence level used to describe the strength of a signal, DAMA’s result shows a whopping 8.9σ detection, far surpassing the usual 5σ threshold to claim “discovery”.

Given all this, it’s somewhat surprising that the community by and large doesn’t regard the DAMA result as evidence of dark matter detection.In my next post, I’ll discuss why this is the case, and give a brief overview of how the story might be resolved in the coming years — stay tuned!

References:
[1] Bernabei, R., Belli, P., Cappella, F. et al., “Final model independent result of DAMA/LIBRA-phase 1”. Eur. Phys. J. C (2013) 73: 2648. doi:10.1140/epjc/s10052-013-2648-7.