Ghosts in the detector: why null results are part of science too

(OK, it’s Halloween, but I don’t have anything particularly Halloweeny to write about. Have a story about ghostly dark matter instead.)

The Large Underground Xenon (LUX) detector. [Credit: Lawrence Berkley National Laboratory]
The Large Underground Xenon (LUX) detector, located nearly a mile underground in South Dakota. The canister contains 368 kilograms of ultra-cold xenon. [Credit: Lawrence Berkley National Laboratory]
Dark matter is the most annoying stuff in the Universe (or maybe second only to people who insist that dark matter doesn’t exist). Even though it comprises about 80 percent of the mass in the cosmos and provides the metaphorical backbone for galaxies, it has eluded direct detection and we have yet to make dark matter particles in our big colliders. A few experiments, including the Cryogenic Dark Matter Search (CDMS), saw some tantalizing hints that could be dark matter, but which contradicted the results of other experiments.

Into this mix, a new detector — the Large Underground Xenon (LUX) detector — has arrived. This experiment is more sensitive than its competitors, so three months of operation were sufficient to begin to make some definitive statements about dark matter detection. Yesterday, the LUX collaboration announced that they had seen…nothing. No dark matter showed up in their detector, and particularly they saw nothing where CDMS had possibly seen a signal; you can read the official research paper here.

(I should point out that the CDMS scientists were wisely cautious: the three signals they saw were statistically marginal, and contradicted earlier results from the XENON experiment. In other words, there was good reason to think these weren’t dark matter particles. Nobody can accuse the CDMS collaboration of over-hyping their results! I’ll have more to say on this in a bit.)

LUX consists of a big tank of 368 kilograms of ultra-pure and very cold xenon, surrounded by arrays of detectors. The entire experiment is located nearly a mile underground in the old Homestake gold mine in South Dakota, to shield it from cosmic rays. (Connoisseurs of particle physics will remember that Homestake was also the site of a groundbreaking neutrino experiment.) If a particle hits the nucleus of one of those xenon nuclei, it can strip an electron away and produce a photon. Electric fields in the tank push electrons toward the top detectors, while photons are picked up by photomultiplier tubes, which amplify the signal from a single photon into a measurable signal.

Different types of particles produce different electron-photon signals inside the tank. Charged particles like electrons or positrons and photons will be common intruders, so researchers treat them as background noise. Neutrons are also noise, but they can mimic dark matter in some ways (being electrically neutral and relatively massive). However, neutrons are likely to hit with more than one nucleus, while dark matter interacts so weakly and rarely that a single collision is going to be as much as we’ll see…if we see even that!

Technically speaking, LUX, CDMS and their ilk are WIMP detectors: looking for weakly interacting massive particles. While WIMPs are the most popular dark matter candidate, they aren’t the only option. Theories of WIMPs are motivated by various particle physics models, though, and so hunting for them makes a lot of sense. The nice thing about LUX is that it can look for high-mass WIMPs — those with masses greater than about 35 billion electron-volts (35 GeV) — and low-mass WIMPs simultaneously, while most detectors can only hunt in one of those regimes. The CDMS candidates fell in the low-mass range: 8.6 GeV, lower than what most WIMP models predict, and where few had expected to see anything.

The relative strengths of the photon and electron signals in the LUX detector. For low-mass WIMPs, the ratio represented by the vertical axis would be reduced into the region bounded by red lines. [Credit: LUX collaboration]
The relative strengths of the photon and electron signals in the LUX detector. For low-mass WIMPs, the ratio represented by the vertical axis would be reduced into the region bounded by red lines. As you can see, most of the black points lie outside that region. [Credit: LUX collaboration]
That’s why the LUX null result is important. (See also this Résonaances blog post for some more technical information.) While it’s too early to rule out all WIMPs, these data put really strong constraints on what’s out there. Sure, there’s a slight possibility that the CDMS results are real, in which case dark matter interacts with the solid silicon detectors in a way that it doesn’t do with xenon. That would be (to use a technical term) friggin’ weird, but not completely impossible.

I have no particular favorite candidate for dark matter (though I’m fond of axions, unlikely as they are), but our Universe is weird enough with dark matter and dark energy. Realizing that dark matter doesn’t correspond to many of our simplest models is an additional weirdness. Perhaps dark matter interacts with normal matter via some new, unknown force that behaves differently than expected. That’s a cool idea! However, coupled with it is a more disturbing prospect: maybe dark matter only interacts with normal matter via gravity, in which case our current dark matter detectors are all doomed to find nothing. (Yet again, remember that the evidence for dark matter is strong from astronomical and cosmological observations.)

Null results are an important part of science, though. Basically, null results come about when you assume something is true and the experiment shows nothing. Sometimes you expect it, sometimes you don’t, but it’s as important as finding something you predict. Many precision experiments have showed that the speed of light is the same in all directions and doesn’t depend on how fast the measurer is moving; that null result is the foundation for the special theory of relativity. Similarly, the general theory of relativity relies on the equivalence between inertial mass (resistance to motion) and gravitational mass (which governs the strength of gravity); finding that equivalence is another form of null result.

We hope to find dark matter in our particle detectors, but null results like these from LUX help us refine our theories. LUX itself will continue to operate through 2015, and its successor will be even more sensitive. If there’s any dark matter hiding in its range of detectability, whether it corresponds to our current models or not, we should know in coming years. Who knows? We might yet find something entirely new. In the meantime, we’ll all wait for the next results from LUX and other detectors, and hope for something other than a null result.

5 responses to “Ghosts in the detector: why null results are part of science too”

  1. Allow me some questions :
    1- what exactly are the range of masses that was excluded ?
    2- what is the null result impact on supersymmetry ?
    3- I read that heavy WIMPS does not solve the hierarchy problem , at what range is this apply ?
    4- what if DM does not exist , what else do we have , MOND ?

    1. 1. LUX hasn’t definitively ruled out any mass ranges yet; the main thing it did was contradict (to a high degree of accuracy) certain earlier candidate detections from CDMS and other experiments. A lot of the actual paper deals with characterizing background noise and the like.
      2. I don’t think LUX has anything new to say about supersymmetry (SUSY). The Higgs boson mass and failure to create certain particles at the LHC that were predicted by the simplest version of SUSY are a bigger deal. SUSY is a big and complicated framework, so I’d hesitate to make sweeping statements about what these null results mean for the theory as a whole.
      3. That question would require a whole other blog post to unwrap. Personally I’m not a particle physicist; I’m interested in learning what dark matter is on its own terms, so stuff like the hierarchy problem is secondary to me.
      4. The problem with alternatives to dark matter is that they have to do everything dark matter does. They have to describe the behavior of the cosmic microwave background, the clustering of matter on the largest scales, the behavior of galaxy clusters, and the motion of material inside galaxies. MOND (to use your example) only handles the last of those, and even then only for some galaxies. That’s not to say that it can’t be done in principle, but it’s going to be an uphill climb. The reason why we’re 99% sure dark matter exists is because the model works. No other model has succeeded.

  2. Q: would an imaginary/hypothetical “time-reversed” particle (e.g. positron considered as a time-reversed electron) still exhibit gravitational mass?

    1. That’s a hard question to answer experimentally, but there’s some preliminary work on it:

  3. Thanks for that link. So, pending experimental verification, if the weak equivalence principle does hold for antimatter, particle inertial mass == gravitational mass, and both are positive. Next crackpot Q: could an imaginary “time-reversed” particle emit photons, or interact in any way with electromagnetic radiation?

%d bloggers like this: