Monday, 7 May 2018

Dark Matter goes sub-GeV

It must have been great to be a particle physicist in the 1990s. Everything was simple and clear then. They knew that, at the most fundamental level, nature was described by one of the five superstring theories which, at low energies, reduced to the Minimal Supersymmetric Standard Model. Dark matter also had a firm place in this narrative, being identified with the lightest neutralino of the MSSM. This simple-minded picture strongly influenced the experimental program of dark matter detection, which was almost entirely focused on the so-called WIMPs in the 1 GeV - 1 TeV mass range. Most of the detectors, including the current leaders XENON and LUX, are blind to sub-GeV dark matter, as slow and light incoming particles are unable to transfer a detectable amount of energy to the target nuclei.

Sometimes progress consists in realizing that you know nothing Jon Snow. The lack of new physics at the LHC invalidates most of the historical motivations for WIMPs. Theoretically, the mass of the dark matter particle could be anywhere between 10^-30 GeV and 10^19 GeV. There are myriads of models positioned anywhere in that range, and it's hard to argue with a straight face that any particular one is favored. We now know that we don't know what dark matter is, and that we should better search in many places. If anything, the small-scale problem of the 𝞚CDM cosmological model can be interpreted as a hint against the boring WIMPS and in favor of light dark matter. For example, if it turns out that dark matter has significant (nuclear size) self-interactions, that can only be realized with sub-GeV particles. 
                       
It takes some time for experiment to catch up with theory, but the process is already well in motion. There is some fascinating progress on the front of ultra-light axion dark matter, which deserves a separate post. Here I want to highlight the ongoing  developments in direct detection of dark matter particles with masses between MeV and GeV. Until recently, the only available constraint in that regime was obtained by recasting data from the XENON10 experiment - the grandfather of the currently operating XENON1T.  In XENON detectors there are two ingredients of the signal generated when a target nucleus is struck:  ionization electrons and scintillation photons. WIMP searches require both to discriminate signal from background. But MeV dark matter interacting with electrons could eject electrons from xenon atoms without producing scintillation. In the standard analysis, such events would be discarded as background. However,  this paper showed that, recycling the available XENON10 data on ionization-only events, one can exclude dark matter in the 100 MeV ballpark with the cross section for scattering on electrons larger than ~0.01 picobarn (10^-38 cm^2). This already has non-trivial consequences for concrete models; for example, a part of the parameter space of milli-charged dark matter is currently best constrained by XENON10.   

It is remarkable that so much useful information can be extracted by basically misusing data collected for another purpose (earlier this year the DarkSide-50 recast their own data in the same manner, excluding another chunk of the parameter space).  Nevertheless, dedicated experiments will soon  be taking over. Recently, two collaborations published first results from their prototype detectors:  one is SENSEI, which uses 0.1 gram of silicon CCDs, and the other is SuperCDMS, which uses 1 gram of silicon semiconductor.  Both are sensitive to eV energy depositions, thanks to which they can extend the search region to lower dark matter mass regions, and set novel limits in the virgin territory between 0.5 and 5 MeV.  A compilation of the existing direct detection limits is shown in the plot. As you can see, above 5 MeV the tiny prototypes cannot yet beat the XENON10 recast. But that will certainly change as soon as full-blown detectors are constructed, after which the XENON10 sensitivity should be improved by several orders of magnitude.
     
Should we be restless waiting for these results? Well, for any single experiment the chance of finding nothing are immensely larger than that of finding something. Nevertheless, the technical progress and the widening scope of searches offer some hope that the dark matter puzzle may be solved soon.

13 comments:

Unknown said...

I wonder how much much of the focus on GeV-to-TeV WIMPs was due to theoretical prejudice and how much due to the experimental bias (look for the easiest first). It seems hard to believe experimentalists would have built-in restrictions in their experiments if they are able to avoid them. We don't blind ourselves to low-mass dark matter, because theorists tell us that the mass must be greater than 1 GeV. I believe this is just as much a matter of advancing detection techniques. Detection of low-mass WIMPs is much harder and requires new types of detectors that we didn't have 10 years ago.

Mad Hatter said...

This is partly true: searching for dark matter is easier at high masses. However, the fact is that for long time experimentalists haven't looked *at all* into the MeV range, even though it was technologically possible to set some limits. I think this was due to the theoretical bias (and it's of course theorists' fault). As soon as the opportunities were pointed out in 1108.5383, things started moving on the experimental side too.

Ervin Goldfain said...

Let’s assume, for the sake of argument, that all searches for particle DM, anywhere from sub-eV to the TeV range, turn out to be empty in the long run. We already know that neither one of the DM models we currently have (Lambda-CDM, particle DM, modified gravity, primordial Black Holes) can account for the full span of DM observations, from cosmological to galactic scales.

At what point are we going to abandon the premise that DM must simply extrapolate the known physics of the SM and GR? At what point are we going to seriously consider that the Dark Sector may be an unforeseen topological property of spacetime?

Anonymous said...

How would a pb-sized scattering cross section of MeV particles be related to pair production cross section in electron/positron colliders? You can't observe that directly, but maybe there is some suitable associated production. Maybe even invisible branching fractions of quarkonia?

Mad Hatter said...

1707.00725 has some discussion of the interplay of the collider, flavor, cosmological, and direct detection bounds on MeV dark matter. Of course, this is highly model-dependent. However, they don't show the results in the plane mDM - \sigma_e, which would indeed be interesting to see.

Cédric Bardot said...

Thank you Jester to remind us that all good data is useful data and to provide the youngests (born after the theoretical completion of the standard model and educated under the supersymmetric standard paradigm) with important historical perspective!

May be Peter Galison could add a 7th chapter like "How experiments with null results necessarily never end" to his classic book. http://press.uchicago.edu/ucp/books/book/chicago/H/bo5969426.html).


This morning J.I Collar's preprint https://arxiv.org/abs/1805.02646 discusses the first experimental limits on strongly interacting dark matter candidates with a mass below 100 MeV, interacting preferentially via nuclear recoils.

Cédric Bardot said...

To Ervin Goldfain :

I think some qualified physicists (and mathematicians) have already considered seriously (and work hard enough) to foresee new spacetime topology/geometry in order to mimic the phenomenology of cosmological dark matter and dark energy (and possibly even galactic scale MOND).

If DM is a "topological/geometric" effect, a consequence of quanta of geometry or at(t)oms of spacetime that possibly cannot be disentangled in an Earth's lab, I guess one has no choice but to complete a genuine convergence model merging and going beyond the concordance cosmological model and the standard one (solving the cosmological constant and hierarchy issues at least) to increase astroparticle physicists attention as it is their natural purpose. One could also wait for the confirmation of EDGES experiment that traditional DM paradigm is 21 cm off the mark and find a way to wove smoothly the cosmic dawn episode in the spectral universe tapestry...

To Jester or anyone else informed : since QCD has been supposed to explain confinement of quarks, has experimental search for free fractional electric charge elementary particles been put to an end?

Mad Hatter said...

Certainly these searches are still being made. There is a whole industry of searching for millicharge particles (Q \lesssim 10^-3). Stable particles with quark-like charge, Q \sim 1/few, are excluded up to masses of \sim 100 GeV, but there could well be such particles with masses \sim TeV.

Ervin Goldfain said...

@ Cedric Bardot,

You say,

“If DM is a "topological/geometric" effect, a consequence of quanta of geometry or at(t)oms of spacetime…”

Note that I am not referring here to the discrete quantization of spacetime along the lines of Quantum Gravity theories (for example, Loop Quantum Gravity).There are alternative ways to describe the nontrivial topology of spacetime near or beyond the electroweak scale, with key implications on the flavor structure and dynamics of the Standard Model.

But this is clearly an off-topic conversation that can be continued elsewhere

Anonymous said...

The sensitivity of DS50 experiment to 10 MeV dark matter mass is not believable. How DS50 can access this much low mass while quenching of Argon is not well understood beyond 0.1 keVnr. Clearly, this limit is the result of extrapolation of quenching to lower nuclear recoil energy.

Urs Schreiber said...

I'd be interested in hearing about the "fascinating progress on the front of ultra-light axion dark matter"! Any chance you could drop a brief hint before you get around to writing a comprehensive post?

Mad Hatter said...

The recent progress by the ADMX axion experiment is described here https://physics.aps.org/articles/v11/34
Other experiments may also bite into the axion window soon, e.g. ABRACADABRA, see here: http://abracadabra.mit.edu/

Mad Hatter said...

Anon, the DS50 results mentioned in this post concern dark matter scattering on *electrons*, so they are not limited by the understanding of the nuclear quenching factors. In fact, the DS50 analysis is very similar to the XENON10 recast also described in the post.