Monday 17 February 2014

Signal of neutrino dark matter

The title of this post is purposely over-optimistic in order to increase the traffic. A more accurate statement is that a recent analysis of X-ray spectrum of galactic clusters claims the presence of a  monochromatic 3.5 keV photon line which can be interpreted as a signal of a 7 keV sterile neutrino dark matter candidate decaying into a photon and an ordinary neutrino. It's a long way before this claim may become a well-established signal.  Nevertheless, imo it's not the least believable hint of dark matter coming from astrophysics in recent years.  
 
First, let me  explain why one would anyone dirty their hands to study X-ray spectra. In the most popular scenario the dark matter particle is a WIMP -- a particle in the GeV-TeV mass ballpark that has weak-strength interactions with the ordinary matter.  This scenario may predict signals in gamma rays, high-energy anti-protons, electrons etc, and these are being searched high and low by several Earth-based and satellite experiments.  But and in principle the mass of the dark matter particle could be anywhere between 10^-30 and 10^50 GeV, and there are many other models of dark matter on the market.  One serious alternative to WIMPs is a keV-mass sterile neutrino. In general, neutrinos are dark matter: they are stable, electrically neutral, and are produced in the early universe.  However we know that the 3 neutrinos from the Standard Model constitute only a small fraction of dark matter, as otherwise they would affect the large-scale structure of the universe in a way that is inconsistent with observations. The story is different if the 3 "active" neutrinos have partners from beyond the Standard Model that do not interact with W and Z bosons -- the so-called "sterile" neutrinos. In fact, the simplest UV-complete models that generate masses for the active neutrinos require introducing at least 2 sterile neutrinos, so there are good reasons to believe that these guys exist. A sterile neutrino is a good dark matter candidate if its mass is larger than ~keV (because of the constraints from the large-scale structure) and if its lifetime is longer than the age of the universe.

How can we see if this is the right model? Dark matter that has no interactions with the visible matter seems hopeless. Fortunately, sterile neutrino dark matter is expected to decay and produce a smoking-gun signal in the form of a monochromatic photon line. This is because, in order to be produced in the early universe, the sterile neutrino should mix slightly with the active ones. In that case, oscillations of the active neutrinos into sterile ones in the primordial plasma can populate the number density of sterile neutrinos, and by this mechanism it is possible to explain the observed relic density of dark matter. But the same mixing will make the sterile neutrino decay, as shown in the diagrams here. If the sterile neutrino is light enough and/or the mixing is small enough then its lifetime can be much longer than the age of the universe, and then it remains a viable dark matter candidate. The tree-level decay into 3 ordinary neutrinos is undetectable, but the 2-body loop decay into a photon and and a neutrino results in production of photons with the energy E=mDM/2. Such a monochromatic photon line can potentially be observed. In fact, in the simplest models sterile neutrino dark matter heavier than ~50 keV would produce a too large  photon flux and is excluded. Thus the favored mass range for dark matter is between 1 and 50 keV. Then the photon line is predicted to fall into the X-ray domain that can be studied using X-ray satellites like XMM-Newton, Chandra, or Suzaku.

Until last week these searches were only providing lower limits on the lifetime of sterile neutrino dark matter. This paper claims they may have hit the jackpot.  The paper  use the XMM-Newton data to analyze the stacked X-ray spectra of many galaxy clusters where dark matter is lurking. After subtracting the background they see is this:


Although the natural reaction here is a resounding "are you kidding me", the claim is that the excess near 3.56 keV (red data points)  over the background model  is very significant, at 4-5 astrophysical sigma. It is difficult to assign this excess  to any known emission lines from usual atomic transitions. If the excess is interpreted as a signal of new physics, one compelling (though not unique) explanation is in terms of sterile neutrino dark matter. In that case, the measured energy and intensity of the line correspond to the the neutrino mass 7.1 keV and the mixing angle of order 5*10^-5, see the red star in the plot. This is allowed by other constraints and, by twiddling with the lepton asymmetry in the neutrino sector, consistent with the observed dark matter relic density.

Clearly, a lot could possibly go wrong with this kind of  analysis. For one thing, the suspected dark matter line doesn't stand alone in the spectrum. The background mentioned above consists not only  of continuous X-ray emission but also of monochromatic lines from known atomic transitions. Indeed,  the 2-10 keV range where the search was performed is pooped with emission lines: the authors fit 28 separate lines to the observed spectrum before finding the unexpected residue at 3.56 keV. The results depend on whether these other emission lines are modeled properly. Moreover, the known Ar XVII dielectronic recombination line happens to be nearby at 3.62 keV. The significance of the signal decreases when the flux from that line is allowed to be larger than predicted by models. So this analysis needs to be confirmed by other groups and by more data before we can safely get excited.


Decay diagrams borrowed from this review. For more up-to-date limits on sterile neutrino DM see this paper, or this plot. Update: another independent analysis of XMM-Newton data observes the anomalous 3.5 keV line in the Andromeda and the Perseus cluster.     

45 comments:

Alex said...

Well, this answers the question of your previous post, i.e. what experimental/observational results are there to look forward to in 2014? There might not be a lot of big experiments coming out with stuff, but one could hope for interesting things from smaller collaborations.

Anonymous said...

I have a graduate degree in HET that I've never really used. I'll freely admit that I'm a little out of touch here. However, I don't remember "is pooped with" being a quantitative term for quantum atomic emission phenomena.

Ervin Goldfain said...

It is intriguing that these results seem to be consistent with some earlier findings on neutrino physics.

For instance, Chandra has previously reported a sterile neutrino signal at 5 keV:

http://arxiv.org/abs/0912.0552

A signal of a fourth family non-standard neutrino was hinted in:

http://arxiv.org/pdf/1101.2755v4.pdf

Jester said...

They are not consistent with either :)
The first reference above is a good reminder that lines like this come around and go around.

Ervin Goldfain said...

It depends on your definition of consistency in neutrino physics...

Of course things are fluid and far from being settled but all this research may be pointing in the same direction. Only time will tell.

Parlyne said...

When you say sterile neutrinos are necessary for neutrino mass generation, isn't that presuming a type I or III seesaw?

Jester said...

Right, the way I wrote it was too strong. I'll correct it.

Kevork Abazajian said...

Nice to see coverage of this on Resonaances. This is a much more significant detection than the 1.8 sigma, after trials factor, detection in 0912.0552. Note they use a fairly outdated version of my parameter space plot from 2007.

An updated parameter space is available at http://www.physics.uci.edu/~kevork/constraints_v5Bulbul.png
with the detection point in yellow and with combined stat+sys errors. This is from our latest thorough study of sterile neutrino dark matter, Horiuchi et al. arXiv:1311.0282. As you can see in the plot, that signal, if real, is on the verge of being detected to low X-ray flux sources like M31 (Andromeda).

Jester said...

Thx Kev. Do you think the M31 DSph count limits are robust, such that they disfavor the x-ray line signal?

Anonymous said...

As a non-physicist amateur, I just want to thank you for your excellent blog.

kashyap vasavada said...

Current standard model has symmetry between no of quarks and no of leptons, except handedness in the case of neutrinos.Will the sterile neutrinos disturb this symmetry i.e now you will need more than 6 quarks? Are there already group structures which can accommodate sterile neutrinos and nothing else or they need a number of new particles?

Jester said...

You can just add sterile neutrinos alone. In the simplest version they are singlets under the SM gauge symmetry group so the local symmetry structure is unchanged. They violate lepton number if they have a Majorana mass, but that doesn't lead to any conflict with current experimental data. They can also fit in the grand unification scheme, e.g. in SO(10) models.

Kevork Abazajian said...

The M31 subhalo constraints we examined are those on thermally-produced sterile neutrinos from active-sterile oscillations in a Dodelson-Widrow type model. Outside of that model, which this potential detection necessarily is, the sterile can be produced from a variety of other mechanisms, some of which are cold enough to evade subhalo count and other structure formation constraints, including the Lyman-alpha forest. However, the "cooler" production mechanisms also lose some of the touted benefits of WDM, namely lower central densities in DM halos, and alleviation of the "Too Big to Fail" problem.

(Note I left this comment on a discussion on fb as well.)

Tories Smell said...

Just a pedantic point. "The dark matter could be anywhere between 10^-30 and 10^50 GeV". The lower limit could be smaller, as low as H0=10^-33 eV=10^-41 GeV. And also, where did you get the upper limit from? It's way above the Planck scale, which on a lenient day I'm happy with, but what is the upper end if not this?

George Djorgovski said...

Oh please... the 7 keV iron emission line complex is the strongest emission "line" seen in the x-ray spectra of galaxy clusters, and it is well known. This is well known among the people who actually bother to learn about these things, which, I guess, does not include many particle physicists (sadly, not an uncommon phenomenon). For a pedagogical intro, see, e.g., http://ned.ipac.caltech.edu/level5/March02/Sarazin/Sarazin5_2_3.html and for the original work, http://adsabs.harvard.edu/cgi-bin/bib_query?1977ApJ...213L..99B (1977!).

Tories Smell said...

Also, you limit of "must be heavier than 1keV" is a bit loose. The 1keV limit applies to thermal warm dark matter. Euivalent sterile neutrino masses are much larger (see e.g. Table I http://arxiv.org/pdf/1401.3769.pdf). The best limit from structure formation is Lyman-alpha which puts pressure on 3.3 keV thermal, which is >16 keV sterile neutrino mass. The same analysis rules out 1keV a 9 astrophysical sigma and 2keV at 4, which I would say rules out this 7 keV neutrino quite convincingly. [Edit to above: Abazajian has already discussed this: thanks!]

Tories Smell said...

Sorry, here is the link to the Lyman-alpha limits http://arxiv.org/abs/1306.2314

Matt Reece said...

It could be worth noting that even a 3 keV line signal from dark matter isn't necessarily from sterile neutrino dark matter. For instance, in models of low-scale SUSY breaking one might have moduli fields around this mass that decay. Such models tend to massively overpredict the abundance of these particles, but in any case the set of constraints and cosmological issues to think about is different from in the sterile neutrino case.

Jester said...

Doddy, i borrowed the DM mass limits from Murayama's talk that I once heard. For the upper limit, on a lenient day you can imagine dark matter is made of macroscopic objects, in which case the upper limit comes from microlensing, roughly 10^-7 solar mass. The lower limit is from the uncertainty principle, see http://arxiv.org/pdf/astro-ph/0003365.pdf

Jester said...

George, no doubt about it that particle physicists are idiots. But.... 1) the analysis was done by legal astrophysicists, 2) what they found could be many things but certainly not the iron line you're talking about (it's 3 keV away, while the instrumental resolution is of order 0.1 keV)

Kevork Abazajian said...

My comments to George Djorgovski, who posted a similar comment on fb:

"George, you're dismissive with no basis. Markevitch, the 2nd author, is a respected X-ray astronomer. What did they miss in their extensive discussion of metal lines in section 5.1?? I'd love to know...

To which he responded:
"I stand partly corrected - I saw 7 keV in the blog post, and did not look at the actual paper; my comment stands as far as the "7 keV sterile neutrino" is concerned. As for the paper itself, they choose to interpret to what to me looks like a noise bump as a real line or a complex thereof; even if it is real, it would be far more sensible, in my mind, to interpret it as some as yet poorly modeled weak line(s) from the intracluster plasma, than to invoke some particles not known to"

And my follow up:
"If physicists did not model anomalies with new particle candidates that are not yet "known to exist", no new particles would ever be discovered. That's a troubling position to take. Sterile neutrinos are a well motivated warm dark matter candidate. Of course, that does not require their existence, just good motivation to search for them. However, I agree that this is likely astrophysical (or instrumental) line(s) as their origin, but still intriguing and should be followed up."

Kevork Abazajian said...

Moving very quickly, the line is now seen in Andromeda observations as well, at 4.4σ when combined with Perseus. It also follows the expected dark matter profile for them... http://arxiv.org/abs/1402.4119

Xezlec said...

So if it can decay to detectable particles, that means it can be produced by them as well, right? So if it's real, then the only reason we haven't seen this 7 keV guy in colliders is that its coupling to those more mundane fields is insanely weak? Am I understanding that right? (Sorry, my engineer brain is not adequately pooped with physics knowledge.)

Jester said...

In principle yes: Z or W bosons can decay into sterile neutrinos. But because of the weak coupling this is extremely rare. Moreover, a sterile neutrino hardly distinguishable from the normal ones, especially at high energy colliders.
I think if the signal is confirmed we will eventually get there (maybe via high intensity neutrino beams) but it's not gonna be easy.

L. Edgar Otto said...

Hi Jester,
When A larger LHC is built do you think it will clarify this. Sterile idea if particles,, IR will it give us a wider version of a standard theory? Is such possible evidence extrodinary or pushes the idea if SUSY deeper still hidden in hints of new physics

Jester said...

As I mentioned above, high energy colliders are not very useful to study super-weakly interacting keV mass neutrinos. So a bigger collider cannot confirm this claim -- it can only somewhat constrain a small fraction of alternative theories of dark matter.

Kevork Abazajian said...

Searches for these types of sterile neutrinos in the laboratory have mainly focused on kink searches in nuclear beta decay, e.g. http://arxiv.org/abs/arXiv:1109.3452

This of course is how the infamous "17 keV" neutrino had several candidate detections:
http://www.nature.com/nature/journal/v366/n6450/abs/366029a0.html

Ervin Goldfain said...

Kevork,

If I recall correctly, the 17 keV neutrino was an artifact of some uncontrolled edge scattering of the outgoing beam near the detector surface.

What is the likelihood of a similar systematic error in searches for sterile neutrinos?

Tories Smell said...

Hi Jester,

Yes, the Hu et al model does put things about where you say at 10^-30 GeV, but the details of that depend on the cosmological and LSS data (see http://arxiv.org/abs/hep-ph/0509257, or http://arxiv.org/abs/1307.1705). The situation is much like the ~1 keV constraint on WDM/neutrinos (but less well studied). With mixed components (like the standard neutrinos/HDM) there are of course only upper limits. What I mean by ~10^-40 GeV is an "in principle" lower bound, like the 10^50 GeV "in principle" upper bound.

In order to call something DM, my only constraint is that it must cluster on scales smaller than the horizon. So, if you allow the fraction of the DM to arbitrarily small, this simply bounds the mass from below by H0. This is precisely using the uncertainty principle as you say. But it is quibbling over percent-level details on my part.

Thanks for a great post, by the way.

Unknown said...

The keV range is hard to believe unless one adheres to a variable mass model for sterile neutrinos, and this is considered in the literature. Of course, the reactor experiments strongly indicate a 1eV neutrino and cosmological bounds are quite tight too. Fun times!

Unknown said...

Intriguingly, under the 'mass temperature' equivalence hypothesis for non local neutrino DM, this keV line corresponds to around 4 million degrees, which is the lowest typical temperature for stellar fusion.

http://www.sun.org/encyclopedia/stars#Introduction

Has anyone tried to link the line to specific stellar processes?

Kevork Abazajian said...

Ervin, There are really only two kinds of systematic uncertainties here: models of the plasma emission lines and an unknown feature the instrumental response. The Bulbul et al. paper addresses both, the first effect is addressed because of the extensive analysis they do of the plasma emission possibilities in Section 5.1, and the second effect is addressed because any instrumental response would be washed out because the clusters have different redshifts, and respond do different observed frame energies. There are no obvious other systematics that remain untested, which is why the Bulbul et al. paper is 25 pages long.

Ervin Goldfain said...

Thanks Kevork.

Bulbul et al. paper covers indeed all the caveats related to the DM interpretation of the spectral line in section 6. It's likely too early to declare victory...

Unknown said...

http://arxiv.org/abs/0906.2968

This Kusenko review is excellent, thanks. If one insists on a local particle DM model, then his case for relic (EW scale generated) sterile neutrinos is very convincing, and a keV line was certainly predicted by this scenario. (But one could envisage other mechanisms besides the hypothetical nu/gamma decay).

Anonymous said...

are you kidding me

Anonymous said...

Interesting... The DAMA modulation energy spectra also is a peak near 3 keV (with not negligible spread).

Jester said...

There's absolutely no connection. Even if sterile neutrinos could for some reason interact with matter strongly enough to produce a scattering signal, the recoil energy of the scattered atom would be many orders of magnitude smaller than keV.

Unknown said...

Jester, sterile neutrinos might be fully absorbed in some QG mechanism.

Jester said...

I don't know what it means

Francesco Nozzoli said...

The idea could be: assume this hypothetical 7 keV long living particle (not necessarily a sterile neutrino) can decay faster in the field of the nucleus/electron exchanging some momentum (so few energy) to this heavy spectator. Then the photon+(small recoil) energy is detected and the approx 3.5 keV light-neutrino would escape the detector.
A similar mechanism would elude the LUX (or similar experiment) limit since they are based on nuclear recoil and electromagnetic background rejection.
A signal could be searched also in the low energy-threshold Germanium detectors (CoGeNT, CDMS) similarly to the axion-like case (a monocromatic peak was expected for axions, a broad peak would be expected in this case)

Jester said...

:)
Looooong shot, but ok...

Unknown said...

At present, everything is a long shot, as far as the theory is concerned. And it might not be a 7 keV sterile particle; that's just one theory. Yes, theorists are definitely idiots, but they do have a knack for sorting out those funny factors of 2.

Unknown said...

But we don't even understand neutrino oscillation properly yet, or whether they have mass or if it is its own anti-particle.I hope we come to understand that by the time sterile neutrinos can be confirmed experimentally.

Anonymous said...

In models which predict sterile neutrino DM in this mass range, for instance those from Shaposhnikov et al., there are also (degenerate) heavier sterile neutrinos in the GeV mass range. Does anyone know if these could affect the anomalous magnetic moment of the muon sufficiently to explain the discrepancy observed in the BNL E821 muon g-2 experiment?

Unknown said...

http://arxiv.org/abs/1402.5837

For reasons beyond my comprehension, this new 7keV theory paper also looks at susy. Arrrgggh!