Do we see dark matter emission from dwarf spheroidal galaxies?

From a dark matter (DM) hunter’s perspective, this year’s Fermi Symposium was highly anticipated. In the six years since the launch of the Large Area Telescope (LAT), we’ve seen our share of ups and downs. An active community, both in and outside the Fermi Collaboration (FC), works hard to fit dark matter to or explain away every deviation in excess of what we expect from the gamma-ray sky. This year’s gathering got the answer to the latest burning question: do we see dark matter emission from dwarf spheroidal galaxies (dSphs)?

Image Credit: NASA, Hubble Heritage Team, (STScI/AURA), ESA, S. Beckwith (STScI). Additional Processing: Robert Gendler

Before we hear the answer, let’s review the mileposts of the LAT DM saga. The first big stir came from the electron/positron spectrum [1], which featured a weaker version of the bump-like feature already measured by the PAMELA experiment. Theorists rushed to explain this with special flavors of ‘leptophillic’ dark matter, tuned to enhance the local flux while producing nothing other detectors would have seen. After an avalanche of 700+ citations and one even more precise measurement (AMS-02), the excitement died, along with most of those models.

Next we had the (in)famous line feature at 135 GeV [2]. With a far more clear-cut interpretation, I held my breath while this result was seen first in the galactic center (GC), then in nearly every control region, and finally at lower and lower significance [3] until its appearance at the Symposium reported a mere 0.72 sigma. A dedicated observation from H.E.S.S. II may rescue it, but for now the line looks tenuous at best.

The latest big hint also comes from the galactic center, but in the continuum emission, rather than as a sharp feature. This makes some sense; the galactic center is nearby and massive, so any dark matter signal ought to be strongest there. On the other hand, it is astrophysically complex. So much so that interpretation of the data there seems to have been saved for last (the Fermi Collaboration has yet to weigh in with a publication). Undaunted, several groups (e.g. [4]) now report that models of known sources do not account for all the emission, leaving what is blandly referred to as the galactic center excess (GCE). Despite a wealth of systematic uncertainty, all parties agree that the galactic center excess is peaked at around 1 GeV and extends fairly symmetrically to about ten degrees from the center of the Milky Way.

Plausible conventional production mechanisms for the galactic center excess include pulsars and cosmic rays, but it can also be neatly accounted for by the existence of a weakly interacting massive particle (WIMP) of about 30 GeV which self-annihilates into b quarks. Without completely ruling out alternatives, this possible dark matter explanation remains just that — possible. What made this dark matter model so exciting was that it also happened to fit a slight excess (1.4 sigma) seen in the dwarf spheroidal galaxies (dSphs) orbiting the Milky Way [5]. As dSphs represent an independent data set, and a nearly background-free one at that, the coincidence was indeed tantalizing. One well-known physicist called the galactic center excess “the most compelling signal we’ve had for dark matter particles – ever.” [6]

The dSph excess was so low that, if it was not just a fluctuation of the background, we would have to wait many years to confirm the dark matter explanation of the galactic center excess. Fortunately, the LAT’s excellent ground team just gave us a big push. A pair-conversion telescope, Fermi relies on extensively calibrated classification algorithms to reconstruct incoming gamma rays from their electronic signatures. These routines have been periodically overhauled throughout the mission as knowledge of the instrument continues to improve. The latest overhaul, known as “Pass 8,” marks the biggest advance yet, boosting the instrument’s effective area while lowering its point spread function like a pair of glasses. Upgrade in hand, the time was right to look again at the dSphs.

Dark matter searches in dwarf spheroidal galaxies have been a specialty of the Oskar Klein Centre. C. Farnier and M. Llena-Garde have both played lead roles in Collaboration papers on the subject ([7], [8]), and J. Conrad introduced the statistical technique now used to combine information from multiple targets [9]. For the latest publication, G. Martinez derived the dSph mass distributions by a nested Bayesian analysis of their hosted stellar populations [10]. Building on this success, yours truly, as part of the FC, took a peek at the Pass 8 data.

Preliminary Fermi-LAT Pass 8 constraints on the WIMP velocity-averaged cross section for one annihilation channel. Using five years of data for the combination of 15 dwarf spheroidal galaxies. Dashed line and bands represent limits expected from blank fields on the sky.” Credits Fermi-LAT Collaboration.

What we found was a whole lot of nothing. The significance of the GCE model dropped drastically, along with all other WIMP annihilation masses and channels. Dropped so far, in fact, that we can now set limits which exclude the annihilation cross section WIMPs need to make up all dark matter out to masses of 100 GeV (see Figure 1). These are now the best limits in the world below 1 TeV, and represent a big bite out of the parameter space left to the indirect dark matter detection field’s favorite class of models. While these constraints do not conclusively rule out the dark matter interpretation of the galactic center excess, they lend no support. “Tension” is the colloquial term.

So at this year’s Fermi Symposium, though debate still raged over the galactic center, I had to report that dSphs had pulled their support from the dark matter interpretation. Like all of us, I was disappointed to find we still have no answer to one of the greatest physics questions of the day, but as I said, dark matter hunters are used to highs and lows. Could gamma rays from annihilating dark matter still be buried in the LAT data? Of course, but there are not many coming from dSphs.

– Brandon Anderson, OKC fellow (

[1] Fermi-LAT Collaboration. Physical Review Letters, vol. 102, Issue 18, id. 181101
[2] C. Weniger. Journal of Cosmology and Astroparticle Physics, Issue 08, article id. 007 (2012)
[3] Fermi-LAT Collaboration. Physical Review D 88, 082002 (2013)
[4] D. Hooper, L. Goodenough. Physics Letters B, Volume 697, Issue 5, p. 412-428
[5] Fermi-LAT Collaboration. Phys. Rev. D 89, 042001 (2014)
[6] New Scientist, April 2014
[7] Fermi-LAT Collaboration. Astrophysical Journal, 712, (2010), 147-158
[8] Fermi-LAT Collaboration. Phys. Rev. Lett. 107, 241302 (2011)
[9] J. Conrad. Astroparticle Physics 62 (2015) 165-177
[10] G. Martinez. eprint arXiv:1309.2641

Leave a Reply

Your email address will not be published. Required fields are marked *