It was in the 1930s when astronomer Fritz Zwicky first showed that the amount of luminous matter in clusters of galaxies was not enough to explain their gravitational binding. More astrophysical evidence has been accumulated since then: astronomical observations of spiral galaxies, gravitational lensing phenomena, the anisotropy in the cosmic microwave background and the large-scale structure of the Universe all show that approximately 95% of the Universe does not consist of ordinary matter. Out of this, ∼ 23% is constituted of dark matter (DM) while the remaining ∼ 72% is called dark energy, a still unknown form of energy which is responsible for the accelerated expansion of the Universe (we will not discuss it here though).
While there exist tantalising evidences for the presence of a dark component of matter in the Universe, its nature remains unknown. Current astrophysical and cosmological observations favour a particle nature of DM, called non-baryonic DM. However, since none of the known particles of the Standard Model (SM) of particle physics can explain it, a plethora of new candidates have been introduced to account for DM. Of particular interest are DM particles with weak interactions, labeled as WIMPs (Weakly-Interacting Massive Particles). WIMPs possess the promising property that their current abundance, when extrapolated from an equilibrium state in the early Universe, can account for all DM. These compelling DM candidates are present in popular extensions of the SM, such as Supersymmetry and Universal Extra Dimensions, for instance. Many proposed theories to solve other longstanding puzzles of the SM have indeed the additional feature of predicting new particles with the properties required by DM.
Identifying the nature and the properties of particle DM is a primary goal of modern theoretical physics. At this purpose, two strategies have been envisaged:
• direct detection which consists in looking for elastic scattering of galactic WIMPs off atomic nuclei in low background underground detectors;
• indirect detection through the observation in the cosmic ray flux of the annihilation and/or decay products of DM from environments of the Universe where it is most concentrated.
Complementary to these techniques, collider searches for the production of new particles through the collision of high energy SM particles (such as protons at the LHC) can provide valuable information concerning new particles which could be identified with the DM.
In the following, we are going to consider indirect searches for WIMP DM. Among the SM particles which can be used for indirect detection experiments there are antimatter particles (like positrons, anti-protons, anti-deuterons..), neutrinos and photons. The latter two can proceed almost unaffected by magnetic fields when propagating in the Universe, and therefore they can carry spatial information about their origin. This makes photons, and in particular γ-rays, promising messengers for DM indirect detection.
The most promising astrophysical targets to look at are those regions of the sky where DM is more concentrated while at the same time the astrophysical background is low or easy to disentangle. Dwarf spheroidal galaxies of the Milky Way provide excellent targets for γ-ray searches for WIMPs. They are characterised by a large DM content, low diffuse Galactic γ-ray foregrounds, and lack of conventional astrophysical γ-ray production mechanisms.
Since from very ancient times humanity wondered to find a rationale in the disposition of objects appearing in the starry sky. We now know that matter is not randomly distributed in the Universe, but is organised in structures and substructures as a result of the DM aggregation under the effect of gravitational attraction. Ordinary matter, which is the only source of visible light we know so far, is in fact a small constituent of the matter component of the Universe, overcounted by DM as approximately five parts to one. This implies that DM is the component which drives the structure formation in the Universe, especially on large scales and at early times, while ordinary matter is dragged behind by the dark component. Eventually, ordinary baryonic matter radiates away its kinetic energy and fall toward the centres of the DM halos thus giving rise to the visible galaxies.
The Universe has a hierarchical structure: at very large scales it is characterised by large voids separated by sheets and filaments of galaxies, with superclusters of galaxies where the filaments intersect. This is a consequence of the fact that DM is indeed distributed in filaments at large scales, as we know from the results of computer simulations performed to trace the evolution of DM under the influence of gravitational interactions. Within the filaments the DM density is not constant and regions of overdensity are present, the so called halos. Ordinary matter is gravitationally dragged inside the halos, eventually accumulating at their centres where galaxies are formed. But this is not the end of the story, the halos themselves are not uniform and posses overdensity regions. If the size of these subhalos is large enough to attract a sufficient amount of ordinary matter additional galaxies are formed inside them; these small galaxies are referred to as dwarf galaxies and orbit around the main halo, and thus around the main galaxy. Dwarf galaxies, ordinary galaxies and cluster of galaxies are all suitable targets when searching for DM annihilation or decay signals. Galaxy clusters are interesting since they are the largest known gravitationally-bound objects in the Universe, but their γ-ray spectra are expected to be contaminated by signals coming from their intergalactic medium and suffer from uncertainties on the determination of the underlying DM profile. Among ordinary galaxies it is naturally to look at the Milky Way and in particular at its center, where the higher density of DM is supposed to be. The galactic center has the advantage of an expected large γ-ray flux from DM annihilation, but has the disadvantage of a comparably high background from the large number of astrophysical sources it contains, from which any putative DM signal must de disentangled.
Figure 1: A result from the Millennium Simulation  describing the DM distribution at large scales. It is easy to spot the presence of voids, filaments and halos which, as described in the text, determine the distribution of galaxies in the Universe.
Dwarf galaxies are excellent probes for DM indirect detection searches: they are thought to lack other γ-ray-producing sources and posses a low stellar population, of the order of some billions of stars, to be compared with the roughly 300 billions present in our Milky Way. Moreover they are DM dominated objects, being composed up to 99% by DM. This makes them very “clean” targets for this kind of observation and a γ-ray flux from a dwarf galaxy would make a very strong case for DM. On the other hand for the same reasons they are very difficult to spot, being low luminosity objects. Up to the beginning of 2015 only 27 Milky Way satellite galaxies were known, but recently 10 new dwarf galaxy candidates have been discovered [5–7]; that is very promising and expands the prospectives for future DM searches in dwarf galaxies.
Figure 2: A result from the Aquarius Project , that simulated a DM halo similar in mass to the one containing our Milky Way galaxy. Notice the presence of subhalos which can possibly host dwarf galaxies.
Among the most powerful experiments measuring the flux of γ-rays from the sky, there is the Large Area Telescope (LAT) onboard of the Fermi satellite, launched in June 2008 into low Earth orbit. The LAT is sensitive to photons in the range of approximately 20 MeV - 300 GeV.
The Fermi-LAT collaboration has recently reported a new analysis based on γ-ray observations of Milky Way dwarf galaxies based on 6 years of Fermi-LAT data . They updated the previous analysis, which had been released in 2013 , using a new dataset, known as Pass 8. This event- level analysis framework shows an upgrade of the event reconstruction and analysis of Fermi-LAT data, which results in a much improved dataset, including better localisations, increased effective area over a wider energy range, and more well reconstructed γ-rays. These improvements, in instrument simulation, reconstruction code, event selection, instrument response functions, systematic uncertainties, isotropic templates. . . , along with two additional years of data taking, lead to a predicted increase in sensitivity of 70%1 relative to the four-year analysis .
The Fermi-LAT collaboration presents a combined analysis of 15 Milky Way dwarf galaxies using this new and improved LAT data set processed with the Pass 8 event-level analysis.
γ-rays from WIMP annihilations can proceed through a variety of SM channels. For the quark and boson channels, the resulting γ-ray spectra are all similar and largely depend on the mass of DM. The leptonic channels have harder spectral energy distributions with a peak in energy flux that is closer to the mass of DM. The Fermi-LAT collaboration performed the analysis for six representative annihilation channels ( ̄bb, τ+τ−, μ+μ−, e+e−, W+W−, and u ̄u), assuming a 100% branching fraction for each of them.
1For the ̄bb channel at 100 GeV.
Figure 3: Distribution of the 27 known Milky Way galaxy satellites (blue points) and of 8 of the DES dwarf galaxy candidates (red) in Galactic coordinates. Figure from .
They find no significant γ-ray excess associated with the Milky Way dwarf galaxies when analysed individually or as a population. Therefore upper limits on the γ-ray fluxes from DM annihilation can be extracted. These can be translated into stringent upper limits on the DM annihilation cross section, which are shown in Fig. 4. Using the new Pass 8 dataset, these bounds show an improvement of a factor 3-5 with respect to the previous analysis .
These bounds (which include the statistical uncertainty on the DM content of the dwarf galaxies) exclude the thermal relic annihilation cross section (∼ 2.2×10−26cm3s−1 ) for WIMPs with masses smaller that 100 GeV annihilating through the b-quark and τ-lepton channels. These results also constrain DM particles with masses greater that 100 GeV, surpassing the best limits from Imaging Atmospheric Cherenkov Telescopes for masses up to 1 TeV.
 M. Ackermann et al. [Fermi-LAT Collaboration], arXiv:1503.02641 [astro-ph.HE].
 M. Ackermann et al. [Fermi-LAT Collaboration], Phys. Rev. D 89, 042001 (2014) [arXiv:1310.0828 [astro-ph.HE]].
 V. Springel, et al., Nature 435 (2005) 629 [astro-ph/0504097].
 V. Springel, et al., Mon. Not. Roy. Astron. Soc. 391 (2008) 1685 [arXiv:0809.0898 [astro-ph]].
 S. E. Koposov, V. Belokurov, G. Torrealba and N. W. Evans, arXiv:1503.02079 [astro-ph.GA].
 K. Bechtol et al. [DES Collaboration], arXiv:1503.02584 [astro-ph.GA].
 B. P. M. Laevens, et al., Astrophys. J. 802 (2015) L18 [arXiv:1503.05554 [astro-ph.GA]].
Text by Valentina De Romeri and Michele Lucente