The Dark Matter Annihilation Signal from Galactic Substructure: Predictions for GLAST

TL;DR

This study assesses the observability of gamma-rays from dark matter annihilation in Galactic subhalos with the GLAST/LAT. By coupling the Via Lactea II high-resolution simulations with analytic extrapolations for unresolved substructure, it derives subhalo boost factors and realistic backgrounds to predict subhalo detectability across a range of WIMP parameters ($M_\chi$ in $50$–$500$ GeV, $igl\u0305igl) angle v$ ~ $10^{-26}$–$10^{-25}$ cm$^3$ s$^{-1}$). The results indicate that GLAST could detect a few to several dozen subhalos, many of which would be spatially extended and resolved, with the exact yield strongly dependent on the subhalo mass function slope, low-mass cutoff, and particle physics. However, considerable uncertainties remain due to halo-to-halo variation, unresolved substructure, and baryonic effects, and the authors emphasize that higher-resolution simulations are needed to obtain converged predictions. Overall, Galactic DM substructure remains a promising GLAST target whose ultimate discovery potential hinges on both particle physics and detailed subhalo statistics.

Abstract

We present quantitative predictions for the detectability of individual Galactic dark matter subhalos in gamma-rays from dark matter pair annihilations in their centers. Our method is based on a hybrid approach, employing the highest resolution numerical simulations available (including the recently completed one billion particle Via Lactea II simulation) as well as analytical models for the extrapolation beyond the simulations' resolution limit. We include a self-consistent treatment of subhalo boost factors, motivated by our numerical results, and a realistic treatment of the expected backgrounds that individual subhalos must outshine. We show that for reasonable values of the dark matter particle physics parameters (M_X ~ 50 - 500 GeV and <sigma*v> ~ 10^-26 - 10^-25 cm^3/s) GLAST may very well discover a few, even up to several dozen, such subhalos, at 5 sigma significance, and some at more than 20 sigma. We predict that the majority of luminous sources would be resolved with GLAST's expected angular resolution. For most observer locations the angular distribution of detectable subhalos is consistent with a uniform distribution across the sky. The brightest subhalos tend to be massive (median Vmax of 24 km/s) and therefore likely hosts of dwarf galaxies, but many subhalos with Vmax as low as 5 km/s are also visible. Typically detectable subhalos are 20 - 40 kpc from the observer, and only a small fraction are closer than 10 kpc. The total number of observable subhalos has not yet converged in our simulations, and we estimate that we may be missing up to 3/4 of all detectable subhalos.

Figures (12)

• Figure 1: Allsky maps of the annihilation signal from VL-II, for an observer 8 kpc from the halo center along the host halo's intermediate principal axis (top two rows) and along its major axis (bottom row). The maps in the left panel show the total annihilation signal from all DM particles within $r_{\rm 200}$, in the right panel only the signal from subhalo particles. In the top row we show the uncorrected signal directly from the simulation, in the bottom two rows the halo center, indicated with a small white ellipse, has been replaced with an artificial $\rho \propto r^{-1}$ cusp, and a mass-dependent boost factor (for $m_0=10^{-6} \,\rm M_\odot$, $\alpha=2.0$) has been applied to the subhalos.
• Figure 2: Sub-substructure in four of VL-II's most massive subhalos. Shown are projections of $\rho^2$ for all particles within a subhalo's outer radius $r_{\rm sub}$. The dashed circle indicates the subhalo's $r_{1000}$. The clumpy sub-substructure boosts the total annihilation luminosity of its host subhalo.
• Figure 3: The annihilation luminosity boost factor due to substructure below VL-II's resolution limit versus subhalo mass, for different subhalo mass functions. Top panel: Dependence on the cutoff mass $m_0$ for slope $\alpha=2.0$. Bottom panel: Dependence on $\alpha$ for $m_0=10^{-6} \,\rm M_\odot$.
• Figure 4: The gamma-ray intensity from DM annihilations in the smooth host halo versus the angle from the Galactic center, directly from the simulated VL-II particles (thick solid line), and with an artificial central cusp with central slope equal to 1.0 (dotted line) and 1.2 (dashed line). For comparison the smooth host halo flux from the VL-I simulation is overplotted (lower thin solid line).
• Figure 5: Diffuse flux due to undetectable subhalos as a function of angle $\psi$ from the Galactic center, for a number of different subhalo mass functions. The thick lines show models with an anti-biased radial distribution, concentrations increasing towards the host center, and different values of the mass function slope $\alpha$ and low mass cutoff $m_0$: $(\alpha,m_0/\,\rm M_\odot)=(2.0,10^{-6})$ (solid), $(2.0,10^{-12})$ (dotted), $(2.0,1)$ (dashed), $(1.9,10^{-6})$ (dot-dashed), $(1.8,10^{-6})$ (triple-dot-dashed). The thin solid line represents the original Pieri2008 model ($B_{\rm ref,z0}$), with $\alpha=2.0$, $m_0=10^{-6}\,\rm M_\odot$, an un-biased radial distribution, and no radial concentration dependence. The flux from the smooth host halo is overplotted with the grey line, see Fig. \ref{['fig:host_flux']}.