Skip to main content

Full text of "Angular Signatures of Annihilating Dark Matter in the Cosmic Gamma-Ray Background"

See other formats


Angular Signatures of Annihilating Dark Matter in the Cosmic Gamma-Ray 











A. Cuoco 1 , J. Brandbyge 1 , S. Hannestad 1 , T. Haugb0Ue 1 ' 2 , G. Miele 3 ' 4 
1 Department of Physics and Astronomy, University of Aarhus, 
Ny Munkegade, Bygn. 1520 8000 Aarhus Denmark 
2 Instituto de Fisica Teorica UAM-CSIC, Universidad Autonoma de Madrid, Cantoblanco, 28049 Madrid, Spain 
3 Universita "Federico II", Dipartimento di Scienze Fisiche, Napoli, Italy & INFN Sezione di Napoli 
4 Instituto de Fisica Corpuscular (CSIC-Universitat de Valencia), 
Ed. Institutos de Investigacion, Apdo. 22085, E-46071 Valencia, Spain 
(Dated: June 19, 2008) 

The extragalactic cosmic gamma-ray background (CGB) is an interesting channel to look for 
signatures of dark matter annihilation. In particular, besides the imprint in the energy spectrum, 
peculiar anisotropy patterns are expected compared to the case of a pure astrophysical origin of 
the CGB. We take into account the uncertainties in the dark matter clustering properties on sub- 
galactic scales, deriving two possible anisotropy scenarios. A clear dark matter angular signature 
is achieved when the annihilation signal receives only a moderate contribution from sub-galactic 
clumps and/or cuspy haloes. Experimentally, if galactic foregrounds systematics are efficiently kept 
under control, the angular differences are detectable with the forthcoming GLAST observatory, 
provided that the annihilation signal contributes to the CGB for a fraction > 10-20%. If, instead, 
sub-galactic structures have a more prominent role, the astrophysical and dark matter anisotropies 
become degenerate, correspondingly diluting the DM signature. As complementary observables 
we also introduce the cross-correlation between surveys of galaxies and the CGB and the cross- 
correlation between different energy bands of the CGB and we find that they provide a further 
sensitive tool to detect the dark matter angular signatures. 

PACS numbers: 95.35.+d, 95.85.Pw, 98.70.Vc 


Astronomical and cosmological observations provide 
overwhelming evidence for the presence of dark matter 
(DM) (see e.g. [1| for a review). In particular, the com- 
bination of various cosmological data sets provides a pre- 
cise measurement of the amount of DM in the universe: 
£l c h 2 ~ 0.11 with a 2a precision of ~ 5% in the mini- 
mal ACDM model @, S [|] and ~ 20% in more extended 
models Q. 

However, despite the noticeable sensitivity to the cos- 
mological abundance of matter (either dark or baryonic), 
such measurements only weakly constrain the properties 
and nature of the particle associated to DM, and very 
weak limits are available on the DM particle mass m x 
and on its couplings. The simplest DM candidate is the 
Weakly Interacting Massive Particle (WIMP) which is 
characterized by having been in thermal equilibrium in 
the early universe (as opposed to for example the ster- 
ile neutrino or super-heavy DM), and having decoupled 
from equilibrium while non-relativistic. In order to get 
the correct DM abundance the mass of such a particle 
cannot be larger than ~ 30 TcV [l], @. On the other 
hand, collider experiments provide a lower bound on the 
mass of ~ 50 — 100 GeV [l(, depending on the specific 
particle candidate. Mass of O(GeV) are however possible 
if more exotic candidates are considered Q- The typical 
thermally averaged DM annihilation cross section in the 
WIMP scenario is <cr x v>~ 10~ 26 cm 3 s _1 However, 
we stress that if the DM is produced out of equilibrium 

in the early universe, no bounds can be given and super- 
massive, GUT scale, DM particles (m x ~ 10 15 GeV) and 
cross sections <a x v> <C 10~ 26 cm 3 s -1 are in principle 

From the point of view of particle physics WIMP can- 
didates are very appealing and emerge naturally in Su- 
persymmetric (SUSY) extensions of the standard model 
or in the Universal Extra Dimensions (UED) model 0]. 
The sensitivity of accelerator experiments, notably the 
Large Hadron Collider, and of direct search experiments 
are approaching the levels required to test the WIMP 
hypothesis, and a direct discovery of DM WIMPs could 
happen in the not so distant future. 

DM WIMP candidates have thus typically a large an- 
nihilation cross section and pair-annihilate into standard 
model particles that subsequently decay and shower pro- 
ducing large numbers of photons and neutrinos. Such 
7-rays from DM annihilation constitute an ideal target 
for astronomical searches. Thus, astrophysical and cos- 
mological observations can provide a crucial test, comple- 
mentary to a direct laboratory detection, in the search 
for the nature of DM particles. Various astrophysical 
environments have been discussed in detail as promising 
sites for observation of DM annihilation, among others 
the galactic center, satellite dwarf galaxies of the Milky 
Way and clumps of DM in the Milky Way halo. In the fol- 
lowing we will focus instead on the all-sky diffuse signal 
expected in the extragalactic cosmic gamma-ray back- 
ground (CGB) SIEflEH. 

Peculiar spectral and angular features can help in dis- 
entangling a signal produced by DM from emission by 


"ordinary" astrophysical sources. The spectrum of pho- 
tons from DM annihilation is in general harder than the 
spectra arising from normal astrophysical processes and 
exhibit a pronounced cutoff at an energy near m x 0, Q . 
The resulting emission thus appear like a "bump" in the 
background astrophysical energy spectrum in the energy 
range in which the DM signal gives a relevant contri- 
bution. However, although this kind of signature would 
constitute a strong hint of DM annihilation, astrophysical 
processes that could mimic such behavior are possible. 

Another signature, which has been widely studied, is 
direct annihilation into a state containing photons, re- 
sulting in a line in the background spectrum that would 
constitute a "smoking gun" signature of DM. However, 
by construction this process is necessarily loop sup- 
pressed and in most models the flux is quite small (see, 
however, [§] for a more thorough discussion). 

Peculiar angular signatures thus offer a complementary 
signature to exclude the remaining degenerate astrophys- 
ical interpretations of a signal. An example is the dumpi- 
ness of DM at sub-galactic scales [ID, [TH, [3 [H| inves- 
tigated by recent zoomed high-resolution N-body simu- 
lations [HI, H3| : Clumpiness would result in a popula- 
tion of high galactic latitude extended gamma emitters 
with a typical annihilating DM gamma spectrum. These 
kinds of objects could hardly be associated to astrophysi- 
cal emitters (but see [HI). In these models the size of the 
clumps is expected to have a characteristic distribution 
and thus the anisotropy of the integrated signal from all 
the clumps also exhibits a characteristic behavior [12]. 

Likewise, the expected angular anisotropics both in the 
case of an astrophysical and of a DM origin of the CGB 
can be calculated, and have received increasing attention 
in the last few years 0, HJ El, H2, H3] . In the following 
we will further pursue this issue addressing the differ- 
ences expected in the two cases and their detectability in 
the light of the improved statistics that will be available, 
when the GLAST observatory is launched and start to 
take data in the near future. We will compare through- 
out the paper our findings in particular with [20l . |22| that 
deal specifically with anisotropies induced by DM anni- 
hilation. Already, there have been claims [T(| HH of a 
DM signal in the CGB as observed by EGRET (see also 
[2lll26[ 127} ). although with the limited EGRET statistics 
and with the uncertainties in the galactic foregrounds, al- 
ternative astrophysical explanations cannot be ruled out. 
On the other hand, with the improved statistics from 
GLAST, a proper analysis of the anisotropy properties 
of the CGB should be able to prove, or disprove, the DM 
interpretation of features in the CGB spectrum. 

Complementary to previous studies we shall employ in 
the following a parametric approach characterizing the 
expected CGB signal in terms of a few key parameters, 
that catch the relevant physical aspects of the problem, 
and varying them in order to asses the robustness and/or 
model dependence of the possible signatures. A further 
advantage of this approach is to make explicit the various 
assumptions employed throughout on which the final sig- 

natures depend. The relevant parameters in the following 
will be the degree of correlation of the CGB sources with 
matter and the absolute normalization of the signal, or, 
equivalently, the expected collected statistics. Further, 
we will also consider complementary anisotropy observ- 
ables like the cross-correlation between surveys of galax- 
ies and the CGB and the cross-correlation between dif- 
ferent energy bands of the CGB. Together with the auto- 
correlations of the CGB these represent a set of indepen- 
dent observables that can be jointly employed improving 
considerably the sensitivity to the DM signal. 

The paper is organized as follows: In section [II] we 
present a discussion of the horizons within which the 
CGB signal is expected to come, relevant for the deter- 
mination of the intensity of the CGB anisotropics itself. 
In section IIIII we introduce the formalism to derive the 
CGB anisotropies in terms of the angular power spec- 
trum. In section llVl we present a forecast for the expected 
statistics from GLAST and we discuss the possibility of 
disentangling the DM annhilation signal from that of as- 
trophysical processes. In sections IVl and IVT1 wc introduce 
the cross-correlation between the CGB and galaxy sur- 
veys and the cross-correlation between different energy 
bands of the CGB and similarly we discuss the different 
behavior and sensitivity in the two cases of interest. In 
section fVIII wc discuss how the previous conclusions ap- 
ply to different possible scenarios for the CGB and DM 
properties. In section [Villi wc summarize and conclude. 


The extragalactic cosmic gamma-ray signal can be pa- 
rameterized as [1, [Til 

J ^' B W dz MWTW ' 


where g{E) = dN^/dE is the photon spectrum of the 
sources, E~ ( is the energy we observe today, p(z,h,r) is 
the matter density in the direction n at a comoving dis- 
tance r, and the redshift z is used as time variable. In 
the following we will interchangeably use p, or p x when 
we want to underline the particle nature of DM. The 
sources are assumed to be distributed proportional to p a . 
The Hubble expansion rate is related to its present z = 
value Hq through the matter and the cosmological con- 
stant energy densities as H{z) = H ox /tt M (l + z) 3 + 
and the reduced Hubble expansion rate h(z) is given by 
H(z) = 100 h(z) km/s/Mpc. We will in the following 
use the parameters of the standard ACDM model 0], i.e. 
Q M = 0.3, fi A = 0.7 and H = 70 km/s/Mpc. The 
quantity r(_E 7 ,z) is the optical depth of photons to ab- 
sorptions via pair production (PP) on the Extra-galactic 
Background Light (EBL). In ref. [l9| an energy threshold 
of E cut = 100 GcV has been considered resulting in a PP 
horizon of about z w 0.5, and a simple extrapolation back 
in time of the present EBL gave a sufficiently accurate 

value of r. In the present work we also consider E cut = 10 
GeV and horizons as large as z w 4 — 5. In this range 
the dynamical evolution of the EBL during the photon 
propagation becomes important for a correct estimate of 
t. To take this into account we use the parametrization 
of t(E 1 , z) from [28| for < z < 5, where evolution effects 
arc included in the calculation. The EBL is expected 
to be negligible at redshifts higher than z w 5 corre- 
sponding to the peak of star formation. Thus, gamma 
photons produced at earlier times experience an undis- 
turbed propagation until z i=a 5, while only in the recent 
epoch they start to loose energy, due to scattering on the 
EBL. Correspondingly, we assume t(E-,, z) = t(E 7 ,5) 
for z > 5 (sec also formula (A. 6) in |19j). 

The case a = 1 is generally representative of astrophys- 
ical sources following the Large Scale Structure (LSS) of 
matter, while the case a = 2 is appropriate for annihi- 
lating DM whose signal follows the square of the matter 
density cx p x through 

<a x v> f°° p x (z,h,r(z)) g[E 7 (l + z)] e~^ z > 
X Snml J Z H(z) (1 + zf 


This last point is however entangled with the exact 
small scale (sub-galactic) clustering properties of DM 
and deserves further discussions. If DM clumps on sub- 
galactic scales, as suggested by various numerical models 
of galaxy formation, or if the DM halo has a very pro- 
nounced spike at its center, the galactic DM signal can 
be greatly enhanced and the overall cosmological contri- 
bution of DM to the CGB would be due to the emission 
from single galaxies. The p x (x) field in Eq.(2) would 
look approximately as a sum of delta functions centered 
on the galaxies' positions and the DM annihilation signal 
would thus trace the matter distribution linearly (actu- 
ally, the galaxy distribution), at least at scales larger 
than the galactic haloes. In this case, however, the DM 
signal expected from the Milky Way itself would prob- 
ably be a more promising observable for signatures of 
DM annihilation, as we will further discuss later. The 
relative contribution of the galactic versus extra-galactic 
DM signal is further discussed in 29}. In principle, if the 
DM clustering properties would be known in the whole 
range from sub-galactic to cosmological scales, the ratio 
of the linear to quadratic contribution can be calculated. 
However, given the still persisting uncertainty in the sub- 
galactic clumping, to be general we will assume a DM an- 
nihilation anisotropy signal 5I X /I X oc P x / p x +£ Px/p~x' 
where £ parameterizes the relative weights of the linear 
and quadratic contributions. In the following we will dis- 
cuss mainly the extreme scenarios ( « 1 and £ ^> 1 in 
which one of the two contributions dominates over the 
other. More precisely, we thus define a "quadratic sce- 
nario" in which DM clustering is relevant only above the 
scale of galactic haloes (~ 10 12 M©), and a "linear sce- 
nario" in which sub-galactic structures dominate the cos- 
mological DM annihilation signal. The mixed scenario is 
discussed further in section IVlTl 


i — i r 

0.1 1.0 10.0 

0.1 1.0 10.0 


FIG. 1: Gamma window functions times the linear growth fac- 
tor D(z) for i?cut = 10, 100 GeV and for various DM masses. 
The curves are normalized to 1 at z = 0. Upper and lower 
curves in each panel refer to the DM annihilation signal corre- 
lating quadratically and linearly (a = 2, 1), respectively, with 

The astrophysical and DM window functions, 
W-y(E-y CUt , z) and W x (E xcut , z), which contain the infor- 
mation about gamma-ray propagation, injection spectra 
and cosmological effects, are defined from Eq. ((T|) as 


I 7 (E Jcut ,h) oc / c\zW 1 (E lcnt ,z) p(z,n) , (3) 


I x (E xcut7 h) oc / dzW x (E xcut ,z)p 2 (z,n), (4) 

where we are using the notation p(z, h, r(z)) = (1+z) 3 x 
p{z, h) to underline that the window function is only 
dependent on the two variables, direction and redshift, 
and to make explicit the (1 + z) 3 behavior of the mat- 
ter density. In principle p s , the density distribution of 
astrophysical sources, should be used in Eq. ©: p s in 
general exhibits a scale and time dependent bias with 
respect to the matter density. However, specific classes 
of astrophysical gamma-ray sources have different biases. 
Blazars, for example, that most likely produce the bulk 
of the CGB signal detected by EGRET, are well known 
to concentrate at the center of clusters of galaxies, thus 


presenting an over-bias with respect to galaxies at high 
densities. On the other hand, galaxies and clusters of 
galaxies quite fairly trace the matter density, at least in 
the recent cosmic epoch. The assumption p s = p for I~ t 
is thus general enough to reasonably describe emission 
from astrophysical sources. 

The window functions can be found from Eq. ([1]) and 
are given by 

W{E mt ,z) = I dE 


g[E{\ + z)} (1 + z) 




where a = 1, 2 applies in the astrophysical and DM cases, 
respectively. When properly normalized W(E cut , z) rep- 
resents the probability of receiving a photon of E 1 > E cut 
emitted at a redshift z. It can be used to define an effec- 
tive horizon, z-h, beyond which the probability of receiv- 
ing a photon is negligible (e.g. < 1%). For i? cu t > 100 
GeV PP losses dominate and the horizon is zt-c < 1 in- 
dependent of the value of a or the shape of g(E). For 
-Ecut 5s 10 GeV, instead, PP losses start to become negli- 
gible (r w 0) and photons propagate freely from arbitrary 
high rcdshifts. However, even in this case a horizon exists 
due to redshifting related this time to the exact shape of 
the injection spectrum g(E) and the value of a. In the 
case of astrophysical sources we take g(E) oc E~ 2 , consis- 
tent with the observed EGRET CGB spectrum and with 
the observed spectra of common astrophysical gamma 
sources like blazars. We found however that for E cut = 10 
GeV the horizon is still mainly settled by the cosmolog- 
ical and PP attenuation effects while the exact shape of 
the spectrum plays a minor role and even choices like 
g(E) cx E^ 1 or g(E) oc E~ 3 change only slightly the 
astrophysical window. Given the poor sensitivity to the 
specific details of the emission spectrum we will thus of- 
ten refer in the following to the term "blazars" , meaning 
in general a representative class of astrophysical gamma- 
emitters tracing linearly the matter density and with 
a power law E~ 2 spectrum. The resulting horizon is 
zu k, 1 as shown in Fig. [TJ The windows are further 
multiplied by the linear growth factor D(z) that takes 
into account the evolution of matter clustering in the 
past (see the next section). D(z) gives a further, al- 
though not crucial, contribution to the determination of 
the exact horizon zu- For E cut = 100 GeV the horizon 
is instead zu « 0.5 and depends exclusively on the EBL 
absorption both in the astrophysical and DM cases. This 
makes this energy range particularly interesting due to its 
limited sensitivity to any particular modelling. Some fur- 
ther effects can in fact contribute to modify the horizon: 
The luminosity of blazars for example can in principle 
change with time due to well known source evolution ef- 
fects introducing a further (l+z) x factor in the window. 
While evolution effects are unimportant for E cut = 100 
GeV, a strong source evolution can in principle affect z-u 
at E cut = 10 GeV. 

In the case of DM, the spectrum g(E) and the cosmo- 
logical factors involved are quite different and the effec- 
tive horizon can be much larger. The different expected 

horizon is in fact an important ingredient for a clear dis- 
crimination through the expected pattern and intensity 
of the anisotropies. A commonly used parametrization 
for the annihilation spectrum of DM is [9( 

exp(-7.76£/A/ x 

(E/M x y 



i.e. a spectrum that is generally harder than the astro- 
physical E~ 2 spectrum, and with a cutoff near the DM 
mass energy (that is the behavior responsible for the 
bump in the overall spectrum). The shape given by 
Eq. ([6]) is almost independent of the details of the an- 
nihilation process, at least for the case of SUSY WIMPs 
where the main contribution comes from decays to qq, 
ZZ. and W + W~ , with subsequent hadronization. A 
slightly different spectrum is expected for the case of 
decay into a lepton-anti-lepton pair or for the annihi- 
lation of UED WIMPs (see Q for details). We will not 
further consider these cases although basically our find- 
ings also apply to them. The resulting windows depend 
on the assumed DM mass, m x , and on the chosen -E cu t. 
Various cases used in the following are shown in Fig. [T] 
At energies above E > 100 GeV the photon absorption 
dominates and, as discussed above, the DM and astro- 
physical horizons are almost identical, z-h w 0.5. For 
^cut = 10 GeV the horizon for astrophysical sources is 
zu ~ 1, while that of DM is generally of order z-u ~ 3 — 4. 
Very interestingly, we can see that the role of absorption 
by the EBL is still quite relevant for DMA at E cut = 10 
GeV limiting the horizon which otherwise would exceed 
z ~ 10 giving much smaller DM anisotropics. Finally, 
for the case of the DM signal correlating linearly with 
matter (a — 1), no appreciable differences are present in 
the windows neither at the high nor at the low energy 
cut. Even if the DM spectral shape is quite peculiar, not 
unexpectedly this seems to play a minor role, as in the 
previously discussed case of astrophysical emission. In 
this case the DM and astrophysical signal have degen- 
erate anisotropy properties and this observable cannot 
further help in disentangling the two contributions. 


A. 3D Power Spectra 

To derive the CGB anisotropies we need first to know 
the spatial clustering properties of the matter field p and 
of its square. To this purpose we use a template of the 
matter distribution derived from a DM N-body simula- 

The N-body simulation was performed with the pub- 
licly available code GADGET-2 [3(| witn 512 3 CDM parti- 
cles in a 128 Mpc/h box. We have assumed a flat ACDM- 
model, with U C dm = 0.30, Q A = 0.70 and h = 0.70 
as well as a scale-invariant primordial power spectrum, 
P cx k. The transfer function was generated using CMB- 
FAST [3l| , and then the initial conditions were computed 




1 1 II llllj 1 1 1 1 llllj 

1 1 1 llllj 1 1 1 1 llllj | 1 1 1 Mil 


N X \ _ 

N \ 




1 1 

i\ ~ 

i i I i i s 

0.001 0.010 

0.100 1.000 
k [h Mpc"'] 

10.000 100.000 

FIG. 2: 3D power spectra of the matter density distribution 
and of its square, as derived from a cosmological N-body sim- 
ulation. Also shown is the linear matter power spectrum, the 
non-linear Halo-model prediction and the cross-correlation 
between the matter distribution and its square. The verti- 
cal dashed line mark the confidence limit on the calculation 
of P(k). All the spectra are normalized at the linear scales to 
the matter power spectrum. 

using second order Lagrangian perturbation theory [32j | . 
The smoothed density field is constructed by interpolat- 
ing the particles to a 2048 3 grid, enforcing mass conser- 
vation, and using the adaptive spline kernel from [33]. 

If p(x) denotes the simulation density field and p(k) its 
Fourier transform, then the matter power spectrum can 
be written as 

Pp(fc) = / d 3 k p*(k)p(k) 


and analogously 

d 3 k 

5a k 

PUk) P 2(k)) 



where p2^k) denotes the Fourier transform of the squared 
density field p 2 (x) and Sai? is a spherical shell of radius k 
and thickness Ak. Finally, it is also possible to estimate 
the cross-correlation spectrum 

P pp 2(k) 

d 3 k 

p* 2 (k)p{k) 


We take into account the time dependence of Pi(k,z) 
(i = p, p 2 , pp 2 ) using the linear growth factor D(z) 

P t {k,z) = P l (k,z = 0)xD 2 (z), (10) 

with D(z) oc h(z) f z °° dz'{l+z')/h 3 (z') and D(0) = l. 1 

In Fig. [2] the various spectra are shown. Notice the 
increase in power at small scales for P p i (k) compared 
to P„(k). For reference P P (k) as calculated in the Halo- 
model [35(] is also shown. It can be seen that the spec- 
tra from the N-body simulation and the Halo-model are 
in quite good agreement. However, the N-body spec- 
trum starts to be affected by numerical noise beyond 
k ~ 20 h Mpc" 1 , shown as a vertical line in the plot, 
and this range is accordingly excluded from the analysis. 
The contribution from higher wave numbers, k > 20 h 
Mpc" 1 , or, equivalcntly, smaller scales, A < 27r/20/i _1 
Mpc, is in any case relevant only for very high multi- 
poles I > 1000 not accessible experimentally, so that for 
the present purposes they can be safely neglected. The 
spectrum of the squared matter distribution is also in 
fair agreement with the Halo-model calculation as de- 
rived in [2(|. The most noticeable feature is an increase 
in the intensity of the anisotropics at the non-linear scales 
k > 1.0 h Mpc" 1 with respect to the matter spectrum, 
understandable in the framework of the Halo-model as a 
dominant contribution from the single-halo term. As ex- 
pected the cross-correlation is in between the matter and 
the matter squared spectra. In the figure all the spectra 
are normalized to the matter spectrum at linear scales, 
while the absolute normalization for the matter squared 
and for the cross-correlation is given by 4 and 2 times 
this value, respectively. 

B. Angular anisotropies 

From Eq.([3])-(|4|) we can now easily construct the angu- 
lar power spectra of the various dimensionlcss fluctuation 
fields SI /I 


c xl 


W 2 (r) P p [ k=-,z(r) 

- z(r) 

W' 2 (r)P p 2 \ k = -,z(r) 

W 2 {r)P p [ k 



for the astrophysical and DM cases following linearly or 
quadratically the matter distribution, respectively. 2 We 
have used the Limber approximation, which is accurate 
for all but the very lowest multipoles. 

In principle the intermediate case of a DM signal oc 
Pxl Px ^ Px/Px can & l so easily be derived, giving a final 




2^x1 + 

(1 + 

2 C X? 


(i + 

2 °X12 


1 D(z) is a good approximation also at non-linear scales where 
P(k,z) grows only slightly faster than the linear growth [34l . 

2 The angular spectra calculations involve an integral over r, 
the comoving distance, while the windows are known in terms 
of the redshift z. We thus use the r-z relation r(z) = 
c/Hq dz'l/ yj fl m (1 + z') 3 + f^Ai an d its inverse z(r). 


1 bill lllllj 1 I llllllj 1 I I lllllj 1 I I lllllj 1 I I lllllj 1 I I I MM 

1 0" 5 1 I I I I I 

10' 10 2 10 3 10 4 10 5 10 6 10 7 
Energy [MeV] 

FIG. 3: EGRET spectrum from [36|] and extrapolation up to 
10 TeV. The solid line shows the expected effect of the PP 

where the p-p 2 spectrum is involved 
C xi2 = J % W x i(r)W x2 (r) P pp2 (k = l -,z(r)^j , (15) 

and < £ < oo weights the relative contribution of the 
linear and quadratic correlation terms. In practice, how- 
ever, in the following we will mainly consider the two 
cases £ = 0, oo, while the intermediate case is easily un- 
derstandable with a qualitative discussion. We will how- 
ever consider quantitatively a mixed scenario in section 


A. Forecast assumptions and sensitivity 

In this section we describe our assumptions to assess 
the sensitivity of the forthcoming gamma-ray detectors, 
in particular GLAST, to the angular signatures in the 
auto-correlation spectrum described in the previous sec- 
tion. Similarly, the sensitivity of the cross-correlation 
observables is discussed in the next sections. 

The observed diffuse gamma emission is constituted 
by the sum of the CGB and of the diffuse galactic emis- 
sion, so that, experimentally, the relevant extragalactic 
signal needs to be separated from the related galactic 
foregrounds. We will assume in the following a perfect 
removal of the galactic gammma foregrounds from the 
CGB. We will thus quote statistical errors only. Indeed, 
foreground separation will be a non trivial issue in the 
analysis of the forthcoming datasets and this is turn is 
expected to propagate to the determination of the CGB 
anisotropics. A detailed analysis of the effects of the 
foregrounds is beyond the scope of this work. We will 
however further discuss this point in section fVIII 

We consider the diffuse energy spectrum as measured 
by EGRET ^ 

/(# 7 ) = fco 





valid from E ~ 10 McV to E ~ 100 GeV, where k = 
(7.32 ±0.34) x lO^cm^s^sr^GeV- 1 , correcting it by 
the effects of EBL absorption as described in section [II] 
We show in Fig. [3] a plot of the EGRET data HI together 
with the fit in Eq. (fT6|) and its extrapolation in the case 
of PP absorption. In agreement with the result of section 
Ullit can be seen again that ~10 GeV is the critical energy 
above which the EBL absorption effects become relevant. 

It is then possible to estimate the number of events, 
N~ n in the relevant energy range to be collected during 
the time t as 

t- dc -n 


dE A eS (E)I y (E) , (17) 


where DC is the duty-cycle of the instrument, flf ov is 
the solid angle of the field of view and A e s(E) is the 
effective collecting area of the instrument (averaged over 
the field of view of the instrument). For GLAST H3 
we will assume a constant A e s(E) = 10 4 cm 2 , DC = 
90%, fifov = 2.4 sr and /f ov = f2f ov /47r. In addition, 
we use the angular resolution of the experiment (<7{, = 
0.115°) and the associated angular window function Wi = 
cxpH 2 a b 2 /2). 

Analogously, the number of photons expected from 
DM annihilation is given by 

N x = t-DC ■ n fov 

dE A eS (E)I x (E) , (18) 


where I X (E) is the DM annihilation spectrum. In our 
parametric approach we calculate the statistics N x nor- 
malizing the DM spectrum to the EGRET spectrum so 
that iV 7 = N x for the relevant energy cut. Later we will 
discuss briefly how the conclusions are affected in the 
case in which the DM signal is reduced to the 20% of the 
EGRET value or in the case in which the CGB itself is 
reduced if parts of it are resolved as sources by GLAST. 

The forecasted error bars on the various CGB angular 
auto-correlation spectra are given by 




l2(l + C Nn /W 2 C 1 ^ 
{21 + l)AZ/fcv ' 

1 2(1 + C N , x /W?Ctf 
(21 + l)A//fov ; 

where C 


^fov/^V 7 and C. 


= n 

fov/iV x 


are the 

gamma and DM random noise levels respectively. 

The resulting spectra and their error bars are shown 
in Fig. U for the case of the pure quadratic scenario. In 
the case of E cnt = 10 GeV DM masses of 100, 500 GeV 
are shown, while for E cut = 100 GeV we consider only 


FIG. 4: Angular spectra for E cut = 10, 100 GeV with 1 a error bars for a 4-year GLAST survey. 

the value 500 GeV, a mass of 1 TeV showing basically 
the same spectrum and error bars. 

The plot on the right (E cut = 100 GeV) shows quan- 
titatively what was anticipated in the previous section. 
The windows are almost identical for the DM and the 
astrophysical cases and the higher intrinsic level of fluc- 
tuations of DM produces a much higher normalization 
in the angular power spectrum. For reasonable values 
of the DM mass the change in the level of anisotropics 
is thus measurable and distinguishable from the astro- 
physical case and provides an important signature of DM 
emission. Further, the shapes of the angular spectra are 
quite different, the DM case giving a further enhance- 
ment of the fluctuations at small scales I > 100 as previ- 
ously found also in [l!| (see in particular Fig. 2). The 
statistics collected above 100 GeV by GLAST in a 4 year 
period (7V 7 f» 10 4 ) is still high enough to allow a more 
than satisfactory measurement and separation of the var- 
ious power spectra. 

In the i? cu t = 10 GeV case there is instead a com- 
petition between the enhanced level of fluctuations and 
their dilution in the wider horizon related to DM. The fi- 
nal normalization of the C/ 's is still greater though, than 
that of blazars although the difference is reduced with 
respect to E cut = 100 GeV. The increased statistics at 
low energy and, more importantly, the different shapes 
of the spectra, however, still make the two contributions 
separable. For E cxlt = 100 GeV, relevant in the case of 
a not too light DM particle m x > 300 GeV, we further 
see that the angular spectrum maintains its diagnostic 
power with the additional advantage that the small hori- 
zon involved, zn « 0.5, considerably reduces the model 
dependence of the signature from cosmological evolution 
or bias effects. 

Finally, an important point to consider is that the as- 
trophysical sources' power spectrum, being almost inde- 
pendent of the energy cut, can be measured at low en- 
ergies, where the collected statistics is high and thus the 
statistical errors are correspondingly small. This calibra- 

tion of the astrophysical signal at low energies can fur- 
ther improve the separation of the two signals especially 
in the case where the DM flux is not at the EGRET 
level but significantly below the CGB flux. The amount 
of separation can be quantified by considering the cross- 
correlation between different energy bins. We will further 
discuss this point in section fVTl 

B. Comparison with previous works 

The above results for -E^it, = 10 GeV are in general in 
good agreement with 2^, 23], confirming the sensitivity 
of the auto-correlation spectrum to the DM signal. In 
particular we confirm that apart the normalization, the 
blazar and DM spectra have a quite different shape with 
the DM case giving much more power to the small scales, 
i.e. for multipoles I > 100. 

In the present work, with respect to [2oT ] we consider in 
much more detail the role of photon absorption showing 
that it is quite relevant also for an energy cut as low as 10 
GeV. We have indeed also improved the treatment of the 
photon absorption process considering the most updated 
results from [281 ] . 

To compare directly our results with that of [2(| it 
should be taken into account that our quadratic model 
shown in Fig. U consider the contribution to anisotropics 
from haloes greater than average galactic haloes, with 
a typical mass of 10 12 Af Q . Ref. [20], instead, consider 
two particular fiducial models with DM clustering un- 
til a sub-halo mass scale of 10 6 M Q and 1O _6 M0. Both 
their model thus consider a certain sub-galactic contribu- 
tion and can be approximately compared to our mixed 
scenario (see section IVII|) with a particular value of the 
mixing parameter £. As expected, indeed, the anisotropy 
spectra in [20T | has a lower normalization corresponding 
to the fact the DM "linear" contribution tends to drag 
the fluctuations to the level of the astrophysical ones. A 
further part of the difference could also arise from our 


improved treatment of the photon absorption although 
in this case, as explained above, a direct comparison is 

case the redshift distribution of the catalogue's galax- 
ies. The related observables in this case are the cross- 
correlation between gamma emission and galaxies and 
the DM-galaxy cross-correlation 


Another observable sensitive to the DM properties can 
be obtained by looking at the cross-correlation between 
the CGB and galaxy catalogues. If the CGB is cosmo- 
logical in origin, clearly a positive cross-correlation is 
expected. Comparing the cross-correlation originating 
from DM annihilation to that of astrophysical emission 
differences are expected, similar to those of the auto- 
correlation spectrum studied in the previous sections. 
The same formalism can be generalized to address these 
differences in detail as we show in the following. Intu- 
itively, the use of the cross-correlation spectrum is a way 
to go beyond the level of the statistical information only 
and the limits imposed by cosmic variance exploiting not 
only the statistical spectrum Ci but also the information 
contained in the whole sky distribution in terms of the 
aim harmonic coefficients [19j. 

Similarly to the CGB we introduce the galaxy intensity 
map of the catalogue 

dz^(z) p g (z,h), 
o dz 


where the galaxy window W g (z) = dnjdz{z) is in this 

w 73 


W^r)W g (r) P p k=-,z(r) 





dr ( I 

^W x i(r)W g (r)P p [k=-,z(r)) , (23) 

C # B = J^W x2 (r)W g (r)P pp , (k= l -,z(r)) -(24) 

As a simplifying hypothesis we again neglect the matter- 
galaxy bias. Notice further that in a galaxy catalogue 
galaxies are observed directly so that the p g from the 
catalogue already contains the redshift evolution and no 
further (1+z) 3 factors are needed. The function W g (z) = 
dn/dz{z) is characteristic of the survey and of its depth 
i.e. the mean observed redshift. In the following we will 
assume the typical shape 

W g (z) = -j-{z) = {z~ -z c ) 2 cxp 



where zq is the mean redshift depth of the survey and z c 
is the low z cutoff. For definitencss we will consider a 
2MASS-likc catalogue with zq = 0.1 and z c = 0. 

The error bars for these observables are this time given 
by a more involved expression 

C 19 
G X9 


{21 + l)AZ/ fov 

73 73 


(21 + l)AZ/ fov 

Ci Cl 

+ cT^T I 1 + C N JW?0 X ){1 + c N , g /wfci) 

X9 X9 



where Cjv,g = fl[ ov /N g is the galaxy random noise, analo- 
gous to Cat i7 and Cn, x , where N g is the number of galax- 
ies in the survey. For the case of the 2MASS survey we as- 
sume f s ky — 0.8 and N g ~ 10 6 . We have further assumed 
that the CGB and galaxy maps have been smoothed to 
the same angular resolution so that the same Wi can be 
used. The use of cross-correlation with galaxies has been 
proven to be a powerful tool in cosmology, in particular in 
the analysis of the Cosmic Microwave Background Radi- 
ation [H, . A cross-correlation with galaxies has also 
been suggested in the study of the MeV gamma back- 
ground [4(|. We refer the reader to these references for 
further details on the formalism employed. 

The results are shown in Fig. [5] with the same assump- 
tions as in Fig. 0] for N 7 and N x . It can be seen that 
in general the galaxy-CGB spectrum is less optimal with 
respect to the auto-correlation spectrum of the CGB it- 
self to look for differences between DM and astrophysical 
sources, but there are still some discerning power. The 
same trend as in Fig. 2] is recognizable: At E cut = 100 
GeV the fluctuations in the DM spectrum are higher with 
respect to the astrophysical case and the statistics and 
angular resolution expected from GLAST can distinguish 
the two cases. At E cut = 100 GeV the balance between 
enhanced DM fluctuations and horizon dilution makes 
degenerate the normalization of the two contributions. 
However, the different shapes at I > 100 still allow to 
disentangle the two cases. Notice that in all cases the in- 
termediate scale multipoles I ~ 100 appear to be optimal 
to disentangle the two cases. 

Although in the single case shown the cross-correlation 
appears to be less sensitive to DM signatures compared 
to the auto-correlation of the CGB, it has to be stressed 
that this is an independent observable and the two can 
be combined and used at the same time improving the 
statistical power of the analysis. Further, different cat- 
alogues can be employed with, possibly, a more suitable 
window that can enhance the sensitivity of the cross- 
correlation. Finally, if the catalogue is sufficiently deep 
(like for example the case of the SDSS main sample and 
the high redshift Luminous Red Galaxies sample 
it can be possible to split the galaxy distribution into 
various redshift bins and perform a tomography analysis 
with different independent cross-correlations. 


The cross-correlation formalism introduced in the pre- 
vious section can be further employed in comparing the 
CGB anisotropics in different energy bands. In partic- 
ular, at the low energy band E < 1 — 10 GeV the DM 
contribution is expected to be negligible. This energy 
range thus represents a natural, high statistics template 
of the astrophysical gamma sky to compare with for the 
higher, E > 10 GeV, DM relevant energy bands consid- 
ered above. 



S< 0.010 


r| 1 — i — i i i 1 1 1 1 1 — i — i i i 1 1 1 1 1 — i — r 

Cross-correlation: (E= 1:10 GeV)-(E cul = 1 GeV) 

m x =500 GeV 
m i( =100 GeV 

I I I I I I 1 1 



0.001 - 


FIG. 6: Cross-correlation between the energy bands 
E = 1 : 10 GeV (1 GeV<i?<10 GeV) and E cut = 10,100 
GeV. The errors refer to a 4-year GLAST survey. 

As an example we plot in Fig. [5] the cross-correlation 
between the energy bands E = 1 : 10 GeV (i.e. 
1 GeV<_E<10 GeV), where the DM contribution is as- 
sumed to be negligible, and E cut = 10, 100 GeV as for 
Figs. Q] and [5] for an astrophysical dominated CGB and 
for a DM dominated CGB for various WIMP masses 
m x . It can be seen that the diagnostic power is sim- 
ilar to that of the auto-correlation of Fig. [4] under- 
standable in the light of the close similarity between 
the i^cut = 10, 100 GeV CGB and the template we are 
comparing with. The cross-correlation between different 
energy bands thus represents a further independent ob- 
servable sensitive to DM signatures. In particular, it acts 
complementary to the auto-correlation spectrum, provid- 
ing an effective, high statistics, calibration of the astro- 
physical background at low energy, thus allowing more 
easily to distinguish the sought DM signal at higher en- 


A. Mixed scenario 

We have seen that a particularly clear signature of DM 
annihilation in the CGB is present in our "quadratic sce- 
nario" . However, a certain contribution from sub-galactic 
clumps and thus a mixing of the linear and quadratic 
scenarios is anyway expected, although, as previously 
discussed, the relative contribution is still quite uncer- 
tain. A contribution to the DM signal from sub-galactic 
clumps is particularly interesting due to the fact that it 
is expected to enhance the overall DM annihilation signal 
of roughly one order of magnitude increasing correspond- 
ingly the chances of detection d, [ID, [l5[ . To give a hint 
of how this contribution affects the previous conclusions 
we show in Fig. [7] the auto-correlation spectrum of DM 
for E cut = 100 GeV and for a m x = 500 GeV WIMP 
in the case in which 80% of the DM signal correlate lin- 
early with matter and only 20% of the DM contribution 
correlate quadratically, the sum being at the level of the 
EGRET flux. Given that DM in the linear scenario is 
almost degenerate with a pure astrophisical emission, an 
equivalent interpretation of Fig. [7] is that of a subdomi- 
nant, 20% level, quadratic DM contribution, and an over- 
all signal dominated by astrophysical emission. We see 
that in both cases the prospects are quite interesting and 
the DM spectrum still differs significantly from that of a 
background generated by astrophysical sources only. 

In the "worst" case, i.e. our "linear scenario" , in which 
sub-galactic clumps dominate the annihilation signal, the 
anisotropies are degenerate with the astrophysical sig- 
nal and the signature in the CGB disappear. In sec- 
ond approximation some difference is still expected due 
to the presence of a bias in the relative distributions of 
DM and astrophysical sources, although the signature be- 
come quite model dependent (see :22j for a more detailed 
discussion). However, in this case, unless our galaxy is 
unrepresentative of an average galactic halo, the best 
chances to detect the DM gamma signal, clearly, would 
come from the Milky Way halo itself for which other kinds 
of anisotropy signatures, due basically to the peculiar 
profile of the galactic halo, are expected (see for example 
[12l.[29j for more details). In this respect, it is interesting 
to notice this sort of complementarity between DM sig- 
natures in the extragalactic cosmological signal and the 
local galactic signal. 

B. CGB normalization 

Part of the population of sources contributing to the 
CGB will likely be resolved by GLAST consequently low- 
ering the level of unresolved emission and thus the inten- 
sity of the CGB. This, indeed, will turn out as an advan- 
tage given that only astrophysical sources are resolved 
and thus the signal to noise ratio for DM is enhanced. 
An estimate in the framework of the blazar model of the 

CGB of [U suggests that GLAST could lower the CGB 
by a factor of 2 [8J] . As an extreme assumption we plot in 
Fig. [5] the error bars in the case in which the CGB (and 
thus the statistics) is reduced by a factor of 5 (i.e. to 20% 
of the present value), assuming the pure quadratic DM 
scenario. We see that even in this case the statistics are 
good enough to separate the two angular spectra. Notice 
that the result is quite conservative given that in the fig- 
ure both the CGB and the DM signal are reduced by a 
factor of 5. 

Although not shown, a very similar result applies for 
the case of a cross-correlation between different CGB 
energy bands that thus equally maintain its diagnos- 
tic power in a low statistics regime. The sensitivity of 
the galaxies-CGB cross-correlation is instead sensibly re- 
duced both in the low and the high energy ranges. Fi- 
nally, if we consider the mixed scenario in the framework 
of this low statistics CGB then the prospects of DM de- 
tection became quite low. A 20% DM quadratic contribu- 
tion in this case would correspond to a 4% contribution 
with respect to the present EGRET intensity, making 
the anisotropy transition signature quite challenging to 

C. Foreground removal 

Finally, we comment on the role of the galactic fore- 
ground on the results. The foreground subtraction re- 
mains a delicate issue, as can be appreciated by the re- 
analysis of the EGRET data performed in [42|, based 
on a revised model for the galactic propagation of cos- 
mic rays that resulted in an appreciably different spectral 
behavior compared to Eq. (|16p and in a slight change in 
the overall normalization. The foreground subtraction is 
also expected to alter and enlarge in a non-trivial way the 
estimate in Eqs. (|19M20[) of the error bars of the angular 
spectrum. A detailed estimate of this effect is beyond 
the scope of the present work. Possibly, however, the 
effect of galactic contamination can be kept under con- 
trol enlarging the galactic cut to higher galactic latitudes 
b > 20° where the galactic emission is expected to rapidly 
decrease [H, [42[ , although at the price of reducing corre- 
spondingly the available statistics. Residual foregrounds 
are anyway expected even at the highest latitudes, at a 
level depending on the foreground model used, making 
non-obvious also this simple first order analysis. An ac- 
curate analysis, further, should eventually rely on a full 
simulation of the data analysis pipeline. 

Further, if, as considered above, dumpiness in the 
Milky Way halo becomes relevant, then, in principle, the 
resulting DM annihilation signal has to be considered as 
a further galactic foreground. CGB extraction, in this 
case, would become more challenging due to the need 
to include in consistent way DM annihilation both in 
the galactic and the extra-galactic signal (see, indeed, 
ref. [27l | where an iterative procedure is applied both 
to the galactic foreground and the extra-galactic back- 




£< 0.100 


0.010 r 

0.001 Lu 

~i — i — i i i 1 1 1 

20% 0^=500 GeV+80% "blazars" 
m.^500 GeV 


FIG. 7: Angular spectra for _E cut = 100 GeV. Shown are the 
cases of CGB dominated by "blazars", CGB dominated by a 
m x = 500 GeV WIMP and CGB contributed by a m x = 500 
GeV WIMP for 20% and by "blazars" for 80%. This latter 
case is degenerate with a CGB contributed entirely by DM 
with a 20% emission tracing the matter quadratically and an 
80% emission tracing the matter linearly. The errors refer to 
the statistics expected from a 4-year GLAST survey. 



0.001 Lu 


FIG. 8: As in Fig. [4] but for a DM and astrophysical signal 
5 times lower (i.e 20% of the present EGRET value) . 

ground for the claim of DM detection in the EGRET 



In the present work we have studied the kind of sig- 
natures that DM annihilation is expected to imprint in 
the anisotropics of the CGB, complementary to the sig- 
natures in the energy spectrum. We have addressed the 
main physical ingredients contributing to the DM signa- 
ture and discussed the robustness of the signature with 
respect to various possible scenarios. We can summarize 
our findings as follows: 

• The DM annihilation signal traces in general the 
matter distribution quadratically due to its p^. de- 
pendence. However, an effective linear correlation 
can arise if the signal is significantly enhanced by 
the presence of cuspy haloes or sub-galactic clumps. 
We have defined the two extreme "linear" and 
"quadratic" scenarios. The first corresponds to the 
case in which the cosmological DM annihilation sig- 
nal is dominated by galactic or sub-galactic struc- 
tures while in the second the signal is dominated 
by emission on scales larger than that of a galactic 
halo. We have chosen a phenomenological approach 
introducing a parameter £ that weights the two rel- 
ative contributions exploring the DM signatures for 
different possible choices of £. 

• The anisotropics are determined both by the in- 
trinsic fluctuations in the source field and by the 
size of the emission horizon zt-l- For E cut > 100 
GeV the horizon z-h is essentially fixed by photon 
absorption in the EBL. The bulk of the gamma- 
rays is expected to originate inside z-h « 0.5, inde- 
pendent of whether they have a DM or an astro- 
physical origin. For E cut > 10 GeV, DM annihila- 
tion in the quadratic scenario has a rcdshift hori- 
zon Zfi ~ 3 — 4. The horizon is still significantly 
limited by PP losses at this energy, otherwise ex- 
ceeding zn w 10. Blazars and DM in the linear 
scenario have degenerate horizons z-h « 1. 

• In the quadratic scenario the DM anisotropy sig- 
nal is sensibly enhanced with respect to blazars for 
E cu t =10 GeV. Further, also the shapes of the an- 
gular spectra differ significantly [20L |22|| . The sig- 
nature remain standing also for E cut = 100 GeV 
despite the decreased statistics and become partic- 
ularly strong, being independent of uncertainties 
related to the blazar-matter bias or to the evolution 
of blazars. This scenario can easily be detected by 
GLAST and would constitute a strong signature of 
DM annihilation. The DM linear scenario, instead, 
exhibit the same level of fluctuations of blazars and 
the two thus have almost degenerate anisotropy fea- 


• The above signature in the angular spectrum re- 
mains quite robust as long as the the quadratic 
DM signal is at least at the 10-20% level with re- 
spect to the linear DM or the blazar component. 
A further uncertainty to take into account is the 
normalization of the CGB (and thus the available 
statistics) that is likely to be reduced if part of the 
sources contributing to the CGB will be resolved 
by GLAST. If the normalization is reduced by an 
extreme factor of 5, (20% of the present EGRET 
value), the pure quadratic DM scenario exhibits 
still a relevant anisotropy transition signature. If 
the quadratic DM contributes for a 20% (thus, 4% 
of the present EGRET value) then the detection of 
the signature becomes quite challenging. 

• The cross-correlation between the CGB and a sur- 
vey of galaxies and the cross-correlation between 
different energy bands of the CGB provide fur- 
ther independent and sensitive observables that can 
be employed in combination with the CGB auto- 
correlation. A joint analysis of all the anisotropy 
observables considerably improves the sensitivity 
to the DM signal and, more in general, the power 
of the statistical analysis. In principle, the exact 
contribution from DM annihilation in sub-galactic 
clumps and cuspy haloes can be treated as free pa- 
rameters (instead of relying on a model) and in- 
ferred from the analysis. 

The above conclusions hold exactly if a perfect clean- 
ing of the galactic foregrounds and a lossless extraction of 
the CGB signal is possible. The analysis of foregrounds 

will be likely the main challenge in the study of the 
CGB. Clearly, given the above shown potential of CGB 
anisotropics in looking for DM signatures, it would be 
worth to perform further detailed studies on the issue. 
The launch of the GLAST satellite is expected by the 
middle of 2008, while the satellite AGILE [H] launched 
in April 2007 is currently already taking data. The im- 
provement in statistics compared to EGRET will allow 
for new, powerful tools to search for exotic contributions 
to the gamma-ray signal. The anisotropy analysis of the 
CGB in particular, if foregrounds contaminations can be 
efficiently kept under control, promises to provide a clear 
signature of DM annihilation or, in the case of a negative 
answer, to obtain new constraints on the DM properties, 
complementary to a pure energy spectrum analysis. 


The authors wish to thank P. D. Serpico for fruitful dis- 
cussions and for valuable comments on the manuscript. 
H. Tu is kindly acknowledged for providing us the Halo- 
model power spectrum. We thank the Danish Centre 
of Scientific Computing (DCSC) for granting the com- 
puter resources used. TH acknowledges partial finan- 
cial support from the Spanish Research Ministry (MEC), 
under the contract FPA2006-05807. GM acknowledges 
supports by Generalitat Valenciana (ref. AINV/2007/080 
CSIC) and by PRIN 2006 "Fisica Astroparticellare: neu- 
trini ed universo primordialc" by the Italian MIUR. 







G. Bertone, D. Hooper and J. Silk, "Particle dark matter: 
Evidence, candidates and constraints," Phys. Rept. 405, 

279 (2005) |arXiv:hep-ph/0404175 
M. Tegmark el al, "Cosmological Constraints from the 
SDSS Luminous Red Galaxies," Phys. Rev. D 74, 123507 
(2006) [arXiv:astro-ph /0608632] . 

D. N. Spergel el al. [WMAP Collaboration], "Wilkin- 
son Microwave Anisotropy Probe (WMAP) three year re- 
sults: Implications for cosmology," Astrophys. J. Suppl. 

170, 377 (2007) [arXiv:astro-ph70 603449 . 
U. Seljak, A. Slosar and P. McDonald, "Cosmologi- 
cal parameters from combining the Lyman-alpha forest 
with CMB, galaxy clustering and SN constraints," JCAP 
0610, 014 (2006) [arXiv:astro-ph /0604335] . 
J. Hamann, S. HannestacT, M. 57 Sloth and 
Y. Y. Y. Wong, "How robust are inflation model and 
dark matter constraints from cosmological data?," Phys. 
Rev. D 75, 023522 (2007) [arXiv:astro-ph/0611582] . 
K. Griest and M. Kamionkowski, "Unitarity Limits on 
the Mass and Radius of Dark Matter Particles," Phys. 
Rev. Lett. 64, 615 (1990). 

J. F. Gunion, D. Hooper and B. McElrath, "Light neu- 
tralino dark matter in the NMSSM," Phys. Rev. D 73 

[8] P. Ullio, L. Bergstrom, J. Edsjo and C. G. Lacey, 
"Cosmological dark matter annihilations into gamma- 
rays: A closer look," Ph ys. Rev. D 66, 123502 (2002) 
[arXiv:astro-p h/0207125| . 
[9] L. Bergstrom, J. Edsjo and P. Ullio, "Spectral gamma-ray 
signatures of cosmological dark matter annihilations," 
Phys. Rev. Lett. 87, 251301 (2001) | astro-ph/0105048l . 

[10] D. Elsaesser and K. Mannheim^ "Supersymmet- 
ric dark matter and the extragalactic gamma ray 
background," Phys. R ev. Lett. 94 (2005) 171302 
[arXiv:astro-ph /0405235| . 

[11] D. Elsaesser and K. Mannheim, "Cosmological gamma 
ray and neutrino backgrounds due to neutralino dark 
matter annihilation," A stropart. Phys. 22 (2004) 65 
[arXiv:astro-ph/0405347| . 

[12] L. Pieri, G. Bertone and E. Branchini, "Dark Matter 
Annihilation in Substructures Revised," arXiv:0706.2101 
[astro- ph]. 

[13] J. E. Taylor and J. Silk, "The dumpiness of cold 
dark matter: Implications for the annihilation sig- 
nal," Mon. Not. Roy. Astron. Soc. 339 (2003) 505 

(2006) 015011 [arXiv:hep-ph/0509024| . 



Berezinsky, V. Dokuchaev and Y. Eroshenko, "De- 
struction of small-scale dark matter clumps in the hier- 










archical structures and galaxies," Phys. Rev. D 73 (2006) 
063504 [arXiv:astro-ph/0511494l . 

L. Bergstrom, J. Edsjo, P. Gondolo and P. Ullio, 
"Clumpy neutralino dark matter," Phys. Rev. D 59 
(1999) 043506 [arXiv:astro-ph/9806072| . 
J. Diemand, M. Kuhlen and P. Madau, "Dark mat- 
ter substructure and gamma-ray annihilation in the 
Milky Way halo," Astrophys. J. 657 (2007) 262 

arXiv:astro-ph /0611370 

J. Diemand, B. Moore and J. Stadel, "Earth-mass dark- 
matter haloes as the first structures in the early uni- 
verse," Nature 433 (2005) 389 [arXiv:astro -ph/0501589] . 

E. A. Baltz, J. E. Taylor and L. L. Wai, "Can Astrophys- 
ical Gamma Ray Sources Mimic Dark Matter Annihila- 
tion in Galactic Satellites?," arXiv:astro-ph/0610731 1 
A. Cuoco, S. Hannestad, T. Haugbolle, G. Miele, 
P. D. Serpico and H. Tu, "The Signature of Large Scale 
Structures on the Very High Energy Gamma-Ray Sky," 
JCAP 0704 (2007) 013 [arXiv:astro-ph/0612559] . 

S. Ando and E. Komatsu^ "Anisotropy of the cosmic 
gamma-ray background from dark matter annihilation," 
Phys. Rev. D 73, 023521 (2006) [astro-ph/0512217] . 
S. Ando, E. Komatsu, T. Narumoto and T. Totani, "An- 
gular power spectrum of gamma-ray sources for GLAST: 
blazars and clusters of galaxies," Mon. Not. Roy. Astron. 
Soc. 376 (2007) 1635 [arXiv:astro-ph/0610155] . 
S. Ando, E. Komatsu, T. Narumoto and T. Totani, 
"Dark matter annihilation or unresolved astrophysical 
sources? Anisotropy probe of the origin of cosmic 
gamma-ray background," Phys. Rev. D 75 (2007) 063519 
[arXiv:astro-ph/061246"7] . 

F. Miniati, S. M. Koushiappas and T. Di Matteo, "Angu- 
lar Anisotropies in the Cosmic Gamma-ray Background 
as a Probe of its Origin," |arXiv:astro-ph/0702083 

W. de Boer, C. Sander, V. Zhukov, A. V. Gladyshev and 
D. I. Kazakov, "EGRET excess of diffuse galactic gamma 
rays as tracer of dark matter," Astron. Astrophys. 444 
(2005) 51 [arXiv:astro-ph/0508617] . 

L. Bergstrom, J. Edsjo, M. Gustafsson and P. Salati, "Is 
the dark matter interpretation of the EGRET gamma ex- 
cess compatible with antiproton measurements?," JCAP 
0605 (2006) 006 [arXiv:astro-ph/0602632] . 
W. de Boer, I. Gebauer, C. Sander, M. Weber and 
V. Zhukov, "Is the dark matter interpretation of the 
EGRET gamma ray excess compatible with antipro- 
ton measurements?," AI P Conf. Proc. 903 (2007) 607 
arXiv:astro-ph/0612462] . 

W. de Boer, A. Nordt, C. Sander and V. Zhukov, "A 
new Determination of the Extragalactic Background of 
Diffuse Gamma Rays taking into account Dark Mat- 
ter Annihilation ," Astron. Astrophys. 470 (2007) 61, 
larXiv:0705.0094l [astro-ph] . 

F. W. Stecker, M. A. Malkan and S. T. Scully, "Inter- 
galactic photon spectra from the far IR to the UV Ly- 














man limit for < z < 6 and the optical depth of the uni- 
verse to high energy gamma-rays," Astrophys. J. 648, 
774 (2006) [astro-ph/05 10449] . See also F. W. Stecker, 
M. A. Malkan and S. T. Scully, "Corrected Table for the 
Parametric Coefficients for the Optical Depth of the Uni- 
verse to Gamma-rays at Various Redshifts," Astrophys. 
J. 658 (2007) 1392 [arXiv:astro-ph/0612048] . 
D. Hooper and P. D. Serpico, "Angular Signatures of 
Dark Matter in the Diffuse Gamma Ray Spectrum," 


JCAP 0706 (2007) 013 [arXiv:ast ro-ph/0702328 
V. Springel, "The cosmological simulation 
GADGET-2, " Mon. Not. Roy. A stron. Soc 
(2005) 1105 [arXiv:astro-ph/0505010] . 
U. Seljak and M. Zaldarriaga, "A line of sight approach to 
cosmic microwave background anisotropies," Astrophys. 
J. 469 (1996) 437 [astro-ph/9603033] . 
M. Crocce, S. Pueblas and R. Scoccimarro, "Transients 
from Initial Conditions in Cosmological Simulations," 

Monaghan, J.J. and Lattanzio, J.C., Astron. Astrophys. 
149 (1985) 135. 

A. Jenkins et al. [Virgo Consortium Collaboration], 
"Evolution of structure in cold dark matter universes," 
Astrophys. J. 499 (1998) 20 [arXiv:astro-ph/9709010] . 
R. E. Smith et al. [The Virgo Consortium Collaboration] , 
"Stable clustering, the halo model and nonlinear cosmo- 
logical power spectra," Mon. Not. Roy. Astron. Soc. 341 
(2003) 1311 [astro-ph/0207664] . 

P. Sreekumar et al. [EGRET Collaboration], "EGRET 
observations of the extragalactic gamma ray emission," 
Astrophys. J. 494, 523 (1998) [astro-ph/9709257|. 
See the URL: 
N. Afshordi, Y. S. Loh and M. A. Strauss, "Cross- 
Correlation of the Cosmic Microwave Background with 
the 2MASS Galaxy Survey: Signatures of Dark Energy, 
Hot Ga s, and Point Sources," P hys. Rev. D 69 (2004) 
083524 [arXiv:astro-ph /0308260] . 

A. Cabre, E. Gaztanaga, M. Manera, P. Fosalba and 
F. Castander, "Cross-correlation of WMAP 3rd year and 
the SDSS DR4 galaxy survey: new evidence for Dark En- 
ergy," Mon. Not. Roy. As tron. Soc. Lett. 372 (2006) L23 
[arXiv:astro-ph /0603690] . 

P. J. Zhang and J. F. Beacom, "Angular Correlations of 
the MeV Cosmic Gamma Ray Background," Astrophys. 
J. 614 (2004) 37 [arXiv:astro-ph/040135l| . 
F. W. Stecker and M. H. Salamon, "The Gamma-Ray 
Background from Blazars: A New Look," Astrophys. J. 
464 (1996) 600 [arXiv:astro-ph/9601120] . 
A. W. Strong, I. V. Moskalenko and O. Reimer, "A 
new determination of the extragalactic diffuse gamma- 
ray background from EGRET data," Astrophys. J. 613, 

956 (2004) [astro-ph/040544T]. 

See the URL: http : / /agile . rm . iasf . cnr . it/