Implications of the positron/electron excesses on Dark Matter properties

1) The data 2) DM annihilations? 3) γ and ν constraints 4) DM decays?

Alessandro Strumia, talk at IPMU, kaizen-ed to December 10, 2009 From arXiv:0809.2409, 0811.3744, 0811.4153, 0905.0480, 0908.1578, 0912.0742 with M. Cirelli, G. Bertone, M. Kadastik, P. Meade, M. Papucci, M. Raidal M. Taoso, E. Nardi, F. Sannino, T. Volansky, www.cern.ch/astrumia/PAMELA.pdf Indirect signals of Dark Matter

DM DM annihilations in our galaxy might give detectable γ, e+, p,¯ d.¯ The galactic DM density profile

DM velocity: β 10 3. DM is spherically distributed with uncertain profile: ≈ − " α#(β γ)/α r γ 1 + (r /rs) − ρ(r) = ρ α r 1 + (r/rs) r = 8.5 kpc is our distance from the Galactic Center, ρ ρ(r ) 0.38 GeV/cm3, ≡ ≈ DM halo model α β γ rs in kpc Isothermal ‘isoT’ 2 2 0 5 Navarro, Frenk, White ‘NFW’ 1 3 1 20 ρ(r) is uncertain because DM is like capitalism according to Marx: a gravitational system (slowly) collapses to the ground state ρ(r) = δ(r). Maybe our galaxy, or spirals, is communist: ρ(r) low constant, as in isoT. ≈

Moore 3 1000

cm NFW 

Einasto, Α = 0.17 Earth GeV 10 in Burkert Ρ ® isoT 0.1 density DM 0.001 0.001 0.01 0.1 1 10 100 Galacto-centric radius r in kpc DM DM signal boosted by sub-halos?

N-body simulations suggest that DM might clump in subhalos:

Annihilation rate R dV ρ2 increased by a boost factor B = 1 100 a few ∝ ↔ ∼ Simulations neglect normal matter, that locally is comparable to DM. Propagation of e± in the galaxy

2 ∂ Φ + = v +f/4π where f = dN/dV dE obeys: K(E) f (E˙f ) = Q. e e − · ∇ − ∂E 2 1  ρ  dN + Injection: Q = σv e from DM annihilations. • 2 M h i dE δ Diffusion coefficient: K(E) = K0(E/ GeV) RLarmor = E/eB. • 2 ∼ 2 Energy loss from IC + syn: E˙ = E (4σ /3m )(uγ + u ). • · T e B Boundary: f vanishes on a cylinder with radius R = 20 kpc and height 2L. • 2 Propagation model δ K0 in kpc /Myr L in kpc Vconv in km/s min 0.85 0.0016 1 13.5 med 0.70 0.01124 12 max 0.46 0.0765 15 5

min med max

Small diffusion in a small volume, or large diffusion in a large volume? Main result: e± reach us from the Galactic Center only in the max case 1

The data ABC of charged cosmic rays e±, p±, He, B, C... Their directions are randomized by galactic magnetic fields B µG. The info is in their energy spectra. ∼

We hope to see DM annihilation products as excesses in the rarer e+ andp ¯.

Experimentalists need to bring above the atmosphere (with balloons or satel- lites) a spectrometer and/or calorimeter, able of rejecting e− and p.

This is difficult above 100 GeV, also because CR fluxes decrease as E 3. ∼ −

Energy spectra below a few GeV are useless, because affected by solar activity. ∼ p/p¯ : PAMELA

Consistent with background

10-1 ç BESS 95+97 ó ÷óæó÷æóó BESS 98 ó óçòóóòóóò÷óæóóæ óóóóôó ôôó òæ BESS 99 10-2 ó óòôóòô ôò÷ ô ó ò ó í BESS 00 ÷ò æ - ô í Wizard MASS 91 ó ô æò CAPRICE 94

GeV ò í æ CAPRICE 98 sr -3 æò 10 ò AMS-01 98

sec PAMELA 08 2 æ m  1 -4 æ in 10 æ flux p

10-5

10-6 10-1 100 101 102 103 p kinetic energy T in GeV

Future: PAMELA, AMS + + e /(e + e−): PAMELA

30% PAMELA 09 PAMELA is a spectrometer + ò à HEAT 94+95 CAPRICE 94 ò calorimeter sent to space. It ò AMS-01 ò MASS 91 + à ò can discriminate e , e−, p, p,¯ . . . ìì à æ à æ 10% ò and measure their energies up ò ì æ ò à ì æ æ to (now) 100 GeV. Astrophys- æ ò ìæà æ à ì fraction æ æ ò æ æ ì ical backgrounds should give a ææ ìàæ ææ ì æ ææ à positron fraction that decreases Positron with energy. This happens below 3% 10 GeV, where the flux is reduced by the present solar polarity. Growing excess above 10 GeV 1% 10-1 1 10 102 103 Energy in GeV

The PAMELA excess suggest that it might manifest in other experiments: if e+/e continues to grow, it reaches e+ e around 1 TeV... − ∼ − + e + e−: FERMI, ATICs, HESS, BETS

+ These experiments cannot discriminate e /e−, but probe higher energy.

0.03 ò ò

ò ò

ò ææ ò æ ò æ æ ò ò ì æ æ æ æ ò æ ò æ æ ò ò ì ì æ ææ æ ææ ò ò òì æ æ æææ ò ì ì à

sec æ ò æ æ

2 æ æ æ æ æ æ æ æ æ æ æ æ æ à æ cm

 æ æ ì 2 à ì 0.01 FERMI ò GeV

L ì

+ ò e HESS08 à + - e

H à

3 HESS09 E ATIC08 à

à

0.003 ò 10 102 103 104 à Energy in GeV

Hardening at 100 GeV + softening at 1 TeV Are these real features? Likely yes. Hardening also in ATICs.

Systematic errors, not yet defined, are here incoherently added bin-to-bin to the smaller statistical error, allowing for a power-law fit. ... Just a pulsar?

A pulsar is a neutron star with a rotating intense magnetic field. The resulting electric field ionizes and accelerates e (and maybe iron) γ e+e , that are − → → − presumably further accelerated by the pulsar wind nebula (Fermi mechanism).

2 2 2 4 3 Epulsar = Iω /2, E˙pulsar = B R ω /6c = magnetic dipole radiation. • − surface

E/M p The guess isΦ Φ +  e− /E where p 2 and M are constants. • e− ≈ e ∝ · ≈

Known nearby pulsars (B0656+14, Geminga, ?) would need an unplausibly (?) large fraction  of energy that goes into e :  0.3. ± ∼

Test: angular anisotropies (but can be faked by local B(~x), pulsar motion). 2

Model-independent theory of DM indirect detection Model-independent DM annihilations

Indirect signals depend on the DM mass M, non-relativistic σv, primary BR:  W +W , ZZ, Zh, hh Gauge/higgs sector  − DM DM e+e , µ+µ , τ +τ Leptons → − − −  b¯b, t¯t, qq¯ quarks, q = u, d, s, c { } No γ because DM is neutral. Direct detection bounds suggest no Z.

The energy spectra of the stable final-state particles

( ) e±, p∓, ν e,µ,τ , d,¯ γ depend on the polarization of primaries: WL or T and µL or R.

The γ spectrum is generated by various higher-order effects:

γ = (Final State Radiation) + (one-loop) + (3-body) We include FSR and ignore the other comparable but model dependent effects The DM spin

Non-relativistic s-wave DM annihilations can be computed in a model-independent way because they are like decays of the two-body D = (DM DM)L=0 state.

If DM is a fundamental weakly-interacting particle, its spin J can be 0, 1/2 or 1, so the spin of D can only be 0, 1 or 2:

1 1 = 1, 2 2 = 1asymm 3symm, 3 3 = 1symm 3asymm 5symm ⊗ ⊗ ⊕ ⊗ ⊕ ⊕ So:

2 D can have spin 0 for any DM spin. It couples to vectors DFµν and to • 2 higgs Dh , not to light fermions: D`L`R is m`/M suppressed.

D can have spin 1 only if DM is a Dirac fermion or a vector. • + PAMELA motivates a large σ(DM DM ` ` ): only possible for Dµ[`γ¯ µ`]. → − DM annihilations into fermions f

Scalar D can only couple as • 2.0 ¯ DfLfR + h.c. = DΨf Ψf Μ+ with negative helicity ¯ with Ψf = (fL, fR) in Dirac notation. It means zero helicity on average, and 1.5

is typically suppressed by mf /M. Huge Μ+ with positive helicity N

weak corrections if .  M MW L x

 d Vector can couple as  1.0 Dµ N d

• H ¯ ¯ Dµ[fLγµfL] or Dµ[fRγµfR] i.e. fermions with Left or Right helicity. 0.5 + + + Decays like µ ν¯µe νe give e with → dN/dx = 2(1 x)2(1 + 2x) |L − 0.0 0.0 0.2 0.4 0.6 0.8 1.0 dN/dx = 4(1 x3)/3 |R − Positron energy fraction x DM annihilations into W, Z

The effective interactions • 2 DFµνµνρσFρσ and DFµν 1.4

give vectors with T ranverse polarization 1.2 (with different unobservable helicity corre- lations), that decay in ff¯ with E = x M as: 1.0 N 

2 L cos = 3(1 + cos ) 8 x 0.8

dN/d θ θ / d  N d H dN/dx = 3(1 2x + 3x2)/2, 0.6 Transverse vector − 0.4 2 DAµ gives Longitudinal vectors (accon- • 0.2 ting for DM annihilations into Higgs Gold- Longitudinal vector stones), that decay as 0.0 0.0 0.2 0.4 0.6 0.8 1.0 dN/d cos θ = 3(1 cos2 θ)/4 − Fermion energy fraction x dN/dx = 6x(1 x). − Final state spectra for M = 1 TeV

Two-body primary channels: e,µ L, µR,τ L, τR,W L,WT ,Z L,ZT ,h, q,b,t .

e+ p 10-5 e 10-3

WT Μ ZT q b 10-4 WΤL fraction fraction L W ΤR T -6 h Z 10 WL T ZL t ZL proton

h - -5

Positron 10 b

Anti t

10-6 10-7 10 102 103 10 102 103 q Energy in GeV Energy in GeV

Γ d 10 -8 10 t sr h q sec

2 b E 1 m W  1

dlog Z  in Γ -9

T 10 d 

-1 d

dlogN 10 F d ´ T

10-2 10-10 10 102 103 1 10 102 103 Energy in GeV Energy in GeVnucleon

Annihilations into leptons give qualitatively different energy spectra. Anti-deuterium

d¯ forms when DM produces ap ¯ and an ¯ with momentum difference below p0 160 MeV. The analytical appoximation assuming spherical-cow events ≈ 2 dN p3 dN ! d¯ = 0 n,¯ p¯ dT ¯ 3k ¯mp dT d d T =Td¯/2 2 misses the jet structure of events, such that N ¯ 1/M is very wrong. d ∝ Relativity demands that higher M boostsp, ¯ n,¯ d¯, leaving N ¯ constant. d ∼ Running PYTHIA on GRID we find orders of magnitude enhancement:

10-3 MonteCarlo W+W- 10-4 ZZ yield -5 d 10 hh Spherical

Total approximation qq 10-6 bb

10-7 tt 102 103 104 ()DM mass in GeV 3

Implications of the data Fitting procedure

PAMELA and FERMI systematic uncertainties? •

multiply each expected e+, e , p+/p backgrounds times A Epi with free • − − i A and p = 0 0.05, and marginalize over A , p . i i ± i i

solar modulation as uncorrelated uncertainty below 20 GeV: • 6% at 10 GeV, 30% at 1 GeV. ± ±

DM halo: marginalize over isoT/NFW/Moore with flat prior. •

Propagation: marginalize over MIN/MED/MAX with flat prior. • (MED is favored?).

Statistical techniques: as reviewed in appendix B of hep-ph/0606054. • Fittingt PAMELA positron data b

20 h e q Μ Z L 15 L ΜR Τ WT L ZT ΤR 2

Χ 10 q b WL t

5 ΜR WT ΜL WL e ΤL,R ZT

0 ZL 100 300 1000 3000 10000 30000 h DM mass in GeV

If M > TeV everything fits. At smaller M only annihilations into leptons or W . The σv needed for PAMELA

e 106 10-20 ΜL 105 ΜR -21 10 ΤL e 4 sec B 10 ΤR 

-22 3 10 q

3 cm factor 10

in b -23 10 v

Σ t 2 e

Boost 10 B W 10-24 T W 10 L 10-25 ZT 1 ZL 100 300 1000 3000 10000 30000 h DM mass in GeV

σv larger than what suggested by cosmology by a factor Be The cosmological σv

2 Thermal DM reproduces the cosmological DM abundance ΩDMh 0.11 for ≈ σv 3 10 26 cm3/sec around freeze-out, i.e. v 0.2. ≈ × − ∼ up to co-annihilations and resonances. Possible extrapolations to v 10 3: ∼ −

-23 10 103

1.3 2 3 5 10 30 100 1000 10-24 Sommerfeld s- wave 102

-25 out

10 - s-wave L Α sec   V 3 M freeze - H

26  cm 10 10 - M

in Sommerfeld p wave = Ε v  Σ 10-27 Α galaxy 1 our -28 -wave cosmological

10 in p at v v ¬ ¬ -29 10 10-1 0.0003 0.001 0.003 0.01 0.03 0.1 0.3 10-1 1 10 102 103 DM velocity v Α v

The Sommerfeld effect is the quantum analogous of this classical effect: the 2 2 2 sun attracts slower bodies, enhancing its cross section: σ = πR (1+vescape/v )

If DM is thermal PAMELA needs s-wave + Sommerfeld and/or a boost factor (DM in sub-halos has small velocity dispersion: Sommerfeld boosts the boost) Non thermal DM

+ E.g. a wino that with M 100 GeV annihilates into W W − with the correct ≈ T T g4(1 M2 /M2)3/2 σv = 2 − W 2πM2(2 M2 /M2)2 − W Problematic with PAMELAp ¯, reconsidered by Kane et al., excluded by FERMI.

DM with M = 150 GeV that annihilates into W+W-

-1 10 10-2 30% HESS08 ò àPAMELA 08 ATIC08 ò PPB-BETS08 ò ò ìà ò à æ EC òìà æ æ ì 10% òì ííííí í í à ò à æ sec ç ç ì ìà æ 2 ç æ ìò ìæà æ à çí ç ô í à à 10-3 ææ ìææ ç òò òçç à ææìàæ òæ ì cm ò à à æ æ ì  ò ç àòà à àà ææ ç2 íá àààôà ìô à æà à çòáá áá á ç ç í ô æ fraction á ô p PAMELA 08 áá -2 ç í ô ô æ 

GeV 10 3% çá á á ò à p L ìô

+ ô à æ æ æ çáò e á æææ + æ æææææ - ô

Positron -4

e æ á H ô æ 10 ææ ç 3 æ ò E æ æ 1% background? á æ à æ áçá æ background? á -5 0.3% áç 10-3 10 1 10 102 103 104 10 102 103 104 1 10 102 103 104 á ò Positron energy in GeV Energy in GeV p kinetic energy in GeV

á

background? ZL q ZT h

WT

b Fitting PAMELA e+ anti p¯ data

Assuming equal boost & propagation for e+ andp ¯ (otherwise everything goes):

40 e ΤR ΜL

35 ΜR t ΤL ΤL ΤR 30 q 2 Χ ΜL b WL 25 e ΜR t

WT

WL 20 ZT

ZL 15 h 100 300 1000 3000 10000 30000 DM mass in GeV

DM must annihilate into leptons or into W, Z with M > 10 TeV ∼

Indeed a W at rest givesp ¯ with Ep > mp. So a W with energy E = M gives Ep > Mmp/MW , above the PAMELA threshold for M > 10 TeV. + + Fitting PAMELA e and FERMI e + e−

DM annihilation hh

50

45 4Μ-sh bb

40 4e

+ - qq

2 Μ Μ

Χ 35

30 4Μ pulsar with 4Τ 25 F = E-pe-EM Τ+Τ-

20 300 1000 3000 10000 30000 DM mass in GeV

Compatible if DM has few TeV mass and annihilates into some leptons Dark Matter best fit

DM with M = 3. TeV that annihilates into Τ+Τ- with Σv = 1.9 ´ 10-22 cm3s

30% 0.03 ò 10-5 FERMI ò FERMI09 PAMELA 09 HESS08 ò sr ò HESS09ò ò ò ò ì sec sec ò æ ò ò 2 2 ò ì ì -6 ææææææ ò ò ò ì ì òì æææòò à 10 æææææææææææ æ æ 10% æ æ cm brehm

cm à æ  

2 æ ì æ æ æ æ à ì fraction æ æ æ æ 0.01 ò GeV IC æ æ æ æ æ æ æ GeV ì L in -7 ò 10 + à E e d + à  CMB - Γ

Positron 3% e F H d 3 E à 2 10-8 E dust à background? star 1% 0.003 ò 2 3 4 10 102 103 104 10 10 10 à 10 0.01 0.1 1 10 100 1000 Positron energy in GeV Energy in GeV Photon energy in GeV New DM theories

(Neutralinos and standard DM models can hardly fit the e± excesses). DM is charged under a dark gauge group, to get the Sommerfeld enhancement. DM annihilates into the new vector. If light, m < GeV, it can only decay into the lighter leptons. Large σ(DM DM `+`+` `∼) obtained. → − − DM with M = 3. TeV that annihilates into 4Μ with Σv = 8.4 ´ 10-23 cm3s

30% 0.03 ò 10-5 FERMI ò FERMI09 PAMELA 09 HESS08 ò sr ò HESS09ò ò ò ò ì sec sec ò æ ò ò 2 2 ò ì ì -6 ææææææ ò ò ò ì ì òì æææòò à 10 æææææææææææ æ æ 10% æ æ cm

cm à æ  

2 æ ì æ æ æ æ à ì fraction æ æ æ æ 0.01 ò GeV IC æ æ æ æ æ æ æ GeV ì L in -7 ò 10 + à E e

d CMB + à  - Γ

Positron 3% e F H d 3 dust brehm E à 2 10-8 E

à background? star 1% 0.003 ò 2 3 4 10 102 103 104 10 10 10 à 10 0.01 0.1 1 10 100 1000 Positron energy in GeV Energy in GeV Photon energy in GeV

Smoother e± spectrum good for FERMI γ brehmstralung reduced from ln M/m` to ln m/m` γ has a mixing θ with the new light vector, giving a σ(DM N) which is too large if elastic or invisible or consistent with DAMA if inelastic thanks to a ∆M >100 keV splitting among Re DM and Im DM induced by the dark higgs. ∼

Sensitivity to θ, m can be best improved by e beam-dump experiments. 3

Bounds from γ, ν indirect detection Bounds on DM from γ and ν

DM DM `+` is unavoidably accompanied by photons: → −

Bremstrahlung from ` (if ` = τ also τ π0 γγ). • ± → → Largest Eγ M, probed by HESS. ∼

˙ Inverse Compton: e±γ e±γ0 scatterings on CMB and star-light: E uγ. • → 2 ∝ Intermediate E Eγ(Ee/me) 50 GeV being probed by FERMI. γ0 ∼ ∼

˙ 2 Synchrotron: e± in the galactic magnetic fit: E uB = B /2. • 6 ∝ Small Eγ 10 eV, probed by radio-observations: Davies, VLT, WMAP. ∼ − γ from bremstrahlung

2 !2 dΦγ 1 r ρ dNγ Z ds ρ(r) = 2 J σv ,J = dΩ dE 24π MDM h i dE line of sight r ρ − −   NFW Einasto isoT region ∆Ω  5 J ∆Ω = 14700 7600 14 Galactic Center 1 10− h i  2400 3000 14 Galactic Ridge 3 · 10 4 · −

106 Moore 104 NFW 3 2

cm 10 

GeV 1 Einasto, Α = 0.17 Center Ridge pc

in -2

r 10 GC Ρ - GC ® Galactic Galactic FERMI -4 - 10 isoT - - - IR radio Γ Γ Γ & & & & & Earth 10-6 10-6 10-4 10-2 1 102 104 r in pc HESS observations

a M = 10 TeV into W+W-, Galactic Center b M = 1 TeV into Μ-Μ+, Galactic Ridge

10-11 10-6

æ L æ sr -7 L sec æ æ 10 æ 2 æ æ æ æ æ TeV cm  æ æ -8 æ æ æ sec 10 æ 2 TeV æ -12 æ æ 10 cm  in æ æ 1 E -9 æ

d 10 æ in  Γ

E æ N d  d Γ

2 -10 æ N

E 10

æ d æ 10-13 10-11 10-1 1 10 10-1 1 10 Γ energy in TeV Γ energy in TeV

23 3 DM signals computed for NFW and σv = 10− cm /sec. We conservatively impose that no point is exceeded at 3σ: so the 1st example above is allowed.

Other bounds from DM-dominated dwarf spheroidals around the Milky Way. Inverse Compton

1 ! 10"5

"6 I%1 approx Galactic e± diffuse (I = 1) while loosing 5 ! 10 FERMI09 6 Full res. most of their energy as sr 2 sec "6 E 2 e 30 GeV 1 ! 10 eγ e0γ0 E Eγ cm γ0 2 ! → ∼ me ∼ 5 ! 10"7 Initial γ: GeV in dE i) Eγ eV from star-light; ! Γ

∼ # 1 ! 10"7 ii) 0 1 eV from dust rescattering; d

Eγ . 2 ∼ E ! "8 iii) Eγ meV from CMB. 5 10 ∼

ICγ dominate over FSRγ at FERMI E 1 ! 10"8 0.01 0.1 1 10 100 1000 Photon energy in GeV FERMI full-sky observations

Point sources and hadron contamination (around 100 GeV) still present. No clear excess. Robust bounds imposing DM < exp in all sky and energy regions:

IC bound on ΣvHDM DM ® Μ+Μ-L in 10-23cm3sec for M = 1.3 TeV isothermal DM profile with L = 4 kpc 90

110 99 86 65 71 63 84 112

45 33 31 30 33 119 85 50 49 133 156 degrees 20 in 20 14 13 15 17 20 b 10 5 16 15 14 18 0 -5 11 14 10 15 -10 15 19

latitude 11 11 13 9 -20 199 103 51 45 100 188 26 25 27 29

Galactic -45

106 84 64 53 59 66 102 117

-90 -180 -135 -90 -45 -20-10 0 10 20 45 90 135 180 Galactic longitude { in degrees

all bins DM exp 2 X (Φ Φ ) exp global fit: χ2 = i − i Θ(ΦDM Φ ) < 9 Φ2 i i i δ − ν observations

( ) ν µ scattering in the rock below the detector produce trough-going µ± r σv ρ2 3G2 M2p Z 1 dN Φ F ∆Ω 2 ν µ h i 2 J dx x ≈ 8π M παµ · · · 0 dx where p 0.125 is the momentum fraction carried by each quark in the nucleon ∼ and αµ = 0.24 TeV/kmwe = dE/d` is the µ energy loss. − ±

The total µ± rate negligibly depends on the DM mass M.

SuperKamiokande got the dominant bounds in cones up to 30◦ around the GC 2 Φµ < 0.02/cm s Radio observations

104 Around the GC magnetic fields equipartition B B contain more energy than 102 light, diffusion and advection 1

Gauss constant B seem negligible, so all the e in 10-2

± B energy goes into synchrotron E 10-4 radiation. The unknown B only 10-6 determines the maximal νsyn: 10-6 10-4 10-2 1 102 104 r in pc 3 2 dWsyn 2e B ν eBE B  p 2 ( 1) where = = 1 4 MHz δ νsyn 3 . . dν ≈ 3me νsyn − 4πme G me Davies 1976 oservations at the lower ν = 0.408 GHz give the robust and dom- inant bound as the observed GC radio-spectrum is harder than synchrotron:

dWsyn σv Z erg = 2 ( ) ( ( )) 4 2 2 10 16 ν 2 dV ρ E ν Ne E ν < πr − 2 dν 2M 400 cone × · cm s

BIG uncertainty in the DM density ρ at 1pc from the GC: NFW or ...? Bounds from cosmology

DM annihilation rate ρ2 is enhanced in the early universe: its products can ∝

1. affect BBN at T MeV fragmenting 4He, D, 3He. . . ∼ Primordial abundances are not safely known.

2. affect CMB reionizing H after matter/radiation decoupling, z <1000. ∼ 13.6 eV ne uγ ionizes all H changing CMB anisotropies × 

3. heat gas after structure formation z 10. ∼ Depends on unknown non-linear small-scale DM clustering.

1, 2 and 3 give comparable constraints at the PAMELA-level, σv 10 23 cm3/sec. ∼ − 2 is stronger and robust and can be improved by PLANCK. e± signals vs bounds

All at 3σ: region allowed by PAMELA e+ and FERMI e+ + e vs bounds on: − • FSR-γ from FERMI full sky, HESS Galactic Center, Ridge, Dwarf Spheroidals; IC γ for L = 4, 2, 1 kpc; CMB; ν; radio observations of the GC • − • • • dSGC--ΓΓ dS--ΓΓ -Γ + - GC GR DMGCDM-radio® Μ Μ , isothermal profile DM DM ® 4Μ, isothermal profile DM DM ® 4e, isothermal profile GC-radio GC-radio GR-Γ GR-Γ - - 10-20 10 20 10 20

Ν Ν FSR-Γ 10-22 10-22 10-22 sec sec sec    FSR-Γ 3 3 3 FSRIC -Γ IC cm cm cm IC in in

in CMB v v v Σ Σ Σ 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI

freeze-out freeze-out freeze-out 10-26 10-26 10-26 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

e excesses can be DM DM 2µ, 4µ, 4e if ρ is isothermal ± → not if Einasto or NFW -Γ dS-Γ dS DM DM ® Μ+Μ-, Einasto profile DM DM ® 4Μ, Einasto profile DM DM ® 4e, Einasto profile

- - 10-20 10 20 10 20 GC-Γ GR-Γ GC-Γ GC-Γ GR-Γ -22 GR-Γ -22 -22 10 10 Ν 10 sec sec sec FSR-Γ Ν    3 3 3 cm cm cm FSR-Γ in in in CMB FSR-Γ IC v v

v IC Σ Σ Σ IC 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI GC-radio GC-radio GC-radio

freeze-out freeze-out freeze-out 10-26 10-26 10-26 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

DM DM ® Μ+Μ-, NFW profile DM DM ® 4Μ, NFW profile DM DM ® 4e, NFW profile dS-Γ - - 10-20 10 20 10 20 GC-VLT GC-VLT dS-Γ dS-Γ GC-VLT GC-Γ GR-Γ -Γ GCGR-Γ GRGC-Γ -22 Ν -22 10-22 Ν 10 FSR-Γ 10 sec sec sec    3 3 3 FSR-Γ cm cm cm FSR-Γ IC in in

in CMB IC v v v IC Σ Σ Σ 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI 10-24 PAMELA and FERMI

GC-radio freeze-out freeze-out freeze-out GC-radio GC-radio 10-26 10-26 10-26 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

The problem is no longer only at small scales not tested by N-body simulations Caveats

L = 1 kpc at the GC (ok?) would relax NFW or Einasto down to isoT: DM annihilations outside the diffusion volume contribute to FSR, but not to IC:

10 kpc 4 =

L isoT

, 3 NFW to 1 NFW

respect Einasto with

v 0.3 Σ 1 2 3 4 5 7 10 L in kpc

Disavored by a) global fits of charged CR; b) abundances of CR with τ τdiff; ∼ c) FERMI sees γ away from the GC. d) realistic smooth growth of K(z). (Can synchrotron) dominate over IC? Only around the GC. not if τ channels

+ - + - + - DMGCDM-radio® Τ Τ , isothermal profile DM DM ® Τ Τ , NFW profile DM DM ® Τ Τ , Einasto profile dSGC--ΓΓ dS-Γ 10-20 10-20 GC-VLT 10-20 GR-Γ dS-Γ Ν

10-22 10-22 Ν 10-22 IC GC-Γ sec sec sec Ν   -Γ  3 3 GC 3 IC GR-Γ

cm CMB cm CMB cm CMB IC GR-Γ

in FSR-Γ in in v v v

Σ Σ FSR-Γ Σ -24 -24 -24 -Γ 10 PAMELA and FERMI 10 PAMELA and FERMI 10 GC-FSRradio PAMELA and FERMI

freeze-out GC-radio freeze-out freeze-out 10-26 10-26 10-26 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

Too many τ π0 γ: FSR direct exclusion for any reasonable profile. → → not if non-leptonic channels

Non-leptonic channels give many FSR γ and can at most be subdominant: − GC-radio GC-radio+ - DMGCDM-radio® bb, isothermal profile DM DM ® W W , isothermal profile DM DM ® hh, isothermal profile GC-Γ GC-Γ GC-Γ -20 -20 -20 dS-Γ 10 10 -Γ 10 dS-Γ Ν dS GR-Γ Ν GR-Γ GR-Γ PAMELA Ν PAMELA PAMELA IC IC -22 10-22 10-22 IC 10 sec sec sec    3 3 3 cm cm cm -Γ -Γ FSR-Γ FSR

FSR in

in CMB in CMB v v v Σ Σ Σ 10-24 10-24 10-24

freeze-out freeze-out freeze-out 10-26 10-26 10-26 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

The SUSY wino or Minimal Dark Matter no longer can fit PAMELA . 4

DM decays DM decays are compatible with NFW

If instead DM decays with life-time τ, replace ρ2σv/2M2 ρ1/Mτ: → DM ® 4Μ, NFW profile DM ® Μ+Μ-, NFW profile DM ® Τ+Τ-, NFW profile

1027 1027 1027 PAMELA and FERMI PAMELA and FERMI PAMELA and FERMI

FSR-Γ exG-Γ exG-Γ sec sec sec 26 26 26

in 10 in 10 in 10 Τ

Τ Τ exG FSR-Γ -Γ IC-Γ IC-Γ time time time - - - Ν IC-Γ Ν life

Ν life life 1025 1025 1025 DM FSR-Γ DM DM

1024 1024 1024 102 103 104 102 103 104 102 103 104 DM mass in GeV DM mass in GeV DM mass in GeV

With DM decay PAMELA/FERMI are allowed for all DM density profiles DM decays are compatible with cosmology

Weak bounds from BBN and CMB, again due to ρ2(t) ρ1(t). →

The extra-galactic γ flux is significant: DM ® Τ+Τ- with M = 6. TeV and Τ = 5.4 ´ 1025 sec

Φcosmo ρcosmoRcosmo 10-6 1 FSR

Φgalactic ∼ ρ R ∼ sr FERMI

sec and can be computed reliably: no de- 2 IC cm  pendence on small-scale DM clustering. GeV 10-7 in E d 

The ‘exG-γ’ bound on FSR+IC is com- Γ F d 2

petitive, helped by FERMI who already E extracted (?) the diffuse γ flux, a few 10-8 times below the less bright sky. 10-1 1 10 102 103 Photon energy in GeV PAMELA and FERMI as DM decay

4e DM decay

65 Μ+ Μ- 60

55 2

Χ 50 ± ¡ W4ΜΤ 45 W± Μ¡ 40 4Τ Τ+Τ-

35 1000 3000 10000 DM mass in GeV e± excesses suggest SU(2) technicolor!?

DM decays suggests M few TeV, which naturally implies the observed ∼ !3/2 ρDM M M M/Tdec 5 e− ∼ ρb ∼ mp Tdec if the DM density is due to a baryon-like asymmetry kept in thermal equilibrium by weak sphalerons down to Tdec 200 GeV. ∼

Possible if DM is a chiral fermion or is made of chiral fermions. The DM mass is M λv 2 TeV for λ 4π: strong dynamics a-la technicolor. ∼ ∼ ∼ GUT-suppressed dimension 6 4-fermion operators give τ M4 /M5 1026 s. ∼ GUT ∼ If the technicolor group is SU(2) with techni-q = (2, 0) under SU(2) U(1) Q L⊗ Y

DM is a bound state, scalar and SU(2)-singlet as suggested by data. • QQ

A 4-fermion LL¯ operator allows a slowDM `+` : no Π W involved. • QQ → − ' L

Usual problems of technicolor: minimal correction to the S parameter... • Conclusions

The PAMELA, FERMI-ATIC, HESS e± excesses attracted most attention. They could be due to astrophysics or to unexpected DM as follows:

2e channel gave the ATIC peak, not the FERMI e+ + e excess. × −

τ channels give too much γ. ×

W, Z, q, b, h, t channels can only fit PAMELA e+ and give too much γ. ×

3 TeV DM that annihilates in 2µ, 4µ, 4e. But only if the injection term is • quasi constant: i) Isothermal profile; ii) DM decays.

DM predicts that the e+ fraction must grow. DM IC-γ must be in FERMI sky.

Next: FERMI, PAMELA, AMS, PLANCK 4

2 GeV < E < 5 GeV 2 GeV < E < 5 GeV residual (SFD) 90 90

45 45

0 0

-45 -45

-90 -90 180 90 0 -90 -180 180 90 0 -90 -180

5 GeV < E < 10 GeV 5 GeV < E < 10 GeV residual (SFD) 90 90

45 45

0 0

-45 -45

-90 -90 11 180 90 0 -90 -180 180 90 0 -90 -180 SFD Haslam -5 -5 10 GeV < E < 20 GeV 10 GeV < E < 20 GeV residual (SFD) 10 10 90 90

45 45 /s/sr] /s/sr] 2 2 10-6 10-6

0 0 dN/dE [GeV/cm dN/dE [GeV/cm 2 2 -45 -45 The FERMI haze?E 10-7 E 10-7

-90 -90Some theorists claim they see a quasi-spherical ‘FERMI0.1 haze’1.0 excess10.0 :100.0 1000.0 0.1 1.0 10.0 100.0 1000.0 180 90 0 -90 -180 180 90 0 -90 -180 Photon Energy [GeV] Photon Energy [GeV] 20 GeV < E < 50 GeV 20 GeV < E < 50 GeV residual (SFD) Haze Uniform 10-5 10-5 90 90

45 45 /s/sr] /s/sr] 2 2 10-6 10-6

0 0 dN/dE [GeV/cm dN/dE [GeV/cm 2 2 -45 -45 E 10-7 E 10-7

-90 -90 0.1 1.0 10.0 100.0 1000.0 0.1 1.0 10.0 100.0 1000.0 180 90 0 -90 -180 180 90 0 -90 -180 Photon Energy [GeV] Photon Energy [GeV]

Fig. 11.— Correlation coefficients for the templates used in the Type 3 template fit (see 3.2). Upper left: SFD-correlated spectrum Fig. 3.— Residual maps after cross-correlating Fermi maps at various energies with the SFD dust map. The mask is described in 3.2. 0 § § which roughly traces π emission. Upper right: Haslam-correlated emission which traces a disky bremsstrahlung and soft ICS component. Cross-correlations are done over unmasked pixels and for 75 ! 285.FERMIons Although the template suggest removes the much haze of is the due emission, to Loop there is I aLower (a left:SNhaze remnant template-correlated emission. close This to component us) has a notably harder spectrum than both the SFD- and Haslam-correlated ≤ ≤ spectra or their model shapes (dashed lines, cf., Figure 2), indicating a separate component. The dashed lines in the upper two panels clear excess towards the Galactic center. This excess also includes a disky component which is likely due to ICS and bremsstrahlung fromare a GALPROP estimate of π0 (left) decay and bremsstrahlung plus ICS emission (right). Lower right: the uniform template-correlated spectrum which traces the isotropic background. Here the dashed line is the result from Figure 9. This high latitude estimate is higher softer electrons (see Figure 5). than the uniform template estimate likely because the π0 emission is non-zero at high latitudes and leaks into our measured background. This is less of an issue for the uniform template which uses themorphological(i.e.,uniform)information. the ICS signal from the haze electrons. For instance, AGN jets and triaxial DM profiles could (to avoid the GC) and width 30◦ (Regions 1-7) as well as the Galactic center and extending roughly 40◦ in b. both produce signals of this shape. Even approximately one region with 75◦ < ! < 285◦ (Region 8). With these The morphological correlation between the WMAP syn- 3.7. Comments on Haze Morphology spherical production could yield such a signal, should the Although a detailed analysis of the possible source of diffusion be somewhat anisotropic. That said, while one correlation coefficients, we define the residual map to be, chrotron and the RSFD(5 10 GeV) is striking as isthe Fermi haze are beyond the scope of this paper, a might have attempted to invoke, e.g., rapid and signif- − few simple comments are in order. First, the profile is icant longitudinal variation of the magnetic field to ex- shown in Figure 4. Here the 41 GHz synchrotron (hazenot well-described by a disk source. While quantifying plain the microwave haze, such approaches are no longer RT (E)=F (E) cT (E) T. (2) this is a subtle task, the success of the template makes tenable (and moreover already had significant tension − × plus Haslam-correlated emission, see Haslam et al. 1982;this point clear. There are many possibilities to explain with an understanding of the Haslam synchrotron maps). To the extent that the templates in T match the morphol- Dobler & Finkbeiner 2008) is shown side by side with thethe oblong shape, should that persist with greater data. The presence of this feature in both gamma rays as well ogy of the π0 and bremsstrahlung gammas, the residual 5-10 GeV Fermi residual map with the mask used in the map will include only ICS emission. Dobler & Finkbeiner (2008) microwave analysis overlaid Figure 3 shows the resultant residual maps using the for visualization. SFD map as a morphological tracer of π0 emission for Figure 5 shows residual maps using the 1-2 GeV Fermi the Region 8 fits. The most striking feature of the dif- maps as a template for the π0 emission. This sort of ference maps is the extended emission centered around