Constraining Dark Matter with Background Light Sam McDermott Dec - - PowerPoint PPT Presentation

Constraining Dark Matter with Background Light

from 1309.4091, with Rouven Essig, Eric Kuflik, Tomer Volansky, and Kathryn Zurek and 1312.0608 with Ilias Cholis and Dan Hooper

Sam McDermott Dec 5, LANL

Prelude

LUX sees nothing

Is this really a big problem? See e.g.: Cirigliano, Graesser, Ovanesyan, and Shoemaker 1311. 5886; Gresham and Zurek 1311.2082

• More and more of our favorite DM parameter

space seems to be getting ruled out

• (Infinitely) large swaths of well-motivated DM

parameter space are currently up for grabs

• We need complementary probes to
• test assumptions of bounds
• explore different parameter space
• This talk will focus on indirect detection

Motivation

• Photons (produced directly, from FSR, from

cascades, etc.) are generic in DM annihilation and decay

• Data are “just sitting there” ready to use

Motivation

This talk two very different methods

• f approaching this problem

Outline

• Light decaying dark matter:
• models
• statistical methodology
• results
• “Light” annihilating dark matter:
• isotropic gamma rays from DM
• astrophysical backgrounds
• results

Part I: Light (MeV-ish) Decaying DM

“Light” Dark Matter

• still cold
• not warm, not an ALP that forms a

galactic scale BEC, etc.

• few keV to few GeV mass
• we assume that all (model-dependent)
• bservables of standard cosmology are

taken care of (i.e., asymmetric or thermal production where appropriate)

Data

• “Diffuse” X-Rays and Gamma-Rays
• HEAO-1 (1977)
• INTEGRAL (2008)
• COMPTEL (1998)
• EGRET (2003)
• Fermi (2012) (21 months)
• Some observations near the galactic

poles, some near the galactic center

Data

0.01 0.1 1 10 102 103 104 10-3 0.01 Eg @MeVD Eg2dFêdEg @MeV cm-2 s-1 sr-1D

HEAO-1: {ŒH58,109L‹H238,289L,»b»ŒH20,90L INTEGRAL:»{»ŒH0,30L, »b»ŒH0,15L COMPTEL: »{»ŒH0,60L, »b»ŒH0,20L EGRET: {ŒH0,360L, »b»ŒH20,60L FERMI: {ŒH0,360L, »b»ŒH8,90L

Analysis methods

• There are many different methods:
• spectral fit plus power law in sliding

energy window (cf. Weniger)

• precise background modeling

(cf. Siegal-Gaskins; newer work here; others)

• “on-off” or template analyses

(cf. Koushiappas + Geringer-Sameth; Finkbeiner + Slatyer; Hooper + Slatyer; Tempel, Hektor, + Raidal; others)

Analysis methods

• There are many different methods:
• spectral fit plus power law in sliding

energy window (cf. Weniger)

• precise background modeling

(cf. Siegal-Gaskins; newer work here; others)

• “on-off” or template analyses

(cf. Koushiappas + Geringer-Sameth; Finkbeiner + Slatyer; Hooper + Slatyer; Tempel, Hektor, + Raidal; others)

requires exceptional energy resolution

Analysis methods

• There are many different methods:
• spectral fit plus power law in sliding

energy window (cf. Weniger)

• precise background modeling

(cf. Siegal-Gaskins; newer work here; others)

• “on-off” or template analyses

(cf. Koushiappas + Geringer-Sameth; Finkbeiner + Slatyer; Hooper + Slatyer; Tempel, Hektor, + Raidal; others)

requires detailed knowledge of astrophysics

• ver very wide

ranges of position and energy space

Analysis methods

• There are many different methods:
• spectral fit plus power law in sliding

energy window (cf. Weniger)

• precise background modeling

(cf. Siegal-Gaskins; newer work here; others)

• “on-off” or template analyses

(cf. Koushiappas + Geringer-Sameth; Finkbeiner + Slatyer; Hooper + Slatyer; Tempel, Hektor, + Raidal; others)

requires exceptional angular resolution

Analysis methods

• There are many different methods:
• spectral fit plus power law in sliding

energy window (cf. Weniger)

• precise background modeling

(cf. Siegal-Gaskins; newer work here; others)

• “on-off” or template analyses

(cf. Koushiappas + Geringer-Sameth; Finkbeiner + Slatyer; Hooper + Slatyer; Tempel, Hektor, + Raidal; others)

• Only direct photon production and primary

FSR; no ICS, synchrotron, etc.

• We simply require (for every energy bin):
• Robust results!

Fluxpredicted ≤ Fluxobserved + 2 × Error Barobserved

Our analysis (1309.4091)

Fluxpredicted ≤ Fluxobserved + 2 × Error Barobserved

Our analysis (1309.4091)

• Only direct photon production and primary

FSR; no ICS, synchrotron, etc.

• We simply require (for every energy bin):
• Robust results!

Can we exclude this?

10.0 5.0 3.0 7.0 1.0 10.0 5.0 2.0 20.0 3.0 1.5 15.0 7.0 EΓ @arb. unitsD FluxΓ @arb. unitsD

Power law background Signal

10.0 5.0 3.0 7.0 1.0 10.0 5.0 2.0 20.0 3.0 1.5 15.0 7.0 EΓ @arb. unitsD FluxΓ @arb. unitsD

With background model, sure

10.0 5.0 3.0 7.0 1.0 10.0 5.0 2.0 20.0 3.0 1.5 15.0 7.0 EΓ @arb. unitsD FluxΓ @arb. unitsD

Power law background Signal

But what if the background does something crazy?

With background model, sure

“Conspiracy” background Signal

10.0 5.0 3.0 7.0 1.0 10.0 5.0 2.0 20.0 3.0 1.5 15.0 7.0 EΓ @arb. unitsD FluxΓ @arb. unitsD

Being conservative: Not ruled out!

10.0 5.0 3.0 7.0 1.0 10.0 5.0 2.0 20.0 3.0 1.5 15.0 7.0 EΓ @arb. unitsD FluxΓ @arb. unitsD

“Conspiracy” background Signal

Being conservative: Not ruled out!

We are being robust at the cost

• f ruling out less

parameter space

Photons from DM decays

Galactic:

• vs. extragalactic:

Galactic dominates, marginally:

dΦγ,G dE = r 4π ρ mDM ΓdNγ dE J(Ω) dΦγ,EG dE = Ω 4π ΓΩDMρc mDMa0H0 Z ∞ dz dN dE(z) 1 p ΩΛ + Ωm(1 + z)3 ρrJ(Ω) ' O(105 GeV3) vs. ρDM/H0 ' 5 ⇥ 106 GeV3

but same order of magnitude

Models of decaying DM

• Hidden Photino – SUSY + hidden U(1). [Higgs a hidden

U(1); break SUSY with messengers from SM; hidden photon/photino with small mass splitting; kinetically mix hidden photon and photon of U(1)EM. Decays involving SM possible if there is a light gravitino. Topology depends on relative masses of hidden photon/photino.]

• Sterile Neutrino – long-lived sterile neutrino
• RPV Gravitino – neutrino/photino mixing. [Planck-scale

suppression gives a naturally small rate for gravitino decays. Fastest decay is gravitino > photon + neutrino.]

• Plus dipole DM, dark scalar, dark pseudoscalar

So the name of the game is...

Particle physics enters through and only: is fixed by decay topology is fixed by the model Ignore the astro/cosmology stuff (be conservative)

Γ

dN/dEγ dN/dEγ

Γ

Dark photino model

⌧e

γd→γ e G ' 3 ⇥ 1023 sec

✓10−8 ✏ ◆2 ✓10 MeV me

γd

◆5 p F 100 TeV !4

τe

γd→γd e G ' 3 ⇥ 1020 sec

✓1 MeV me

γd

◆5 p F 104 TeV !4 1 m2

γd

m2

DM

!−4

Two different decays, depending on whether the dark photon or dark photino is more massive: MeV scale comes out naturally:

m2

e γd = ✏ · gd hDY i ' (5 MeV)2 ⇣

✏ 10−8 ⌘ ⇣ gd 0.2 ⌘ p hDY i 50 GeV !2

Dark photino DM,

⌧e

γd→γ e G ' 3 ⇥ 1023 sec

✓10−8 ✏ ◆2 ✓10 MeV me

γd

◆5 p F 100 TeV !4

Short- Lived

0.1 1 10 102 103 104 10-12 10-10 10-8 10-6 10-4 0.01 mg

é

d @MeVD

e g é

dÆgG

é

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

me

γd < mγd

√ F = 104 TeV

me

γd > mγd

τe

γd→γd e G ' 3 ⇥ 1020 sec

✓1 MeV me

γd

◆5 p F 104 TeV !4 1 m2

γd

m2

DM

!−4

Short- Lived mg

é

d<mG

é +mgd

1 10 102 104 105 106 mg

é

d @MeVD

F @TeVD g é

dÆgdG

é Æ f + f -G é

HEAO-1 INTEGRAL COMPTEL EGRET

Dark photino DM,

Sterile neutrino DM

Three-body and radiative decays contribute to photon background at similar levels: mixing angle between and

νs νe

τνs→νγ ' 7.2 ⇥ 1017 sec ✓10 MeV mχ ◆5 ✓sin2 2θ 10−8 ◆−1

τνs→ναe+e− ' 9.6 ⇥ 1015 sec ✓10 MeV mχ ◆5 ✓sin2 2θ 10−8 ◆−1

Sterile neutrino DM

Relic abundance is model-dependent, but a UV-insensitive contribution comes from late-time oscillations (Dodelson-Widrow mechanism) Bounds are different depending on whether or .

Ωνs = ΩDW Ωνs ≥ ΩDW

Sterile neutrino DM

Ly-a Short- Lived Ws>WDM CMB BBN SN1987A

0.01 0.1 1 10 102 10-32 10-30 10-28 10-26 10-24 10-22 10-20 10-18 10-16 10-14 10-12 10-10 10-8 mns @MeVD sin22q ns Hradiative and three-body decaysL Wns≥WDW

HEAO-1 INTEGRAL COMPTEL EGRET

Ly-a Short- Lived Ws>WDM CMB BBN SN1987A

0.01 0.1 1 10 102 10-24 10-22 10-20 10-18 10-16 10-14 10-12 10-10 10-8 mns @MeVD sin22q ns Hradiative and three-body decaysL Wns=WDW

HEAO-1 INTEGRAL COMPTEL EGRET

very strong!

RPV gravitino DM

Short- Lived

0.1 1 10 102 103 104 10-8 10-7 10-6 10-5 10-8 10-7 10-6 10-5 10-4 10-3 0.01 0.1 1 m3ê2 @MeVD Ug

é n

G é Æng

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

×

• G

γ

• γ

ν

τ e

G→νγ =

1 32π |Ue

γν|2 m3 3/2

m2

Pl

!−1 ' 3.8 ⇥ 1028 sec ✓10 MeV m3/2 ◆3 ✓10−4 Ue

γν

◆2

Dipole DM

dimension 5 operator can be strongly constrained L ⊃ λ Λ ¯ χ2σµνχ1Fµν

τdipole ' 4.1 ⇥ 1020 sec ✓10 MeV m1 ◆3 ✓ Λeff 1019 GeV ◆2

(Λeff ≡ Λ/λ)

Short- Lived

0.01 0.1 1 10 102 103 104 1017 1018 1019 1020 1021 1022 1023 1024 1025 m1 @MeVD Leff @GeVD Dipole DM

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

Dark (pseudo)scalars

τπd→γγ ' 1.1 ⇥ 1020 sec ✓10 MeV mπd ◆3 ✓ fπd 1015 GeV ◆2 τφ→e+e− ' 8.3 ⇥ 1018 sec10 MeV mφ ✓10−20 ga ◆2

Short- Lived

1 10 102 103 104 10-25 10-24 10-23 10-22 10-21 10-20 mf @MeVD g fÆe+e-+FSR

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

Short- Lived

0.01 0.1 1 10 102 103 104 1014 1015 1016 1017 1018 1019 1020 1021 1022 mpd @MeVD fpd @GeVD pdÆgg

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

Part 1 Halftime

• Those were the model-dependent bounds
• bounds on model-specific parameters (mixing

angles, decay constants, etc.)

• very strong for dimension<6,

non-Planck-suppressed operators

• About to show model-independent bounds
• just the lifetime – mass plane from now on
• lifetime bounds from 6 (FSR photons) to 10

(direct photons) orders of magnitude stronger than 1/H0

Photon line

0.01 0.1 1 10 102 103 104 1026 1027 1028 mf @MeVD t @secD fÆgg

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

τUniv × 1010

e+ e- (FSR)

1 10 102 103 104 1022 1023 1024 1025 mf @MeVD t @secD fÆe+e-+FSR

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

τUniv × 107

e+ e- (FSR), boosted

τUniv × 107

1 10 102 103 104 1022 1023 1024 1025 m1 @MeVD t @secD f1Æf2f3Æf2 e+ e-

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

e+ e- (FSR), three-body

1 10 102 103 104 1022 1023 1024 1025 1026 m1 @MeVD t @secD c1 Æ c2 e+ e- + FSR

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

τUniv × 108

three-body, with photons

τUniv × 1010

0.01 0.1 1 10 102 103 104 1026 1027 1028 m1 @MeVD t @secD f1 Æ f2 g g

HEAO-1 INTEGRAL COMPTEL EGRET FERMI

Part 1, Recap

• The decaying DM bounds:
• model-specific parameters
• model-independent parameters

(lifetime, mass)

• (very robust)

Part II: “Light” (GeV-ish) Annihilating DM

Part 1I, Outline

Annihilating DM:

• effect of structure on DM annihilations
• relative contributions
• astrophysical backgrounds
• results

Where does DM annihilate?

First guess: mostly in the MW halo smooth galactic: smooth extragalactic: ρ2

srs ' O(10−47 GeV7)

vs. ρ2

DM/H0 ' O(10−52 GeV7)

Structure

Structures (overdensities) dramatically change this picture because the rate of annihilations scales quadratically with density (not linearly, as for decays) Where the DM is matters for annihilations!

Structure

ρcrit ' 5.5 ⇥ ρav

Critical density with which a sphere collapses instead

• f growing indefinitely:

Final density after virializing:

δ ⌘ ρstructure/ρav ' 32 · 5.5 ' 178

ρ2

eg ' δ2ρ2 DM/H0 ' O(few ⇥ 10−48 GeV7) ' ρ2 srs

ρf = 32 × ρi

Improved estimate:

• The smooth galactic and clumpy extragalactic

intensities are roughly similar in magnitude

• this makes sense: the MW is (mostly) a

DM halo!

• But there is additional structure:
• clusters at all redshifts have subclusters

(analogously: the MW has subhalos)

• Add it all up!

Different contributions

satellite mass function

Milky Way

dIsm dE = hvi 2m2

χ

dNγ dE Z

Vobs

dVMW ⇢2(s, b, `) 4⇡s2

(from satellites) (from smooth distribution) (uncertain)

dIsat dE = hvi 2m2

χ

dNγ dE Z dVMW dM 1 4⇡s2 ⇥ ⇥ dnsat(s, b, `, M) dM Z dVsat⇢2

sat(M)

Extragalactic

halo mass function boost from substructure

dIeg dE = hσvi 2m2

χ

Z dz dNγ[E(1 + z)] dE ¯ ρ2(z) 4π ⇥ ⇥ (1 + z)3 H(z) e−τ[E(1+z),z] ¯ ρ2(z) = Z dM dn(M, z) dM [1 + bsh(M)] × × Z dV ρ2

host(r, M)

(uncertain) (uncertain)

Uncertainties

• dN/dE from PPPC DM ID (Pythia+EW corrections)
• optical depth from semi-analytic modeling

(Gilmore, Primack, et al)

• halo mass function and subhalo boost factor
• semi-analytic fits (Phoenix simulation)
• Anderhalden and Diemand

for example...

Ellipsoidal Collapse Tinker et al ‡

z dn

dM dz' 1 + z' dn dM

2 4 6 8 10 0.0 0.5 1.0 1.5 z arbitrary units Planck+WMAP HSolidL vs. WMAP HDashedL

Relative intensities

MW, smooth Extragalactic MW, subhalos Total

0.1 1 10 102 10-10 10-9 10-8 10-7 10-6 Eg @GeVD E2◊XdIêdE\ @GeVêcm2◊s◊srD mDM=100 GeV, Xsv\=3¥10-26 cm3ês channel: DM DM Æ b b

default substructure calculation

Relative intensities

MW, smooth Extragalactic MW, subhalos Total

0.1 1 10 102 10-10 10-9 10-8 10-7 10-6 Eg @GeVD E2◊XdIêdE\ @GeVêcm2◊s◊srD mDM=100 GeV, Xsv\=3¥10-26 cm3ês channel: DM DM Æ b b

conservative substructure calculation

Relative intensities

NO substructure

MW, smooth Extragalactic Total

0.1 1 10 102 10-10 10-9 10-8 10-7 10-6 Eg @GeVD E2◊XdIêdE\ @GeVêcm2◊s◊srD mDM=100 GeV, Xsv\=3¥10-26 cm3ês channel: DM DM Æ b b

Many backgrounds

• star forming galaxies
• blazars (resolved and unresolved)
• radio galaxies (BL Lactaea objects,

FSRQs, etc.)

• ultra-high energy cosmic rays
• millisecond pulsars (...)

Backgrounds

Astrophysics can account for all of the EGB

Fits without DM

Example fit with DM

Cross-section bounds

dash-dot (projection)

Important probe

• Complementary to observations of stacked dwarf

spheroidals, galactic center, ... (similarly strong or stronger, with very different systematics)

• Very strong bounds!
• currently bounds thermal DM up to ~ 20 GeV
• Projected to be the strongest bound by end of

the Fermi mission

• can bound thermal DM up to ~ 400 GeV

Conclusions

• Bounds on light decaying DM from the

galactic diffuse background are strong (and robust!) even though observations are not DM-centric

• Looking outside the galaxy is a promising

method for putting strong constraints on more massive annihilating DM

Thank you!

Other in-progress projects I’m excited to talk about:

• thoughts on relic neutrino detection
• new work related to DM direct detection
• extensions of work on DM in neutron stars
• collider phenomenology of a light Higgs partner