WIMPs: An Introduction - Physics Institute of Bonn ... · WIMPs: An Introduction Manuel Drees Bonn University & Bethe Center for Theoretical Physics Introduction to WIMPs – p. 1/59

WIMPs: An IntroductionManuel Drees

Bonn University & Bethe Center for Theoretical Physics

Introduction to WIMPs – p. 1/59

Contents

1 The “WIMP Miracle”


Contents


2 Have WIMPS been Detected (Directly)?


Contents



3 Asymmetric Dark Matter?


Contents




4 WIMPs and Colliders


Contents




4 WIMPs and Colliders

5 Summary


Conditions for Dark Matter Candidates

Requirements for a good DM candidate χ:

Must have lifetime τχ ≫ τU





Must be electrically neutral (otherwise not dark)






Must have correct relic density: Ωχ ≃ 0.22







If DM consists of thermally produced “elementary” particles:Leads to events with missing ET at colliders!







If DM consists of thermally produced “elementary” particles:Leads to events with missing ET at colliders!

Counter–examples: axions; dark atoms; primordial black holes; keV

neutrinos: not covered in this talk. Note: Proves that LHC does not “recreate

conditions of the early universe”!


The “WIMP Miracle”

Assume χ was in full thermal equilibrium with SMparticles at sufficiently high temperature T :

χ production rate nχ〈σ(χχ → SM)vχ〉 > expansion rate H





nχ ∝ e−mχ/T , 〈σ(χχ → SM)v〉 ∝ T 0 or2, H ∝ T 2/MPlanck






=⇒ equality (“freeze-out”) reached at TF ≃ mχ/20







=⇒ Ωχh2 ≃0.1 pb · c

〈σ(χχ → SM)v〉







=⇒ Ωχh2 ≃0.1 pb · c

〈σ(χχ → SM)v〉

Indicates weak–scale χχ annihilation cross section:〈σ(χχ → any)v〉 ≃ 3 · 10−26cm3s−1


WIMPs and Early Universe

Ωχh2 can be changed a lot in non–standard cosmologies(involving T ≫ TBBN):

Increased: Higher expansion rate H(T ∼ TF );additional non–thermal χ production at T < TF ; . . .





Decreased: Reduced expansion rate H(T ∼ TF );entropy production at T < TF ; . . .





Decreased: Reduced expansion rate H(T ∼ TF );entropy production at T < TF ; . . .

Determining σ(χχ → SM) allows probe of very earlyUniverse, once χ has been established to be “the” DMparticle! e.g. MD, Iminniyaz, Kakizaki, arXiv:0704.1590


High−T Production of DM Particles

Sometimes χ production from thermalized SM particles iscalled “thermal production” even if χ never was in thermalequilibrium:




If σ(χχ → SM) ∝ 1/s: Production maximal at T ≃ mχ

(“freeze–in”, Hall et al., arXiv:0911.1120)






If σ(χχ → SM) ∝ s/Λ4: Produced dominantly at highestpossible temperature, TR.







Either way, χ interactions with SM particles are too weak togive missing ET signal, unless χ has “partners” that can beproduced via gauge interactions (ex.: gravitino G, axino a)







Either way, χ interactions with SM particles are too weak togive missing ET signal, unless χ has “partners” that can beproduced via gauge interactions (ex.: gravitino G, axino a)Interactions are too weak for direct detection; indirectdetection is possible only for unstable DM candidates.


Have WIMPs Been Detected Directly?

Direct detection ≡ search for elastic WIMP–nucleonscattering




Kinematics: vχ ∼ vSun ∼ 10−3c





=⇒ energy transfer to nculeus A:

Q <∼ min

(10−6 m2

χ

mA, 10−6mA

)<∼ 100 keV






Q <∼ min

(10−6 m2

χ

mA, 10−6mA

)<∼ 100 keV

=⇒ Cannot excite (most) nuclei!






Q <∼ min

(10−6 m2

χ

mA, 10−6mA

)<∼ 100 keV

=⇒ Cannot excite (most) nuclei!

Momentum transfer <∼ 100 MeV =⇒ may need to worry

about elastic form factors; quite well understood (for spin–indep.

scattering)


Recoil Spectrum

dRdQ ∝ |F (Q)|2

∫ vmax

vmin

dvv f1(v)


Recoil Spectrum

dRdQ ∝ |F (Q)|2

∫ vmax

vmin

dvv f1(v)

f1(v) : WIMP velocity distribution. Usually assumedMaxwellian in rest frame of the galaxy, cut off atvesc =⇒ vmax. Gives roughly exponentially falling spectrum.


Normalized Recoil Spectra

0 20 40 60 80 100Q [keV]

0.0001

0.001

0.01

0.1

1

1/R

dR

/dQ

[1/

keV

]

mχ=100 GeV, A=73

mχ=100 GeV, A=136

mχ=10 GeV, A=73

mχ=10 GeV, A=16


Experimental Challenges

Spectrum “backed up” against instrumental thresholdQmin




Rates of current interest ≪ background rate, e.g. fromradioactive decay (for most materials)=⇒ try to discriminate between nuclear recoil (signal)and e/γ induced events (background)!




Rates of current interest ≪ background rate, e.g. fromradioactive decay (for most materials)=⇒ try to discriminate between nuclear recoil (signal)and e/γ induced events (background)!

Will go through three claimed signals: DAMA(/LIBRA),CoGeNT, CRESST.


DAMA

Pure scintillation detectors (doped NaI) in Gran Sasso:6 years with 100 kg (DAMA)6 years with 250 kg (DAMA/LIBRA)


DAMA

Pure scintillation detectors (doped NaI) in Gran Sasso:6 years with 100 kg (DAMA)6 years with 250 kg (DAMA/LIBRA)All events are counted.


DAMA

Pure scintillation detectors (doped NaI) in Gran Sasso:6 years with 100 kg (DAMA)6 years with 250 kg (DAMA/LIBRA)All events are counted.Observe few percent modulation of total rate


DAMA

Pure scintillation detectors (doped NaI) in Gran Sasso:6 years with 100 kg (DAMA)6 years with 250 kg (DAMA/LIBRA)All events are counted.Observe few percent modulation of total rateCompatible with ∼ 50 GeV WIMP scattering off I, or ∼ 10GeV WIMP scattering off Na.


DAMA Results

2-6 keV

Time (day)

Res

idua

ls (

cpd/

kg/k

eV)

DAMA/NaI (0.29 ton×yr)(target mass = 87.3 kg)

DAMA/LIBRA (0.53 ton ×yr)(target mass = 232.8 kg)

2-6 keV

Time (day)

Res

idua

ls (

cpd/

kg/k

eV)

DAMA/LIBRA ≈ 250 kg (0.87 ton×yr)


DAMA: Problems

No e/γ discrimination is attempted, although some (statistical)discrimination should be possible using light curve


DAMA: Problems


Deduced shape of spectrum weird: Falls towards smallQ! (Depends on 3–dimensional WIMP velocity distribution.)


DAMA: Problems



Amplitude of modulation is getting smaller!E.g. in 2–6 keVee bin (in units of 10−3/kg · day · keVee):DAMA 1995–2001: 20.0 ± 3.2LIBRA 2003–2007: 10.7 ± 1.9LIBRA 2007–2009: 8.5 ± 2.2

Ratio LIBRAIIDAMA = 0.43 ± 0.13

More than 4σ away from 1! Results for 2–4, 2–5 keVee bins similar.


DAMA: Problems



Amplitude of modulation is getting smaller!E.g. in 2–6 keVee bin (in units of 10−3/kg · day · keVee):DAMA 1995–2001: 20.0 ± 3.2LIBRA 2003–2007: 10.7 ± 1.9LIBRA 2007–2009: 8.5 ± 2.2

Ratio LIBRAIIDAMA = 0.43 ± 0.13

More than 4σ away from 1! Results for 2–4, 2–5 keVee bins similar.

No convincing non–WIMP interpretation of modulationknown.


CoGeNT: Time–Averaged Analysis

Operate cryogenic Ge detectors with very low thresholdQmin.



Operate cryogenic Ge detectors with very low thresholdQmin.Originally: Found excess relative to simple bckgd model atvery low Q =⇒ small mχ;




χ2 fit with signal not significantly better than with purebackground: no claim of signal in published paper!




χ2 fit with signal not significantly better than with purebackground: no claim of signal in published paper!This September: More data, re–evaluateed background =⇒size of possible “signal” reduced by ∼ factor 5!


CoGeNT: Results


CoGeNT: Results


CoGeNT: Modulation

After 15 months of data taken: Find 2.8σ “evidence” forannual modulation


CoGeNT: Modulation


Much too large to be compatible with time–averaged“signal”, for standard halo


CoGeNT: Modulation


Much too large to be compatible with time–averaged“signal”, for standard halo

No event–by–event e/γ rejection at these low energies


CoGeNT: Summary

No signal claimed in time–averaged analysis!


CoGeNT: Summary


There is a large, poorly understood background


CoGeNT: Summary


There is a large, poorly understood background

Modulation “signal” statistically very weak, and way toolarge


CRESST

Uses cryogenic CaWO4 crystals; detect scintillation lightand heat: Allows event–by–event discrimination! See 67events after cuts.


CRESST

Uses cryogenic CaWO4 crystals; detect scintillation lightand heat: Allows event–by–event discrimination! See 67events after cuts.Unfortunately, quite large backgrounds (rough estimates, not final fit):

e/γ events: 8

α background: 9.2

n background: 1.5 to 11.4

Pb recoil (from 210Po decay): 17


CRESST


e/γ events: 8

α background: 9.2



Much of fitted excess has essentially no light: only “half asignal”


CRESST


e/γ events: 8

α background: 9.2



Much of fitted excess has essentially no light: only “half asignal”No. of α events is correlated with no. of signal events afterα subtraction.


CRESST: Results


CRESST: Results


CRESST: Results

What is negative light yield?


CRESST: Correlation

5 10 15 20events in α reference region

0

5

10

15

20

even

ts in

sig

nal r

egio

n (α

sub

trac

ted)

y = 0.44x + 2.1 (χ2 = 7.3 / 6 d.o.f.)

Data

y = 7.2 (χ2 = 11.4 / 7 d.o.f.)


Exclusion Limits from Other Expts

Best limit for larger masses from Xenon100. Uses ionization

and scintillation. Very few events after cuts. Alas, not safe for

mχ ≤ 12 GeV: bound strongly depends on high−v tail of f1(v),

and on experimental energy resolution.







Xenon10 more robust at small Q; excludes DAMA, CRESST,

original CoGeNT “signal” for “usual WIMP”









CDMS (+ EDELWEISS) second best for not too small mχ.

Uses phonons and ionization. Very few events after cuts. Not

safe below 12 GeV.











safe below 12 GeV.

CDMS low−Q analysis uses phonons only; sizable number of

events. Excludes DAMA, original CoGeNT for “usual WIMP”.











safe below 12 GeV.

CDMS low−Q analysis uses phonons only; sizable number of

events. Excludes DAMA, original CoGeNT for “usual WIMP”.

SIMPLE heated droplet detector: Challenges DAMA.


Theory of WIMP–Nucleus Scattering

Leff = cN NNχχ + aN NγµNχγµχ + bN Nγµγ5Nχγµγ5χ

For scalar χ: γµ → i∂µ in 2nd term; 3rd term absent

For Majorana χ: 2nd term absent

1st, 2nd term give spin–independent (s.i.) interaction, 3rd

term gives spin–dependent (s.d.) interaction.

“Usual WIMP”: same s.i. scattering on p and n!


Isospin violation in s.i. interaction?

σs.i.χp ≃ σs.i.

χn true for Higgs exchange (in particular, in(N)MSSM): massive quarks are same for p, n!





Not true for q, q(1) exchange: quark charges matter!But: M(χq → χq) has same sign for all quarks: nocancellations =⇒ isospin violation in praxis notimportant, since all nuclei have similar n/p ratio.





Not true for q, q(1) exchange: quark charges matter!But: M(χq → χq) has same sign for all quarks: nocancellations =⇒ isospin violation in praxis notimportant, since all nuclei have similar n/p ratio.

Gauge boson exchange can break isospin: coefficientsap, an may differ in sign! M(χq → χq) is now linear in(new) quark charges.


Large isospin violation in s.i. interaction?

|M(χA → χA)|2 ∝ |Zap + (A − Z)an|2

=⇒ need apan < 0 for significant isospin violation:arrange for cancellation in unwanted nuclei (e.g. Xe).



|M(χA → χA)|2 ∝ |Zap + (A − Z)an|2


Does not work for CDMS vs. CoGeNT: same target!



|M(χA → χA)|2 ∝ |Zap + (A − Z)an|2



Increases required |ap| even more, to describe claimedsignals=⇒ Need new light gauge bosons!



|M(χA → χA)|2 ∝ |Zap + (A − Z)an|2



Increases required |ap| even more, to describe claimedsignals=⇒ Need new light gauge bosons!

Combined analyses: (e.g. Kopp, Schwetz, Zupan, arXiv:1110.2721

[hep-ph]) Still cannot explain all data consistently!


Weisskopf’s (?) Theorem

A theory that explains all data must be wrong, since at anygiven point some data are wrong.


Weisskopf’s (?) Theorem

A theory that explains all data must be wrong, since at anygiven point some data are wrong.

Competition between null experiments with few(background) events after cuts, and claimed “signals” withlarge, not always well understood backgrounds!


Asymmetric Dark Matter?

Cosmology: ΩDM ≃ 5Ωbaryon: strange coincidence?




Ωbaryon determined by baryon–antibaryon asymmetry,not by thermal decoupling of baryons: try same thing forWIMPs?





Most attractive: χ − χ asymmetry related to baryonasymmetry [O(50) papers]

=⇒ number density nχ ≃ np after annihilating awaysymmetric component





Most attractive: χ − χ asymmetry related to baryonasymmetry [O(50) papers]

=⇒ number density nχ ≃ np after annihilating awaysymmetric component

Needs 2 to 3 times larger χχ annihilation cross sectionthan for thermal WIMPs!


“Explaining” ΩDM/Ωbaryon

Follows if mχ ≃ 5mp!




Alas: mχ has no known relation to mp. mχ ≃ 5mp just asmysterious as ΩDM ≃ 5Ωbaryon

=⇒ Haven’t explained anything!




Alas: mχ has no known relation to mp. mχ ≃ 5mp just asmysterious as ΩDM ≃ 5Ωbaryon

=⇒ Haven’t explained anything!

Circular logic: ADM with mχ ≃ 5 GeV interestingbecause there are “signals” for low–mass WIMPs;low–mass WIMPs interesting because ADM “explains”ΩDM/Ωbaryon??


WIMPs and Colliders

1 Generalities: WIMP DM Production and Missing ET


WIMPs and Colliders


2 Light Gauge Bosons


WIMPs and Colliders



3 SUSY DM and the “LHC Inverse Problem”


WIMPs and Colliders



3 SUSY DM and the “LHC Inverse Problem”

4 Higgs Searches and Direct DM Detection


Cannot predict missingET from χχ production

Thermal WIMP: Only know total χχ → SM crosssection; contribution of specific final states(e+e−, uu + dd) not known


Cannot predict missingET from χχ production

Thermal WIMP: Only know total χχ → SM crosssection; contribution of specific final states(e+e−, uu + dd) not known

Ωχh2 determined from σ(χχ → SM) near threshold(TF ≃ mχ/20 =⇒ s ≃ 4m2

χ). At colliders need ≥ 3 bodyfinal state to get signature (e.g. e+e− → χχγ, qq → χχg)=⇒ typically need σ(χχ → SM) at s ∼ 6 to 10m2

χ!


“Model-independent” approach

Goodman et al., arXiv:1005.1286 and 1008.1783; Bai, Fox, Harnik, arXiv:1005.3797; Wang,

Li, Shao, Zhang, arXiv:1107.2048; Fox, Harnek, Kopp, Tsai, arXiv:1103.0240

Parameterize χ interaction with relevant SM fermionthrough dim–6 operator; e.g. for hadron colliders:

Leff = GχχΓχχqΓqq







χ Majorana =⇒ Γχ ∈ 1, γ5, γµγ5

Γq ∈ 1, γ5, γµ, γµγ5

If Γχ, Γq ∈ 1, γ5 : Gχ = mq/(2M3∗ ) (chirality violating!), else

Γχ = 1/(2M2∗ ) Rajamaran, Shepherd, Tait, Wijango, arXiv:1108.1196.







χ Majorana =⇒ Γχ ∈ 1, γ5, γµγ5

Γq ∈ 1, γ5, γµ, γµγ5

If Γχ, Γq ∈ 1, γ5 : Gχ = mq/(2M3∗ ) (chirality violating!), else

Γχ = 1/(2M2∗ ) Rajamaran, Shepherd, Tait, Wijango, arXiv:1108.1196.

Compute monojet signal from qq → χχg, compare withmonojet limits (current bound) and background (ultimatereach)!


Γχ = γµγ5 (corr. to spin-dep. interact.)


Γχ = 1 (corr. to spin-indep. interact.)


Remarks

For Γχ = 1 (spin-indep. interact.): Current bound poor;ultimate LHC reach interesting only for mχ ≤ 5 GeV.


Remarks


For Γχ = γµγ5 (spin-dep. interact.): LHC bound betterthan (comparable to) direct search limit for mχ ≤ (≥) 20

GeV; future reach factor 103 better, if no other BSMsource of missing ET exists.


Remarks




Γχ = γ5 similar to first case; cannot be probed in directWIMP detection (rate ∝ v2

χ)


Remarks





χ)

Bound does not hold if mass of mediator particle≤ max(mχ, ET/ )!


Remarks





χ)

Bound does not hold if mass of mediator particle≤ max(mχ, ET/ )!

Altogether: very limited usefulness for most actual WIMPmodels.


2 DM and Light (Gauge) Bosons

(At least) 3 kinds of WIMP models require light (m ≤ fewGeV) (gauge) bosons U :

MeV DM: Suggested as explanation of 511 keV line(=⇒ slow e+) excess from central region of our galaxy(Boehm et al., astro-ph/0309686). Should have mχ ≤ 10 MeV (γconstraints)=⇒ mχ ≤ mU ≤ 200 MeV to mediate χχ → e+e−; fixesgUχχgUe+e−/m2

U ! (Unless 2mχ ≃ mU .)


2 DM and Light (Gauge) Bosons

(At least) 3 kinds of WIMP models require light (m ≤ fewGeV) (gauge) bosons U :

MeV DM: Suggested as explanation of 511 keV line(=⇒ slow e+) excess from central region of our galaxy(Boehm et al., astro-ph/0309686). Should have mχ ≤ 10 MeV (γconstraints)=⇒ mχ ≤ mU ≤ 200 MeV to mediate χχ → e+e−; fixesgUχχgUe+e−/m2

U ! (Unless 2mχ ≃ mU .)

PAMELA/FermiLAT inspired TeV DM: Needs lightboson for Sommerfeld enhancement (e.g. Arkani-Hamed et al.,

arXiv:0810.0713(4)) (χχ → UU → 4l is also somewhat lessconstrained by γ spectrum than χχ → 2l.)


DAMA/CoGeNT inspired few GeV DM: Needs lightmediator to achieve sufficiently large σχp. (2 differentmediators for isospin violation to evade bounds: Cline, Frey,

arXiv:1108.1391)


Light Gauge Bosons (cont’d)

In all cases: U couplings to (most) SM particles must be≪ 1 to evade bounds! (gµ − 2, meson decays, ν crosssections, APV, . . . ).




Possible explanation: kinetic mixing with γ/B boson! Is1-loop effect =⇒ squared Uff coupling is O(α3).




Possible explanation: kinetic mixing with γ/B boson! Is1-loop effect =⇒ squared Uff coupling is O(α3).

Uχχ coupling may well be large.


Signatures of light gauge bosons

If mU > 2mχ: U → χχ dominant! Is invisible =⇒ need extra

tag, e.g. e+e− → γU → γ+ nothing.





Physics background ∝ s =⇒ lower energy is better!Borodatchenkova, Choudhury, MD, hep-ph/0510147





Physics background ∝ s =⇒ lower energy is better!Borodatchenkova, Choudhury, MD, hep-ph/0510147

Instrumental backgrounds (not from e+e− annihilation)seem large


Sensitivity at B−factories (100 fb−1)

0.001 0.01 0.1M

U [GeV]

1e-08

1e-06

0.0001

0.01

1

g χg e R

mχ < MU

/2

mχ > MU

/2

upper boundfrom g

e-2

max. sensitivity(e

+e

- channel)

max. sensitivity(invisible channel)

Red, black: Regions allowed by Ωχ, σ(χχ → e+e−).Introduction to WIMPs – p. 38/59

Signatures of light gauge bosons (cont.d)

If mU < 2mχ: U → ℓ+ℓ−




Sufficiently light U can even be produced in fixed–targetexperiments: e−N → e−e+e−N (tridents), with peak inMe+e−





First exptl. results from MAMI A1 arXiv:1101.4091 and JLABAPEX arXiv:1108.2750 Excludes new mass ranges around 200to 300 MeV for A′ ≡ U kinetically mixed with photon.





First exptl. results from MAMI A1 arXiv:1101.4091 and JLABAPEX arXiv:1108.2750 Excludes new mass ranges around 200to 300 MeV for A′ ≡ U kinetically mixed with photon.

Also, KLOE-2 performed search, mostly for φ → Uη: nosignal. arXiv:1107.2531


A1 and APEX results


3 SUSY DM and LHC “Inverse Problem”

Saw above: WIMP searches at colliders not promising, ifWIMP is only accessible new particle. Fortunately, in manycases the WIMP is the lightest of many new particles! Truein SUSY. (Also in Little Higgs.)



Saw above: WIMP searches at colliders not promising, ifWIMP is only accessible new particle. Fortunately, in manycases the WIMP is the lightest of many new particles! Truein SUSY. (Also in Little Higgs.)Recall: Primary motivation for SUSY not related to DM!




Stabilizes hierarchy m2Higgs ≪ M2

Planck





Planck

Allows unification of gauge couplings





Planck


In scenarios with unified Higgs masses: EWSB requiressizable hierarchy! (Not in NUHM2.)





Planck


In scenarios with unified Higgs masses: EWSB requiressizable hierarchy! (Not in NUHM2.)

HLS theorem, relation to superstrings: don’t single out weakscale.


Features of SUSY

Need superpartner for each SM particle: Same rep. of gauge

group, spin differs by 1/2


Features of SUSY



Need at least 2 Higgs doublets (anomalies, mt · mb 6= 0)


Features of SUSY




SUSY implies equal masses for partners =⇒ SUSY must be

broken


Features of SUSY





broken

Naturalness: sparticle masses should be at weak scale (strictly

true only for 3rd generation, elw gauginos)


Features of SUSY





broken

Naturalness: sparticle masses should be at weak scale (strictly

true only for 3rd generation, elw gauginos)

In simplest, R−parity invariant scenario: lightest superparticle

LSP is stable: satisfies one condition for DM candidate!


SUSY DM candidate: sneutrinoν

Disfavored theoretically: not LSP in constrained models




Excluded experimentally by direct searches




Excluded experimentally by direct searches

νR can be candidate in extended theories, e.g. withgauged U(1)B−L at TeV scale


SUSY DM candidate: neutralino χ01

Mixture of B, W3, h0u, h0

d




d

In constrained models: often is lightest sparticle invisible sector! (Other possibility: lightest stau τ1)




d


In “most” of parameter space: χ01 ≃ B, and predicted

Ωχ01h2 too large! O(1 to 10) rather than O(0.1) in

standard cosmology,




d


In “most” of parameter space: χ01 ≃ B, and predicted

Ωχ01h2 too large! O(1 to 10) rather than O(0.1) in

standard cosmology,

but DM–allowed regions of parameter space do existeven in constrained models!


Regions with correctΩχ0

1h2

Co–annihilation region: mχ01≃ mτ1



1h2


Higgs funnel(s): mχ01≃ mh/2, mA/2



1h2



Well–tempered neutralino: µ − M1 ≤ MZ =⇒ χ01 is

B − h0 mixture. (Requires mq ≫ mg in cMSSM; can bearranged “anywhere” in NUHM.)



1h2



Well–tempered neutralino: µ − M1 ≤ MZ =⇒ χ01 is

B − h0 mixture. (Requires mq ≫ mg in cMSSM; can bearranged “anywhere” in NUHM.)

Note: DM–allowed region of (m0,m1/2) plane of cMSSMdepends on A0, tan β!


Impact of LHC searches

Is model dependent: Only probe g, q sector so far! Here:Assume cMSSM for defineteness.




Well–tempered neutralino, A−pole need large mq: limitsstill fairly weak: mg,min increased from ∼ 400 GeV to∼ 550 GeV




Well–tempered neutralino, A−pole need large mq: limitsstill fairly weak: mg,min increased from ∼ 400 GeV to∼ 550 GeV

τ1 co–annihilation requires mq ≤ mg: good for LHCsearches; still plenty of allowed region left.


Impact of Indirect DM Searches

No halo signal in co–annihilation region; ν from Sunsmall




Signals very small in A−funnel




Signals very small in A−funnel

Well-tempered neutralino most promising, especially νfrom Sun, but present limits not constraining


Impact of direct WIMP Searches

XENON, CDMS⊕EDELWEISS begin to probewell–tempered neutralino


Impact of direct WIMP Searches

XENON, CDMS⊕EDELWEISS begin to probewell–tempered neutralino

Signals in other regions very small


Impact of Future WIMP Discovery at Collider

Generically: could determine:

WIMP mass: Very useful for indirect searches (greatlyreduced “look elsewhere” problem); less so for directsearches, once mχ ≥ mN





WIMP couplings: Determine cross sections and finalstates in indirect searches; determine cross sections indirect searches





WIMP couplings: Determine cross sections and finalstates in indirect searches; determine cross sections indirect searches

Most interesting to me: Predict Ωχh2, compare withobservation: Constrain very early universe!


Fitting SUSY parameters at LHC (“inverse problem”)

Feasibility (resulting errors) very strongly depend on wherewe are in parameter space!




Good: Low sparticle masses, many leptons e, µ in finalstates.





Not so good: Large masses, mostly hadronic finalstates.





Not so good: Large masses, mostly hadronic finalstates.

Two approaches: Case studies, broad scans of parameterspace


Case study:τ1 co–annihilation region in cMSSMArnowitt et al., arXiv:0802.2968

Needs mτ1− mχ0

1≤ 15 GeV

=⇒ χ02 → τ1τ, χ±

1 → τ±1 ντ have nearly unit branchingratio=⇒ no di–lepton edges!



Needs mτ1− mχ0

1≤ 15 GeV

=⇒ χ02 → τ1τ, χ±


τ1 → χ01τ gives rather soft τ : Difficult to detect!



Needs mτ1− mχ0

1≤ 15 GeV

=⇒ χ02 → τ1τ, χ±



Study three classes of final states:(i) 2τ + 2j + ET/(ii)4 non − b j + ET/(iii) leading b + 3j + ET/



Needs mτ1− mχ0

1≤ 15 GeV

=⇒ χ02 → τ1τ, χ±



Study three classes of final states:(i) 2τ + 2j + ET/(ii)4 non − b j + ET/(iii) leading b + 3j + ET/

Fit many kinematical distributions simultaneously,including slope of softer pτ

T spectrum in sample (i) =⇒

predict Ωχ01h2 to 10%!


Result of fit

M (GeV)∆8 9 10 11 12 13

0.08

0.09

0.1

0.11

0.12Ω

~ χ2 h 10


Result of fit

M (GeV)∆8 9 10 11 12 13

0.08

0.09

0.1

0.11

0.12Ω

~ χ2 h 10

Unfortunately, chosen benchmark point (mg = 830 GeV,mq ≃ 750 GeV) is most likely excluded!


Scan of Parameter SpaceArkani-Hamed et al., hep-ph/0512190

15 parameter description of weak–scale MSSM



15 parameter description of weak–scale MSSMFix mA = 850 GeV, A3 = 800 GeV



15 parameter description of weak–scale MSSMFix mA = 850 GeV, A3 = 800 GeVRandomly choose mq,g ∈ [600, 1000] GeV,other masses ∈ [100, 1000] GeV, tan β ∈ [2, 50]



15 parameter description of weak–scale MSSMFix mA = 850 GeV, A3 = 800 GeVRandomly choose mq,g ∈ [600, 1000] GeV,other masses ∈ [100, 1000] GeV, tan β ∈ [2, 50]Consider 1808 observables, both counting rates and binneddistributions



15 parameter description of weak–scale MSSMFix mA = 850 GeV, A3 = 800 GeVRandomly choose mq,g ∈ [600, 1000] GeV,other masses ∈ [100, 1000] GeV, tan β ∈ [2, 50]Consider 1808 observables, both counting rates and binneddistributionsIntroduce “χ2−like” variable

(∆SAB)2 =1

nsig

nsig∑

i=1

(sAi − sB

i

σABi

)2

Only “significant” signatures included




(∆SAB)2 =1

nsig

nsig∑

i=1

(sAi − sB

i

σABi

)2

Only “significant” signatures includedAllow 1% syst. error (15% on total signal rate), 10 fb−1 of14 TeV data, no background




(∆SAB)2 =1

nsig

nsig∑

i=1

(sAi − sB

i

σABi

)2

Only “significant” signatures includedAllow 1% syst. error (15% on total signal rate), 10 fb−1 of14 TeV data, no backgroundMC: (∆SAB)2 > 0.285 =⇒ models differ at > 95% c.l.


Results and Remarks

Found 283 degenerate pairs, with (δSAB)2 < 0.285, for43,026 “models” (i.e., sets of parameters)


Results and Remarks

Found 283 degenerate pairs, with (δSAB)2 < 0.285, for43,026 “models” (i.e., sets of parameters)Note

Their observables are correlated=⇒ (δSAB)2 is no true χ2

=⇒ need MC to intepret it, from comparing runs withdifferent random no. seed: is this reliable estimator forcomparing different parameter sets?


Results and Remarks

Found 283 degenerate pairs, with (δSAB)2 < 0.285, for43,026 “models” (i.e., sets of parameters)Note

Their observables are correlated=⇒ (δSAB)2 is no true χ2

=⇒ need MC to intepret it, from comparing runs withdifferent random no. seed: is this reliable estimator forcomparing different parameter sets?

Statistics looks weird! Comparing two simulations ofsame “model”, get 611 (out of 2600) cases where some2ℓ observable has > 5σ discrepancy: way too many!


Simpler approach

Bornhauser and MD, in progress

Define 12 disjunct event classes, depending on no.,charge, flavor of leptons


Simpler approach



Consider 7 mostly uncorrelated observables for eachclass: No. of events; 〈nτ±〉; 〈nb〉; 〈nj〉; 〈n2

j〉; 〈HT 〉


Simpler approach



Consider 7 mostly uncorrelated observables for eachclass: No. of events; 〈nτ±〉; 〈nb〉; 〈nj〉; 〈n2

j〉; 〈HT 〉

Define proper χ2, incl. corr. between 〈nj〉, 〈n2j〉, only

including significant observables: test with MC.


Results of simpler approach

W/o syst. error: only one of 283 “degenerate” pairs hasp > 0.05, and two parameter sets really are very similar!(Except for heavy sleptons.)




Introducing syst. errors as Arkani-Hamed et al.: 41 outof 283 pairs have p > 0.05; still are pretty similarphysically





Introducing SM background, but no syst. error: 12 pairshave p > 0.05





Introducing SM background, but no syst. error: 12 pairshave p > 0.05

With systematic errors and background: 71 pairs havep > 0.05.


4 Impact of Higgs Searches

In many WIMP models, Higgs exch. dominates χp

scattering, in which case σχp ∝ 1/m4H : crucial to know

Higgs mass!





Higgs mass!

In SUSY at large tan β: σχ01p

∝ tan2 β/m4A: need info on

heavy Higgses!





Higgs mass!



heavy Higgses!

TeVatron and CMS searches for H,A → τ+τ−

significantly increase lower bound on DM–allowed mχ01

in general MSSM (Aborno Vasquez, Belanger, Boehm, arXiv:1108.1338);exclude scenarios with very large σχ0

1p.





Higgs mass!



heavy Higgses!

TeVatron and CMS searches for H,A → τ+τ−

significantly increase lower bound on DM–allowed mχ01

in general MSSM (Aborno Vasquez, Belanger, Boehm, arXiv:1108.1338);exclude scenarios with very large σχ0

1p.

Higgs searches can also be used to distinguishbetween WIMP models and to help determineparameters. E.g. mh in MSSM constrains stop sector.


Summary 1

Rumours of the direct detection of WIMPs are greatlyexaggerated


Summary 1

Rumours of the direct detection of WIMPs are greatlyexaggerated

“Asymmetric Dark Matter” doesn’t explainΩDM ≃ 5Ωbaryon


Summary 2

Well–motivated WIMP models can be tested atcolliders!


Summary 2


Scenarios with new light gauge bosons withsuppressed couplings to SM fermions are now beingprobed at low−E colliders, fixed–target expts.


Summary 2



LHC not very good for “model–independent” WIMPsearch. (Signal is O(α2αS), background is O(ααS).)


Summary 2




If WIMP signal is found at LHC, LHC experiments willbe better at extracting parameters than indicated bytheory analyses that have been published so far; manyavenues remain to be explored.


Summary 2




If WIMP signal is found at LHC, LHC experiments willbe better at extracting parameters than indicated bytheory analyses that have been published so far; manyavenues remain to be explored.

Higgs sector also very important for WIMP physics!


Documents

WIMPs: An Introduction - Physics Institute of Bonn ... · WIMPs: An Introduction Manuel Drees Bonn University & Bethe Center for Theoretical Physics Introduction to WIMPs – p. 1/59