Collider searches for dark matter. Joachim Kopp. SLAC, October 5, Fermilab

Transcription

1 Collider searches for dark matter Joachim Kopp SLAC, October 5, 2011 Fermilab based on work done in collaboration with Patrick Fox, Roni Harnik, Yuhsin Tsai Joachim Kopp Collider searches for dark matter 1

2 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 2

3 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 3

4 Search strategies for dark matter (DM) Direct searches Look for DM nucleus scattering. Uncertainties: Local DM density Backgrounds Calibration Indirect searches Look for astrophysical signatures of DM annihilation or decay. Uncertainties: Profile of Milky Way s DM halo Cosmic ray propagation Collider searches Look for missing energy signatures. Problem: Can only find DM candidate (no proof that it is DM) Model-dependent strategy: Cascade decays with /E T Less model-dependent strategies: THIS TALK Joachim Kopp Collider searches for dark matter 4

5 Dark matter in an effective theory approach Assumption: DM interactions described by effective field theory Sample operators: (χ = dark matter, f = SM fermion, Λ = suppression scale) O V = ( χγ µχ)( f γ µ f ) Λ 2, (vector, s-channel) O S = ( χχ)( f f ) Λ 2 (scalar, s-channel) O A = ( χγ µγ 5 χ)( f γ µ γ 5 f ) Λ 2 (axial vector, s-channel) O t = ( χp Lf )( f P R χ) Λ 2 + (L R), (scalar, t-channel) can be Fierz ed into s-channel operators O g = α s( χχ)(gµνg a aµν ) Λ 2 (scalar, s-channel) f χ In a full, UV complete theory: Λ = M/ g f g χ f Joachim Kopp Collider searches for dark matter 5 χ

6 Mono-jet and mono-photon signatures of dark matter Idea: Pair production of DM + some visible particles Tevatron, LHC: Mono-jets χ q coupling probed in jet(s) + /E T q q χ CDF (1.1 fb 1 ): , ATLAS (1 fb 1 ): ATLAS-CONF , CMS (1.1 fb 1 ) : CMS-PAS-EXO Goodman Ibe Rajaraman Shepherd Tait Yu , Rajaram Shepherd Tait Wijangco Bai Fox Harnik, Fox Harnik JK Tsai χ LEP, Tevatron, LHC: Mono-γ χ f coupling probed in photon + /E f f DELPHI (650 pb 1 ): hep-ex/ , CDF (2 fb 1 ): DØ(1 fb 1 ): CMS (1.14 fb 1 ): CMS-PAS-EXO χ Fox Harnik JK Tsai , χ Joachim Kopp Collider searches for dark matter 6

7 Mono-jet and mono-photon signatures of dark matter Idea: Pair production of DM + some visible particles Tevatron, LHC: Mono-jets χ q coupling probed in jet(s) + /E T q q χ CDF (1.1 fb 1 ): , ATLAS (1 fb 1 ): ATLAS-CONF , CMS (1.1 fb 1 ) : CMS-PAS-EXO Goodman Ibe Rajaraman Shepherd Tait Yu , Rajaram Shepherd Tait Wijangco Bai Fox Harnik, Fox Harnik JK Tsai χ LEP, Tevatron, LHC: Mono-γ χ f coupling probed in photon + /E f f DELPHI (650 pb 1 ): hep-ex/ , CDF (2 fb 1 ): DØ(1 fb 1 ): CMS (1.14 fb 1 ): CMS-PAS-EXO χ Fox Harnik JK Tsai , χ Joachim Kopp Collider searches for dark matter 6

8 Mono-jet and mono-photon signatures of dark matter Idea: Pair production of DM + some visible particles Tevatron, LHC: Mono-jets χ q coupling probed in jet(s) + /E T q q χ CDF (1.1 fb 1 ): , ATLAS (1 fb 1 ): ATLAS-CONF , CMS (1.1 fb 1 ) : CMS-PAS-EXO Goodman Ibe Rajaraman Shepherd Tait Yu , Rajaram Shepherd Tait Wijangco Bai Fox Harnik, Fox Harnik JK Tsai χ LEP, Tevatron, LHC: Mono-γ χ f coupling probed in photon + /E f f DELPHI (650 pb 1 ): hep-ex/ , CDF (2 fb 1 ): DØ(1 fb 1 ): CMS (1.14 fb 1 ): CMS-PAS-EXO χ Fox Harnik JK Tsai , χ Joachim Kopp Collider searches for dark matter 6

9 Limit setting procedure 1 Simulate signal process pp/p p j + χχ or pp/p p/e + e γ + χχ in MadGraph, Pythia, Delphes/PGS (Simulate also SM backgrounds for verification) Events GeV ATLAS 7 TeV, 1 fb 1, veryhighpt (Z νν) + j (W l inv ν) + j T ATLAS data ATLAS BG our MC DM signal Events 650 pb DELPHI 650 pb 1 DELPHI MC our MC DM signal Our simulations in excellent agreement with the collaborations (after correcting normalization by 10 30%) Signal and background have different spectral shape (Z νν) + γ Joachim Kopp Collider searches for dark matter 7

10 Limit setting procedure 1 Simulate signal process pp/p p j + χχ or pp/p p/e + e γ + χχ in MadGraph, Pythia, Delphes/PGS (Simulate also SM backgrounds for verification) 2 Compare signal + BG prediction to data: (require χ 2 = 2.71 to derive 90% C.L. limit on Λ) LHC: LEP: χ 2 [Nobs NSM NDM(mχ, Λ)]2 N DM(m χ, Λ) + N SM + σsm 2 χ 2 2 X j=bins h N j SM + Nj DM (mχ, Λ) Nj obs + Nj obs log N j i obs N j SM + Nj DM (mχ, Λ) Only total rate analysis for LHC mono-jet (cannot model systematic uncertainties in background shape) LEP mono-photon analysis is statistics dominated can use spectral information Joachim Kopp Collider searches for dark matter 7

11 Limit setting procedure 1 Simulate signal process pp/p p j + χχ or pp/p p/e + e γ + χχ in MadGraph, Pythia, Delphes/PGS (Simulate also SM backgrounds for verification) 2 Compare signal + BG prediction to data: Cutoff scale GeV Solid : Observed Dashed : Expected ΧΓ Μ Χ uγ Μ u ΧΓ Μ Χ dγ Μ d 90 C.L. 400 ATLAS 1 fb pp j + /E T veryhighpt Cutoff scale GeV ΧΧ ee Χe e Χ ΧΓ Μ Χ eγ Μ e DELPHI 650 pb 1 e + e γ + /E T ΧΓ Μ Γ 5 Χ eγ Μ Γ 5 e 90 C.L Joachim Kopp Collider searches for dark matter 7

12 Limit setting procedure 1 Simulate signal process pp/p p j + χχ or pp/p p/e + e γ + χχ in MadGraph, Pythia, Delphes/PGS (Simulate also SM backgrounds for verification) 2 Compare signal + BG prediction to data: Cutoff scale GeV Solid : Observed Dashed : Expected veryhighpt HighPt LowPt 90 C.L. 300 CMS ΧΓ Μ Χ uγ Μ u High jet pt and /E T cuts are very beneficial Joachim Kopp Collider searches for dark matter 7

13 Limit setting procedure 1 Simulate signal process pp/p p j + χχ or pp/p p/e + e γ + χχ in MadGraph, Pythia, Delphes/PGS (Simulate also SM backgrounds for verification) 2 Compare signal + BG prediction to data: 3 Convert limits on Λ into limits on DM nucleus scattering cross section or DM annihilation cross section Joachim Kopp Collider searches for dark matter 7

14 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 8

15 LHC limits on the DM nucleon scattering cross section WIMP nucleon cross section ΣN cm ATLAS 7TeV, 1fb 1 VeryHighPt Solid : Observed Dashed : Expected ΧΓ Μ Χ qγ Μ q CoGeNT Α s ΧΧ G ΜΝ G ΜΝ DAMA q ± 33 CRESST 90 C.L Spin independent WIMP mass m Χ GeV XENON 10 CDMS XENON 100 WIMP nucleon cross section ΣN cm ATLAS 7TeV, 1fb 1 VeryHighPt Solid : Observed Dashed : Expected DAMA q ± 33 Spin dependent ΧΓ Μ Γ 5 Χ qγ Μ Γ 5 q 90 C.L WIMP mass m Χ GeV PICASSO XENON 10 COUPP SIMPLE Assumption here: Equal coupling to all quark flavors Extremely competitive limits for See also work by Rajaraman et al Light dark matter (below direct detection threshold) DM coupled to gluons (high gluon luminosity at the LHC) Spin-dependent DM interactions (DD suffers from loss of coherence) γ + /E T final state slightly less sensitive, but could provide confirmation or discriminate between models if signal is observed in j + /E T Joachim Kopp Collider searches for dark matter 9

16 LEP limits on the DM nucleon scattering cross section WIMP nucleon cross section ΣN cm Equal couplings to all SM fermions ΧΧ f f Χf f Χ ΧΓ Μ Χ f Γ Μ f CoGeNT Spin independent DAMA q ± C.L. XENON 100 CDMS WIMP mass m Χ GeV WIMP proton cross section Σp cm Equal couplings to all SM fermions DAMA q ± 33 ΧΓ Μ Γ 5 Χ f Γ Μ Γ 5 f Spin dependent 90 C.L. PICASSO XENON 10 COUPP SIMPLE WIMP mass m Χ GeV LEP only constrains DM electron coupling Additional assumptions needed to set limit on DM nucleon scattering Here: Equal coupling to all SM fermions assumed Limits only slightly weaker than those from the LHC, but slightly more model-dependent Joachim Kopp Collider searches for dark matter 10

17 Direct detection phenomenology of leptophilic DM What if dark matter couples only to electrons at tree level? Scattering on an electron Outer-shell electrons can be kicked out (WIMP-electron scattering) Inner-shell electrons will remain bound (elastic WIMP-atom scattering) recoil transferred to nucleus Electrons can be excited to an outer shell, but remain bound (inelastic WIMP-atom scattering) recoil partly transferred to nucleus Problem: Typically very small recoil energies (mχ m e) Visible events probe high-momentum tail of e wave functions Iodine Sodium p 2 Χ p 2 ev s 3d 2p 2s 4d 3p 3s 4p 4s 5p 5s p ev p 2 Χ p 2 ev s 2s 2p 3s p ev JK Niro Schwetz Zupan arxiv: Joachim Kopp Collider searches for dark matter 11

18 Direct detection phenomenology of leptophilic DM What if dark matter couples only to electrons at tree level? Scattering on an electron Strongly suppressed Loop-induced scattering on the nucleus Dominant if allowed Forbidden e.g. for axial vector operator ( χγµγ 5χ)( f γ µ γ 5f ) Suppressed by loop factor χ p + l γ χ l p + χ, k µ χ, kµ l q γ l γ N, p µ N, pµ χ, k µ l q γ l χ, kµ γ N, p µ N, pµ JK Niro Schwetz Zupan arxiv: Joachim Kopp Collider searches for dark matter 11

19 LEP constraints on leptophilic dark matter Scattering on an electron (Loop forbidden), e.g. ( χγ µ γ 5 χ)(ēγ µ γ 5 e): LEP provides the only meaningful limit on DM SM scattering Loop-suppressed scattering on the nucleus, e.g. ( χγ µ χ)(ēγ µ e): LEP still has a great advantage: WIMP proton cross section Σp cm Couplings to leptons only 90 C.L. DAMA q ± 33 CoGeNT ΧΓ Μ Χ Γ Μ Χ Χ Spin independent XENON 100 CDMS WIMP mass m Χ GeV χ χ l l Fox Harnik JK Tsai, arxiv: γ p + p + Joachim Kopp Collider searches for dark matter 12

20 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 13

21 Collider limits on DM annihilation χχ qq (LHC limits) Cross section Σ vrel for Χ Χ q q cm 3 s C.L. Solid: Observed Dashed: Expected Χ Γ Μ Χ q Γ Μ q Χ Γ Μ Γ 5 Χ q Γ Μ Γ 5 q Thermal relic v 2 rel 0.24 freeze out χχ l + l (LEP limits) Cross section Σ vrel for ΧΧ cm 3 s Equal coupling to all charged leptons 90 C.L Hooper Goodenough Fermi LAT, WMAP haze Only v rel independent models WIMP mass m Χ GeV ΧΓ Μ Χ Γ Μ Χ Χ Fermi Γ: ΧΧ Μ Μ Draco Fermi e e : ΧΧ Thermal relic Light thermal relic ruled out Constraints weakens by 1/BR( χχ qq) resp. 1/BR( χχ ll) if DM has also other annihilation channels Joachim Kopp Collider searches for dark matter 14

22 Velocity-dependent annihilation cross sections σ S v rel = 1 r 1 m2 f (m 2 8πΛ 4 mχ 2 χ mf 2 ) vrel 2 scalar r! σ A v rel = 1 48πΛ 4 1 m2 f m 2 χ 24m 2 f + 8m4 χ 22m 2 χm 2 f + 17m 4 f m 2 χ m 2 f For some operators, σv rel is suppressed by v 2 rel or m2 f /m2 χ. advantage for colliders compared to astrophysical searches v 2 rel axial vector Cross section Σ vrel for ΧΧ cm 3 s Χ C.L ΧΓ Μ Γ 5 Χ Γ Μ Γ Χ v rel 69.3 km sec 2 Draco ΧΓ Μ Χ Γ Μ Fermi Γ: ΧΧ Μ Μ ΧΧ Draco dwarf galaxy Joachim Kopp Collider searches for dark matter 15

23 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 16

24 Beyond effective field theory Assume DM interactions mediated by light particle effective field theory breaks down, have to include mediator explicitly f χ f χ φ f χ f χ Collider cross section σ coll 1 (q 2 M 2 med )2 + Γ 2 med /4 ŝ Direct detection cross section σ scatter 1 mn 2 m2 χ Mmed 4 (m N + m χ ) 2 For light mediators, colliders have a relative disadvantage... unless a narrow mediator can be produced on-shell and decays to DM Joachim Kopp Collider searches for dark matter 17

25 Constraints on suppression scale Λ for light mediators Continue to use Λ M med / g χ g f as measure for DM interaction strength Cutoff scale GeV min mχ min mχ 1 GeV 1 GeV min mχ 1 GeV 90 C.L. DELPHI 650 pb 1 M 200 GeV contact op. M 100 GeV M 50 GeV M 10 GeV m χ > M med /2: Limit is weaker than for the effective field theory (contact operator) case m χ < M med /2: Mediator produced on-shell Limit improves again... but depends on the partial width for Mediator χχ Note: If Mediator SM SM is possible, other constraints may apply Joachim Kopp Collider searches for dark matter 18

26 DM nucleon scattering with light mediators WIMP nucleon cross section ΣN cm C.L. 1 GeV contact op. 1 GeV min mχ min mχ M 50 GeV ΧΓ Μ Χ f Γ Μ f DELPHI 650 pb 1 M 200 GeV M 100 GeV Spin independent Equal couplings to all SM fermions WIMP mass m Χ GeV m χ > M med /2: Limit weaker m χ < M med /2: Limit stronger or weaker, depending on width Joachim Kopp Collider searches for dark matter 19

27 DM annihilation with light mediators Cross section Σ vrel for ΧΧ e e cm 3 s C.L. Thermal relic 1 GeV 1 GeV min mχ min mχ contact op. DELPHI 650 pb 1 M 50 GeV M 100 GeV M 200 GeV v 2 rel 0.24 freeze out Annihilation only into e e m χ > M med /2: Limit weaker m χ < M med /2: Limit stronger or weaker, depending on width Spikes due to resonant annihilation Joachim Kopp Collider searches for dark matter 20

28 Are our bounds within the regime of validity of EFT? Question: Are there UV-complete models that saturate our limits and can still be described by effective operators? For LEP: No problem s 200 GeV, Λ lim O( ) GeV. models with M med 500 GeV and g χ, g f 1 are OK Joachim Kopp Collider searches for dark matter 21

29 Are our bounds within the regime of validity of EFT? Question: Are there UV-complete models that saturate our limits and can still be described by effective operators? For LEP: No problem For LHC: Models with Mmed 5 TeV and g χ, g f 5 required For smaller (but not too small) Mmed, g χ, g f, limits derived in EFT are overly conservative 90 CL limit on cutoff scale lim GeV Vector coupling m Χ 50 GeV m Χ 500 GeV Shading: M 3 M 8 Π g Χ g q contours Joachim Kopp Collider searches for dark matter 21

30 Are our bounds within the regime of validity of EFT? Question: Are there UV-complete models that saturate our limits and can still be described by effective operators? For LEP: No problem For LHC: No problem Joachim Kopp Collider searches for dark matter 21

31 Are our bounds within the regime of validity of EFT? Question: Are there UV-complete models that saturate our limits and can still be described by effective operators? For LEP: No problem For LHC: No problem EFT might work even better for more sophisticated UV completions Joachim Kopp Collider searches for dark matter 21

32 Are our bounds within the regime of validity of EFT? Question: Are there UV-complete models that saturate our limits and can still be described by effective operators? For LEP: No problem For LHC: No problem EFT might work even better for more sophisticated UV completions Note: EFT may be problematic for scalar operators (SU(2) invariance requires Higgs insertions) Fox Harnik JK Tsai Joachim Kopp Collider searches for dark matter 21

33 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 22

34 Invisible Higgs decays to dark matter A special case of a light mediator is the SM Higgs boson. Best limits come not from the j + /E T or γ + /E T channels, but from invisible Higgs searches Vector boson fusion: forward jets + /E T Associated production with a Z : Z + /E T Limits on BR(h inv) can be translated into Higgs DM coupling y χ BR(h χχ) = Γ(h χχ) = y 2 χ 8π m h Γ(h χχ) Γ(h χχ) + Γ(SM), [ ( ) ] 2 3/2 2mχ 1, m h Joachim Kopp Collider searches for dark matter 23

35 Projected limits from invisible Higgs decay No invisible Higgs search yet compute future sensitivity (based on projected BR(h inv) sensitivity from Gagnon et al. ATLAS CSC NOTE 10) ATLAS 30 fb 1 upper bound projected WIMP nucleon cross section ΣN cm CoGeNT VBF m h 250 GeV ZH inv m h 120 GeV VBF m h 120 GeV DAMA q ± 33 CRESST LHC 14 TeV, 30 fb 1 Spin independent future sensitivity WIMP mass m Χ GeV XENON 10 CDMS XENON 100 Joachim Kopp Collider searches for dark matter 24

36 A lower limit on DM nucleon scattering Assume DM interacts through Higgs exchange Assume specific value for m h Non-obervation of Higgs at m h can be interpreted as a lower limit on BR(h inv) WIMP nucleon cross section ΣN cm CoGeNT CMS Higgs combined lower bound CMS lower bound m h 250 GeV m h 350 GeV m h 150 GeV DAMA q ± 33 CRESST Spin independent WIMP mass m Χ GeV XENON 10 CDMS XENON 100 Joachim Kopp Collider searches for dark matter 25

37 Outline 1 Mono-jet and mono-photon signals of dark matter 2 Collider limits on direct detection cross sections 3 Collider limits on annihilation cross sections 4 Beyond effective field theory 5 Invisible Higgs decays to dark matter 6 Summary & Conclusions Joachim Kopp Collider searches for dark matter 26

38 Summary & Conclusions Mono-jet and Mono-photon searches at LEP and LHC provide strong constraints on dark matter properties in an effective field theory formalism Colliders are superior to direct searches if dark matter is very light ( O(5 GeV)) or if interactions are spin-dependent or leptophilic Colliders are superior to indirect searches if dark matter is light ( few 10 GeV))... and always independent of astrophysical uncertainties Limits can become stronger or weaker if mediator is light Special case: DM interacting through the Higgs Invisible Higgs searches will be very sensitive Joachim Kopp Collider searches for dark matter 27

39 Thank you!

40 Simulation of the DELPHI detector CompHEP: Event generation Hacked MadAnalysis: Detector response DELPHI, hep-ex/ , arxiv: Joachim Kopp Collider searches for dark matter 29

41 Simulation of the DELPHI detector CompHEP: Event generation Hacked MadAnalysis: Detector response Three EM calorimeters: High Density Projection Chamber (HPC) 45 < θ < 135 xγ = E γ/e beam > 0.06 Trigger efficiency: Eγ = 6 GeV, 30 GeV, 100 GeV Cut efficiency: 6 GeV, 80 GeV Resolution: / E DELPHI, hep-ex/ , arxiv: Joachim Kopp Collider searches for dark matter 29

42 Simulation of the DELPHI detector CompHEP: Event generation Hacked MadAnalysis: Detector response Three EM calorimeters: High Density Projection Chamber (HPC) Forward EM calorimeter (FEMC) 12 < θ < 32 xγ = E γ/e beam > 0.1 Trigger eff.: 10 GeV, 15 GeV Cut efficiency: 10 GeV, 100 GeV Noise/machine bg: 11% loss Resolution: / E 0.11/E) DELPHI, hep-ex/ , arxiv: Joachim Kopp Collider searches for dark matter 29

43 Simulation of the DELPHI detector CompHEP: Event generation Hacked MadAnalysis: Detector response Three EM calorimeters: High Density Projection Chamber (HPC) Forward EM calorimeter (FEMC) Small Angle Tile Calorimeter (STIC) 3.8 < θ < 8 xγ = E γ/e beam > 0.3 Efficiency: 48% Resolution: / E DELPHI, hep-ex/ , arxiv: Joachim Kopp Collider searches for dark matter 29

44 Simulation of the DELPHI detector CompHEP: Event generation Hacked MadAnalysis: Detector response Three EM calorimeters: High Density Projection Chamber (HPC) Forward EM calorimeter (FEMC) Small Angle Tile Calorimeter (STIC) In addition (fudge factors): 90% efficiency fudge factor Lorentzian energy smearing, width E DELPHI, hep-ex/ , arxiv: Joachim Kopp Collider searches for dark matter 29

45 LHC mono-photon limits WIMP nucleon cross section ΣN cm CMS 7TeV, 1.14 fb 1 Mono photon Solid : Observed Dashed : Expected ΧΓ Μ Χ qγ Μ q CoGeNT DAMA q ± 33 CRESST 90 C.L Spin independent WIMP mass m Χ GeV XENON 10 CDMS XENON 100 Joachim Kopp Collider searches for dark matter 30