Constraining CMSSM dark matter with direct detection results

29
Constraining CMSSM dark matter with direct detection results Chris Savage Oskar Klein Centre for Cosmoparticle Physics Stockholm University with Yashar Akrami, Pat Scott, Jan Conrad & Joakim Edsjö JCAP 1104:012, 2011 [arXiv:1011.4318] JCAP 1107:002,2011 [arXiv:1011.4297]

description

Constraining CMSSM dark matter with direct detection results. Chris Savage Oskar Klein Centre for Cosmoparticle Physics Stockholm University. with Yashar Akrami, Pat Scott, Jan Conrad & Joakim Edsjö JCAP 1104:012, 2011 [arXiv:1011.4318] JCAP 1107:002,2011 [arXiv:1011.4297]. Overview. - PowerPoint PPT Presentation

Transcript of Constraining CMSSM dark matter with direct detection results

Page 1: Constraining CMSSM dark matter with direct detection results

Constraining CMSSM dark matter with direct detection results

Chris SavageOskar Klein Centre for Cosmoparticle Physics

Stockholm University

with Yashar Akrami, Pat Scott, Jan Conrad & Joakim EdsjöJCAP 1104:012, 2011 [arXiv:1011.4318]JCAP 1107:002,2011 [arXiv:1011.4297]

Page 2: Constraining CMSSM dark matter with direct detection results

Overview

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Direct detection signalN, Ek=1..N , Sm, , , ...

WIMP parameters m , SI,p , SD,p , SD,n

Direct detection signalN, Ek=1..N , Sm, , , ...

CMSSM (e.g.) parameters m0, m1/2, A0, tanβ, sign(μ)

WIMP parameters m , SI,p , SD,p , SD,n

Phenomenology Particle (SUSY) TheoryMessy parameter space:

statistical scanning requiredWell behaved parameter space:

analytical methods for constraints?

Pheno space notfully mapped out

by CMSSM

This talkExperimental groupsJi-Haeng Huh talk

Page 3: Constraining CMSSM dark matter with direct detection results

Overview

• How will future direct detection results constrain dark matter from supersymmetric theories?

Realistic reconstruction of dark matterproperties using CMSSM as case study

• Outline Basics: CMSSM, direct detection Analysis: likelihoods, statistics and scanning Phenomenological parameter constraints

• Individual/combined experimental results• Statistical/scanning issues• Halo model, hadronic uncertainties

CMSSM parameter constraints9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Page 4: Constraining CMSSM dark matter with direct detection results

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Basics andAnalysis Procedure

Page 5: Constraining CMSSM dark matter with direct detection results

Basics

CMSSM (Constrained Minimal Supersymmetric Standard Model) Simplest SUSY model: four parameters + one sign Complicated parameter space: disconnected regions, sharp peaks,… Results/issues representative of generic SUSY models

(e.g. MSSM-7, BMSSM, NMSSM, etc.)

Direct detection:future ton-scale experiments XENON1T (Xe, neutron odd) [LUX, PANDA-X] CDMS1T (Ge, neutron odd) [EDELWEISS, CRESST?] COUPP1T (CF3I, proton odd)

Not included: CoGeNT, CDEX, DAMA, KIMS -like (higher backgrounds)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

DetectorWIMP

WIMPScatter

Page 6: Constraining CMSSM dark matter with direct detection results

Analysis

Realistic analysis Typical thresholds and efficiencies Finite energy resolution Backgrounds at target levels (~ 2 events), known spectrum Uncertainties in halo model (density, velocity distribution) Hadronic uncertainties: WIMP-quark → WIMP-nucleon couplings

Likelihoods Direct detection

Nuisance parameters Halo model Nucleon structure SM parameters

…also physicality constraints9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

See paper for technical details

COUPP: no spectrum

Number of events (Poisson) Event energies (spectrum)

Page 7: Constraining CMSSM dark matter with direct detection results

Analysis

Procedure Select CMSSM models that give particular m and SI,p :

benchmark models Generate random experimental results Reconstruct CMSSM model by scanning

over CMSSM parameter space• DarkSUSY + SuperBayeS (MultiNest)

Statistics Scan: Bayesian (SuperBayes) Results: Frequentist or Bayesian

• Profile likelihood (frequentist)• Marginalized PDF (Bayesian)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

www.darksusy.orgwww.superbayes.org

Most experimental analyses

Page 8: Constraining CMSSM dark matter with direct detection results

Benchmark Models

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

BM1: low m , high SI,p O(100-400) signal events

BM2: low m , low SI,p O(1-3) signal events

BM3: moderate m and SI,p O(20-30) signal events

BM4: high m , high SI,p O(20-30) signal events

+ 2 background events (on average)

Benchmarks still below most recent XENON constraints

Page 9: Constraining CMSSM dark matter with direct detection results

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Results(Constraints)

Page 10: Constraining CMSSM dark matter with direct detection results

BM1: low m , high SI,p

Spin-independent/dependent cross-sections vs. mass XENON: ~ 200 signal events (~ 7 SD events)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

true value max likelihood posterior mean

Page 11: Constraining CMSSM dark matter with direct detection results

BM1: low m , high SI,p

Spin-independent/dependent cross-sections vs. mass CDMS: ~ 140 signal events (~ 2 SD events)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 12: Constraining CMSSM dark matter with direct detection results

BM1: low m , high SI,p

Spin-independent/dependent cross-sections vs. mass COUPP: ~ 390 signal events (~ 120 SD events)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 13: Constraining CMSSM dark matter with direct detection results

BM1: low m , high SI,p

Spin-dependent couplings: neutron vs. proton an ≈ -ap : CMSSM prediction (not experimental constraint) O(5) [CDMS/XENON] vs. O(100) [COUPP] SD events

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 14: Constraining CMSSM dark matter with direct detection results

BM2: low m , low SI,p

Spin-independent/dependent cross-sections vs. mass ~ 1.4 / 2.1 / 3.0 signal events (~ 0 / 0 / 0.1 SD)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 15: Constraining CMSSM dark matter with direct detection results

BM3: moderate m and SI,p

Spin-independent/dependent cross-sections vs. mass ~ 17 / 23 / 32 signal events (~ 0 / 0 / 0.6 SD)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 16: Constraining CMSSM dark matter with direct detection results

BM4: high m , high SI,p

Spin-independent/dependent cross-sections vs. mass ~ 19 / 25 / 36 signal events (~ 0 / 0 / 0.3 SD)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 17: Constraining CMSSM dark matter with direct detection results

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Issues

Page 18: Constraining CMSSM dark matter with direct detection results

Issue: sampling/coverage Mass constraint from energy spectrum:

degeneracy for heavy WIMPs

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Phenomenological parameter scanPato et al., PRD 83, 083505 (2011)

BM3

BM4

Page 19: Constraining CMSSM dark matter with direct detection results

Issue: sampling/coverage

• Scan points without DD likelihood BM4 in poorly sampled region BM3 in higher sampled region

• Degeneracy: BM3 & BM4 should give

similar DD signals (N, Ei)

• BM4 scan: Good fit around BM3 Nothing to draw scan towards

BM4 region Too few models to properly

evaluate profile likelihood9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Page 20: Constraining CMSSM dark matter with direct detection results

Issue: sampling/coverage

• Real priors and/or effective priors affect scan region

• Scan may miss some regions of interest or cover them too coarsely

• Can lead to significant over/under-coverage of confidence regions (frequentist) or credible regions (Bayesian)

• Possibly improved by higher statistics …if higher statistics gives sharper likelihood contours

(can overcome real/effective priors) Not for previous case

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Page 21: Constraining CMSSM dark matter with direct detection results

Issue: nuisance parameters

• Halo model Local density, velocity distribution Standard Halo Model (SHM): isothermal sphere

• 3 velocity parameters: v0, vobs, vesc

Structure?• Annual modulation (DAMA, CoGeNT)• Directional detection (DRIFT)

• Hadronic matrix elements Used in calculating SI & SD from -quark couplings 6 relevant matrix elements (only 3 are important) Affect CMSSM constraints, not pheno constraints (at least not

directly)9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Halo models + direct detection:see Strigari & Trotta (2009)

and various works by A. Green

See Ellis, Olive & CS, PRD 77, 065026 (2008)

Page 22: Constraining CMSSM dark matter with direct detection results

Halo model uncertainties

With / without uncertainties in halo model (nuisance parameters) Local DM density most significant

• See e.g. Strigari & Trotta, JCAP 11, 019 (2009)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 23: Constraining CMSSM dark matter with direct detection results

Hadronic uncertainties

With / without hadronic uncertainties (nuisance parameters) No change: affects only CMSSM parameter constraints

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 24: Constraining CMSSM dark matter with direct detection results

Hadronic uncertainties

With / without hadronic uncertainties (nuisance parameters) Only directly affects CMSSM parameter constraints

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 25: Constraining CMSSM dark matter with direct detection results

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

CMSSMConstraints

Page 26: Constraining CMSSM dark matter with direct detection results

CMSSM constraints

No direct detection likelihood (priors and nuisance only)

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 27: Constraining CMSSM dark matter with direct detection results

CMSSM constraints (BM1)

With direct detection likelihood Gaugino mass (m1/2) best constrained (related to m) Weaker constraints on m0, A0, tanβ

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

Profile likelihood: ■ 1σ ■ 2σ

Page 28: Constraining CMSSM dark matter with direct detection results

CMSSM constraints

• Can combine with other observational data: Indirect detection: cosmic-rays, neutrinos Accelerators Relic density, etc.

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection

-rays (Fermi-LAT)Segue 1 analysisScott et al. (2009)

Neutrinos (IceCube)IC collab + Edsjö, Scott, CS, in prep.

Accelerator (LHC: ATLAS)SU3 benchmark analysis

Bridges et al. (2010)

See Trotta et al., JHEP 0812:024 (2008)

Page 29: Constraining CMSSM dark matter with direct detection results

Summary

• Examined realistic reconstruction of darkmatter properties in SUSY (e.g. CMSSM)theories using direct detection results

• Can reconstruct WIMP properties reasonablywell in some cases, not so well in others Coverage, sampling issues:

Accuracy affected by scanning technique Nuisance parameters

• Combine DD results with other observationsto better constrain SUSY theory parameters

9/27/2011 C Savage - DSU 2011 - CMSSM and Direct Detection