COmpact Detector for EXotics at LHCb: CODEX-b

Transcription

1 COmpact Detector for EXotics at LHCb: CODEX-b Vladimir V. Gligorov, Simon Knapen, Michele Papucci, Dean Robinson LHCb Implications Workshop, CERN Nov 2017 Based on: Leonardo da Vinci: Codex Leicester (V. Gligorov, S. Knapen, M. Papucci, & DR)

2 LLPs are generic! Common consequence of: small couplings or; scale (or loop) hierarchies or; small mass splittings R-parity violation Gauge mediation (mini-)split SUSY stealth SUSY Asym. Dark Matter Freeze-in Composite Dark Matter Baryogenesis Neutrino masses Flavor puzzle Neutral Naturalness Hidden Valleys LLP lifetimes as long as τ 1 s are broadly consistent with BBN Masses can plausibly range from sub-mev to O(100 GeV). A large parameter space to explore! Dean Robinson dean.robinson@uc.edu CODEX-b 2 19

3 LHC coverage ATLAS/CMS will set best limits for charged, colored and/or heavy LLPs LHCb: probes O(GeV) neutral LLPs with significant muon BR and cτ VELO scale: can trigger on softer µ s and softer DVs. In some cases, ATLAS, CMS, LHCb coverage is already complementary (See Carlos talk) Pierce, Shakya, Tsai, Zhao: Large BGs in hadronic collisions: lighter, neutral, longer-lived LLPs are hard for them to see! Dean Robinson dean.robinson@uc.edu CODEX-b 3 19

4 Complementarity There is a growing landscape of proposals for LLP searches: including MATHUSLA, MilliQan, FASER and CODEX-b. There are no good theory priors in the huge space of allowed LLP branching ratios, lifetimes or masses. Hard to build a single detector to cover all these cases! If we are building a LLP detection programme, rather than an exclusion programme, we need data from multiple experiments with decorrelated backgrounds Complementarity is key for an LLP detection in the broadly unexplored parameter space! Dean Robinson dean.robinson@uc.edu CODEX-b 4 19

5 LLP Prelude CODEX-b Setup NP Benchmark Reach Next Steps Dean Robinson CODEX-b 4 19

6 x LHCb Cavern DELPHI DAQ LHCb IP8 Dean Robinson CODEX-b 5 19

7 x LHCb Cavern Pre-Run 3 (2020): Data AcQuistion will be moved to surface DELPHI CODEX-b box SM SM ϕ shield veto UXA shield Pb shield IP8 General strategy: Look for decays-in-flight of LLPs from IP8 Dean Robinson dean.robinson@uc.edu CODEX-b 5 19

8 Instrumentation As a proof-of-concept: Fiducial volume ( the box ) is m; angular acceptance 1%. 6 (RPC) tracking layers on all faces 5 sets of 3 vertical tracking layers equally spaced in box 1cm strip granularity 10m Tracking simulation: Tracking effs are all O(1) for benchmarks we consider (more later) 10m Dean Robinson dean.robinson@uc.edu CODEX-b 6 19

9 Capabilities and possibilities Distance is only 4 bunch crossing times for relativistic objects: Integrate CODEX-b into the DAQ & readout, and treat as subdetector Identification and at least partial reconstruction of the LLP event. E.g. tag a VBF jet for Higgs decays, or an associated K ( ) for B decays. Phase II pileup is manageable with precise enough timing info Modest size of the fiducial volume: Consider more ambitious detection technologies such as calorimetry or time-of-flight Momentum reconstruction and particle identification Aids confirmation of a discovery! Precision timing and spatial resolution, 100 ps or futuristic 50 ps resolution possible Required for LLP mass reconstruction Dean Robinson dean.robinson@uc.edu CODEX-b 7 19

10 Primary Backgrounds Primary muons may scatter on air. Attenuated with extra shielding, and remainder vetoable by front tracking faces Air scatter µ UXA shield Primary neutral hadrons suppressed with additional shielding Pb shield K 0, n,... Primary: suppressed K 0, n,... UXA shield (In practice Pb is not ideal for neutrons: other materials to be considered) ν-air inclusive inelastic σ: O(3) events, but actual fake rate likely much smaller. Dean Robinson dean.robinson@uc.edu CODEX-b 8 19 Pb shield

11 Secondary Backgrounds Muon or neutron secondary production in shield can be large! Vetoable reducible by active veto in shield K 0, n,... Reducible: vetoed µ shield veto UXA shield Pb shield Active veto placed so that active veto eff/rejection rate is minimized, while neutral irreducible BGs are suppressed µ K 0, n,... Irreducible: suppressed shield veto UXA shield Pb shield Dean Robinson dean.robinson@uc.edu CODEX-b 9 19

12 Geant4 (20 + 5)λ simulation To estimate BGs we use a (preliminary) Geant4 simulation. Includes muons, kaons, pions, neutrinos, neutrons, gammas, protons,... BG species irreducible by shield veto Particle yields reducible by shield veto Baseline Cuts n + n Ekin > 1 GeV K 0 L Ekin > 0.5 GeV π ± + K ± Ekin > 0.5 GeV ν + ν E > 0.5 GeV Events/bin L = 300 fb 1 Incident μ Flux IP μ Flux E kin [GeV] These are yields not scattering rates! n air scattering prob. 5% Normalization is set for min bias σ 100 mb Muon-air interactions can be vetoed using front detector faces Shield veto event rejection rate 10 4 No use (yet) of timing or spatial information Estimates validated with simplified propagation model using muon CSDA and kaon scattering length, and scattering/muoproduction cross-sections, from data Events/bin L = 300 fb 1 K L ε veto = 10 5 K L irred. n ε veto = 10 5 n irred E kin [GeV] Dean Robinson dean.robinson@uc.edu CODEX-b 10 19

13 Data-driven BG calibration Cosmics will be used for spatial & time detector alignment. (Signal is from horizontally displaced source, not vertical!) Backgrounds can be measured by putting a small telescope in the LHCb cavern Measure background rates with different shield thicknesses Is being considered for an engineering run well ahead of full detector construction. Looks promising to do very soon! Dean Robinson dean.robinson@uc.edu CODEX-b 11 19

14 LLP Prelude CODEX-b Setup NP Benchmark Reach Next Steps Dean Robinson CODEX-b 11 19

15 Benchmark Scenarios Consider two benchmark LLP scenarios Spin-1 massive gauge boson γ d produced by Higgs decays h γ d γ d γ d decays via kinetic mixing with SM hypercharge light O(GeV) scalar ϕ produced in inclusive B decays b sϕ via Higgs mixing portal ϕ decay through same Higgs portal Coming soon: h dark glueballs, twin Higgs mixing portal Suggestions for other portals/signatures welcome! Dean Robinson dean.robinson@uc.edu CODEX-b 12 19

16 b sϕ: Higgs-scalar mixing Single parameter portal: Higgs-scalar mixing angle, θ, controls production rate and lifetime 10 6 LHCb, 3 fb CHARM LHCb, 300 fb 1 sin 2 θ CODEX-b SHiP MATHUSLA m ϕ (GeV) Large theory uncertainties for m ϕ 1 GeV! (cf. Evans ) Blue dashed includes tracking sim; dot-dashed for L = 1/ab Dean Robinson dean.robinson@uc.edu CODEX-b 13 19

17 b sϕ: Higgs-scalar mixing Single parameter portal: Higgs-scalar mixing angle, θ, controls production rate and lifetime short lifetime regime: ϕ s decay before reaching box 10 6 LHCb, 3 fb CHARM LHCb, 300 fb 1 sin 2 θ CODEX-b SHiP LHCb downstream tracking? MATHUSLA long lifetime regime: B production limited m ϕ (GeV) Large theory uncertainties for m ϕ 1 GeV! (cf. Evans ) Blue dashed includes tracking sim; dot-dashed for L = 1/ab Dean Robinson dean.robinson@uc.edu CODEX-b 13 19

18 General b sϕ Reach Relax constraint between lifetime and inclusive BR Br[B X s φ] B X s inv m φ = 0.5 GeV LHCb B K(φ μμ) m φ = 1 GeV LHCb B K(φ μμ) m φ = 2 GeV m φ = 3 GeV cτ [m] Max efficiency at cτ 10 m: parent B s and daughter ϕ s only mildly boosted. Dean Robinson dean.robinson@uc.edu CODEX-b 14 19

19 h γ d γ d : Higgs-Dark photon portal m γd = 0.5 GeV m γd = 10 GeV 10 1 Br[h γ d γ d ] ATLAS 2DV CODEX-b CODEX-b & μsh 1 ab 1, m MATHUSLA cτ [m] ATLAS 1DV ATLAS 2DV CODEX-b CODEX-b & μsh 1 ab 1, m cτ [m] For m ϕ = 0.5 GeV, ATLAS reach is sys limited: Assume factor of 5 improvement. (Could scale as much as L) h inv reach anticipated to be few% Dean Robinson dean.robinson@uc.edu CODEX-b 15 19

20 Tracking Efficiencies Implement a tracking simulation for above geometry: Six hits required for a track Assume sensitivity down to 600 MeV momentum Dominated by assumption p > 600MeV. Needs proper simulation of turn-off. cτ (m) m ϕ [B X s ϕ] m γd [h γ d γ d ] Dominated by opening angle resolution. Requires optimization using station spacing and granularity Lesson: Proof-of-concept tracking effs are O(1). Can be further optimized. Dean Robinson dean.robinson@uc.edu CODEX-b 16 19

21 Boost Resolution Reconstruct parent boost from the measured decay vertex (no timing!), assuming 2-body decay with relativistic products (only need spatial info!) β 1 θ1 θ2 β φ 0.5 GeV, B X s φ 2.0 GeV, B X s 0.5 GeV, H γ γ d d 20.0 GeV, H γ γ d d 2 10 LLP, β = sin(θ1+θ2) sin θ1+sin θ see also Curtin & Peskin: βγ resolution (%) The resolution is < 1% dominated by distance to first measured point, not detector granularity Boost distribution is dominated by the spread of boosts, not resolution. Dean Robinson dean.robinson@uc.edu CODEX-b 17 19

22 Boost and Mass Reco GeV, 10m 1.0 GeV, 10m 2.0 GeV, 10m GeV, 1m 1.2 GeV, 1m 5.0 GeV, 10m 20.0 GeV, 10m ln(βγ ) reco ln(βγ ) reco Mild resolution even for b sϕ! For b sϕ, use time-of-flight to reconstruct LLP mass. Assume 100 ps and 50 ps resolution per hit. LLPs slow enough for mild mass reconstruction! GeV, 10m 1.0 GeV, 10m 2.0 GeV, 10m GeV, 10m 1.0 GeV, 10m 2.0 GeV, 10m m φ (GeV} m φ (GeV} Dean Robinson dean.robinson@uc.edu CODEX-b 18 19

23 Thoughts & Next Steps No showstoppers so far. CODEX-b can significantly enhance the NP reach and capabilities of the LLP programme, complementing or exceeding the reach of other LHC experiments, with largely decorrelated backgrounds Lots to do: Develop a more realistic proposal for the detector, including BG shield analysis and tracking setup/other technologies, and achievable resolution for reconstruction Develop the NP physics case for other models Participation welcome! Thank you! Dean Robinson dean.robinson@uc.edu CODEX-b 19 19

24 Extras and Details Dean Robinson CODEX-b 19 19

25 Why do we care about LLPs? Long lived particles are generic consequence of theories with: Small couplings Scale (or loop) hierarchies Phase space suppression broken sym weak mixing/ marginal operator technically natural m M, typically n 4 loop factors ( ) n m Γ ε 2 PS M SM provides a template in weak decays: Multiple scales (G F ) Approx symmetries (e.g. isospin) 3-body final states squeezed spectra approx sym multibody decays Dean Robinson dean.robinson@uc.edu CODEX-b 20 19

26 LHCb (projected) reaches Two example portals: Generically not excluded: Applies for light inflaton (Bezrukov and Gorbunov ) Higgs mixing portal: B K ( ) (ϕ µµ) Dark photon portal: D DA LHCb: Reach in cτ is limited by size of VELO/TT distance and/or statistics Ilten, Soreq, Thaler, Williams, Xue: Dean Robinson dean.robinson@uc.edu CODEX-b 21 19

27 Some LLP Proposals MilliQan: Image: D.Curtin & R. Sundrum CODEX-b physics reach will be most directly comparable to MATHUSLA. The LLP detection strategies and technologies necessarily have various commonalities, though the backgrounds and configuration will differ in several critical aspects FASER: Dean Robinson CODEX-b 22 19

28 LLP Reach Intuition Number of LLP decay vertices lumi 300/fb or more! N box = L LHCb (σ Br) pp ϕx ε box Fiducial efficiency: portal dependent sensitive to η, φ, β distribution ε box = box dv 2πr 2 cτ dβ w(β, η) e r/(cτβγ) βγ for distance cτ βγ ε depth /cτ βγ linear suppression at long lifetimes for distance cτ βγ ε e rnear/cτ βγ exponential suppression at short lifetimes Dean Robinson dean.robinson@uc.edu CODEX-b 23 19

29 Other Capabilities and Possibilities If DELPHI is removed, fiducial volume doubles! Precision timing and spatial resolution, 100 ps or futuristic 50 ps resolution possible Multi-track signatures could be very compelling signals of NP Can fully reconstruct events even with one light invisible state: semileptonic ϕ f lν or ϕ ϕ ff Empty space ( muon shadow ) between IP8 and UXA shield might also be exploited if BR(LLP µµ) 1. box rad. shield µ shadow µ + ϕ µ µ veto Pb shield IP8 Dean Robinson dean.robinson@uc.edu CODEX-b 24 19

30 b sϕ Single parameter portal: Higgs-scalar mixing angle, θ Inclusive b sϕ branching ratio Br[B X s ϕ] 6. s 2 θ (1 m 2 ϕ/m 2 b) 2 ϕ width also set by s 2 θ, from data-driven estimate (Fradette and Pospelov ). Somewhat large theory uncertainties for m ϕ 1 GeV! (cf. Evans ) B distribution generated with Pythia8 hardqcd; inclusive decays modelled by B Kϕ exclusive bb production cross-section 500 µb Dean Robinson dean.robinson@uc.edu CODEX-b 25 19

31 h γ d γ d Higgs-Dark photon portal: yhf µνf µν + ɛf µνb µν Br[h γ d γ d ] and γ d lifetime controlled by separate parameters γ d branching ratios fixed by e + e data GF Higgs production, simulated with Pythia8 Daughter muons can be quite hard, can use muon shadow if Br[γ d µµ] is significant Dean Robinson dean.robinson@uc.edu CODEX-b 26 19