Search for 3rd generation superpartners with the ATLAS experiment - PowerPoint PPT Presentation

Search for 3rd generation superpartners with the ATLAS experiment Keisuke Yoshihara (University of Pennsylvania) DPF2017 July 31 (Fermilab)

Introduction The SM of particle physics is incomplete. Supersymmetry can be a new theory solving various problems remained in the SM. 1. Higgs mass divergence 
 λ f h h at Planck scale due to 
 radiative corrections 
 Λ : UV cut-off ~ Planck scale s (scalar partner) (Hierarchy problem) ~ ~ h h λ s -> t and b are a key 
 (large Yukawa coupling) arXiv:1110.6926 2. Naturalness (Natural sparticle mass ˜ g stop mass (1-loop order) SUSY) suggests the − m 2 presence of light 3rd gen. ˜ ˜ t L t R Z = | µ 2 | + m 2 ˜ b L squarks together with the H u 2 ˜ H higgsino LSP(s). 2 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Large Hadron Collider (LHC) LHC was constructed to perform various searches (Higgs boson and BSM physics) at TeV energy scale. 3 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Challenging environment at LHC • The cross section for the SUSY proton - (anti)proton cross sections is very small. 9 9 10 10 MSTW 2008 NLO PDFs 8 8 10 σ σ σ σ tot 10 HE 7 7 10 10 Tevatron LHC LHC 20 pile-up 2 pile-up 6 6 10 10 -1 5 5 10 10 -2 s σ σ b σ σ 33 cm 4 4 10 10 3 3 10 10 events / sec for L = 10 jet > √ σ σ σ σ jet (E T √ √ s/20) √ 2 2 10 10 ( nb ) ) ) ) σ σ W σ σ 1 1 10 10 σ ( ( ( σ Z σ σ σ σ σ σ 0 0 10 jet > 100 GeV) 10 σ jet (E T σ σ σ • As the luminosity increases, number -1 -1 10 10 of interactions per crossing (pile-up -2 -2 10 10 σ WW σ σ σ -3 -3 or μ ) and detector occupancy 10 10 σ σ σ σ t σ σ σ ZZ σ -4 -4 10 σ σ ggH σ σ 10 increases. { σ σ σ WH σ M H =125 GeV -5 -5 10 10 σ VBF σ σ σ -6 -6 • Collecting important physics 10 10 WJS2012 -7 -7 10 10 events in this difficult environment 0.1 1 10 √ √ s (TeV) √ √ is a key at the LHC. m t ~1 TeV ~ 4 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

SUSY decay and production at LHC Not reviewed, SUSY production and decay ~ �� �� �� � � , ˜ g/q q � ˜ t b �� �� ~ ~ �� � g/q ~ �� �� �� � � � � g/q q � � � Basic Event topologies of SUSY ! • The stop/sbottom is pair-produced (in RPC scenario) from the gg ~ ~ process if the mass is light (m t,b < 1 TeV). As the SUSY mass goes high, the qq contribution gets larger (PDF is very steep). � ~ ~ • The stop/sbottom decays into intermediate states ( χ 20 , χ 1± ) if exists, �� ~ otherwise the stop/sbottom decays directly into the LSP ( χ 10 ). 5 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Search strategy t 1 , (˜ t 1 , ˜ ˜ ˜ ˜ ˜ b 1 ) b 1 , t 1 , t 1 , sparticle masses χ ± 0 χ 0 1 , ˜ 1 , ˜ 2 , χ ± 0 χ 0 χ 0 1 , ˜ 1 , ˜ 2 , ˜ 3 χ ± χ 0 χ 0 1 , ˜ 1 , ˜ 1 , ˜ 2 , χ 0 χ 0 1 , ˜ 1 , ˜ 1 , 1 , χ 0 1 , ˜ 1 , a) pure bino LSP P b) wino NLSP P c) higgsino LSP P d) bino / higgsino mix • Various pMSSM (or simplified) models are built to cover the various physics models (GUT, Naturalness, etc…) and the LSP scenarios. • The event selection is optimized for the various final states, 
 ~ ~ ~ ~ e.g. t -> t χ 10 , b χ 1± , b -> b χ 10 , t χ 1± … • Both RPC and RPV stop/sbottom searches are performed in ATLAS. This talk focuses on RPC scenario . 6 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Bino LSP models Pure bino LSP (simplified) model: New technique: 
 1 [GeV] ˜ t 1 , BDT and shape-fit > ∆ m = m ˜ t 1 − m ˜ Sparticle masses χ 0 m 1 χ 0 m ˜ t 1 ] m ˜ ∆ m b 0 > m t > + m > ∆ m m ˜ t m W m m ˜ t 1 < m ˜ ∆ χ 0 ∆ 0 100 > 1 > χ 0 m + χ 0 t 1 → bff 0 ˜ 1 m ∆ t 1 → bW ˜ 1 χ 0 ∆ m W 0 Decay is governed t 1 → c ˜ 1 ˜ 1 χ > t m → ˜ ˜ ˜ ˜ 1 ~ ~ t by Δ m(t 1 , χ 10 ) . χ 0 1 , ˜ 1 , ~ 0 0 100 200 0 100 200 300 ] m ˜ t 1 [GeV] t + χ 10 a ) Pure Bino LSP ~ ~ Wino NLSP ( m( χ 1± )~ 2m( χ 10 )) (pMSSM) model: GUT (cMSSM/mSUGRA) b t t 1 , ˜ ˜ b 1 , ~ W h b+ χ 1± signature: 
 p p Sparticle masses ˜ ˜ t 1 t χ 0 χ 0 ˜ ˜ 1 1 χ ± χ 0 ˜ ˜ 2 χ ± 1 high pT b-jets, jets, 0 χ 0 1 , ˜ 1 , ˜ 2 , χ 0 χ ⌥ ˜ ˜ 2 1 χ 0 χ 0 ˜ ˜ ˜ ˜ t 1 t 1 1 p p and large MET Z W χ 0 1 , ˜ t b 1 , ~ ~ b + χ 1± t + χ 20 b) Wino NLSP 7 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Higgsino LSP models ATLAS-CONF-2017-037 Higgsino LSP (simplified) model: Naturalness Two benchmark models: ˜ t 1 , Events / 1 GeV ATLAS Preliminary Data Total SM 500 -1 s = 13 TeV, 36.1 fb t t 2L t t 1L ~ ~ Preselection (soft lepton) W+jets Single top a) Δ m( χ 10 , χ 1± ) = 5 GeV t t +V Diboson 400 Sparticle masses ~ ~ 300 b) variable Δ m( χ 10 , χ 1± ) = 0-30 GeV 200 100 Signature: 
 χ ± χ 0 χ 0 1 , ˜ 1 , ˜ 1 , ˜ 0 Data / SM 2 , 1.5 1 Soft-objects and large MET 0.5 5 10 15 20 25 30 c ) Higgsino LSP lepton p [GeV] T Well-tempered (pMSSM) model: DM relic density ( Ω h2 ~ 1.12) t 1 , (˜ ˜ b 1 ) • Admixture LSP (bino/higgsino) satisfying M 1 ~ -|µ| Sparticle masses ~ ~ • Typical Δ m( χ 10 , χ 1± ) ~ 20-50 GeV χ ± 0 χ 0 χ 0 1 , ˜ 1 , ˜ 2 , ˜ 3 • Interpretation only (no event selection optimized) χ 0 1 , ˜ 1 , d ) Bino / Higgsino mix 8 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Top quark reconstruction ATLAS-CONF-2017-037 • In the decay of heavy stop, the top quark is highly boosted. As a consequence jets from the top decay tends to form a large radius jet. Events / 10 GeV Data Total SM ATLAS Preliminary -1 t t 2L t t 1L s = 13 TeV, 36.1 fb Single top W+jets miss Preselection (high E ) 4 10 Others T q R=1.0 3 10 q q t q t 2 10 b b Data / SM 1.5 Resolved top Boosted top 1 0.5 100 150 200 250 300 350 reclustered m [GeV] top Events / 20 GeV • The analysis benefits from reconstructing hadronically decaying top quark (“hadronic top reconstruction”). 9 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Background estimate ATLAS-CONF-2017-037 30 Events / 40 GeV ATLAS Preliminary Data -1 s = 13 TeV, 36.1 fb Total SM tN_med ttZ CR 25 t t +V CR Diboson Variable B Single top 20 Others SR 15 ttZ(ll) CR 10 5 0 Data / SM 1.5 1 0.5 0 100 200 300 400 500 600 Variable A (ll) p [GeV] p T of Z(ll) [GeV] T • Use control region (invert one or two SR selections) • Simulation uncertainties (PS+hadronization, hard- scattering, PDF, …) need to be assessed and propagated when extrapolating the norm to the SR. • Minor backgrounds are normalized to the SM prediction. 10 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Results: Validation region ATLAS-CONF-2017-037 Validation Regions Events ATLAS Preliminary Data Total SM -1 s = 13 TeV, 36.1 fb Signal regions t t 2L t t 1L 3 10 W+jets t t +V Single top Diboson 2 10 Blinded 10 SR 1 tot 0 0.077 0.154 0.231 0.308 0.385 0.462 0.538 0.615 0.692 0.769 0.846 0.923 1 2 σ ) / exp 0 - n 2 − obs (n TVR WVR TVR T1LVR T2LVR WVR T1LVR T2LVR WVR bffN bWN tN_med tN_high bffN bffN bWN tN_med tN_med tN_med tN_high tN_high tN_high • VRs are monitored while blinding SRs. 11 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)

Results: Signal region ATLAS-CONF-2017-037 Validation Regions Events ATLAS Preliminary Data Total SM Signal -1 s = 13 TeV, 36.1 fb Signal regions t t 2L t t 1L 3 10 W+jets t t +V Region Single top Diboson 2 10 10 1 tot 0 0.077 0.154 0.231 0.308 0.385 0.462 0.538 0.615 0.692 0.769 0.846 0.923 1 2 σ ) / exp 0 - n 2 − obs (n TVR WVR TVR T1LVR T2LVR WVR T1LVR T2LVR WVR bffN bWN tN_med tN_high bffN bffN bWN tN_med tN_med tN_med tN_high tN_high tN_high • No significant excess is observed. 12 DPF2017, July31 2017, Keisuke Yoshihara (University of Pennsylvania)