SUSY Search at CMS - University of Michiganmctp/SciPrgPgs/events/2009/LHC/talks/DarkMatterWorkshopV3(bhatti).pdfSUSY Search at CMS Anwar A Bhatti The Rockefeller University. On behalf

SUSY Search at CMS

Anwar A BhattiThe Rockefeller University

On behalf of CMS CollaborationLHC Dark Matter Workshop

Michigan Center of Theoretical PhysicsJanuary 6-10, 2009

• Jet+MET+0 lepton analysis• Jet+MET+leptons analysis• MET independent analysis• Conclusions

SuperSymmetry

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 2

Avoids fine-tuning of SM, can lead to GUTs Assume LSP is stable ⇒ possible dark matter candidate SUSY breaking mechanism is unknown ⇒ many parameters

A symmetry between fermions and bosons |S=0 or S=1⟩ ↔ |S=½⟩

mSUGRA:supergravity inspired model5 free parameters: m0, m½, A0, tan(β) and sign(µ)

Supersymmetry at CMS

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 3

Final state typically has multiple jets and large missing transverse energy.Cross sections depend on the SUSY parameters, specially masses of squarks and gluinos.

LM160,250GeV

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 4

Benchmark Point LM1

m0=60 GeV, m½=250 GeV, tan(β)=10,A0=0,sign(µ)>0Gluino mass = 600 GeV, squark mass =320 GeV

Inclusive Jet+MET Analysis

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 5

Standrad Model Processes

SUSY LM1 ~50 pb• Require large MET and multi-jets to suppress the SM backgrounds.• Use data-driven techniques to estimate backgrounds.Event Selection:• Cleanup• No leptons (no e,µ, isolated tracks, EM rich jets)• Three leading jets with Pt>180, 110, 30 GeV• HT = PtJet2+PtJet3+PtJet4+MET >500 GeV• MET> 200 GeV

QCD 1010 pb

W+jet (leptons) 7x104 pb

Z+jets (leptons) 7x103 pb

800 pbtt

Missing Transverse Energy

Event Cleanup ≥ 1 primary vertex Activity in Ecal, Hcal and tracker

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 6

Energy: 250 GeV (ECAL)

Halo Muons in CSCs and HB

Not a big issue as long as such events do not overlap with real pp collision.

Cosmic Ray Muon Air Shower

Cosmic Muon

>0.175

> 0.1

Lepton Veto (remove W/Z/tt background)

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 7

QCD Background Events

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 8

MET in QCD is due mis-measured jets and is in the same direction as the jet.

Even after these cuts, QCD is the largest background at LM1.

QCD

SUSY@LM1

1,2 METjet1,jet2

2 21,2 1,2

|1,2

|where ( ) 0.5R

ϕ ϕ

ϕ π ϕϕ = −

= ∆ + − >

Data-driven technique: Matrix Method

BBackground

CSignal+BG

ABackground

DBackground

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 9

Variable X

Varia

ble Y

• For uncorrelated X,Y, background in C: C= D (B/A)

• Signal contamination in A,B,D should be small.• Additional corrections if X and Y are correlated.• Possible pair: MET and Δφ(Jet,Met) min• Working on optimizing the procedure.

Data Driven Technique: Jet Smearing

MET in QCD events arises from the fluctuations in detector response to jets.

The high MET tail in multi-jet events can be determined by smearing the jets in the low MET region using detector response functions.

Response can be measured in well-understood data:• multijet events• dijet events• photon-jet events

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 10

Use response function to predict high MET region

Response Function

Based on earlier ATLAS work

MET (GeV)

PtCaloJet/PtParticleJet

Top Background

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 11

• MET Template: Events passing SUSY cuts but require a lepton and

MT(lepton, MET) <100 GeV• MET distribution same as the background except lepton.

• Normalize the MET template to the data in low MET region 100< MET<200 GeV.• Works quite well if no SUSY contamination in normalization region.• Effect of SUSY contamination in normalization region on the background estimate is under study.

not-identifiedNon-isolated

Semi-leptonic decays are dominant.

Based on earlier ATLAS work Template Data

Norm. Region

MET MET

Irreducible Background: Zνν

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 12

More details in talk by James Lamb, later in the session.

Data-driven Estimation I (Z+jets)Standard Candle: use Z→µµ

Replace leptons by neutrinosTotal uncertainty ~20% for 1 fb-1

Statistically limited • Br(Z→µµ)= 1/6 Br(Z→νν)

Data-driven Estimation II (W,γ+jets)Gain in statistics:σ(Z+2jet)=σ(W+2jet)/3=σ(γ+2jets)/0.8

• Only Z→µµ (ee) are usable. 3.3%Approximation: V+jets events at high pT have similar event shapes.

Z→νν background estimate (100 pb-1)

MC-truth 35

From γ+jet 29±3(stat)±5(sys)

From W+jets 35±10(stat)±8(sys)±3(theoy)

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 13

Discovery Potential

SUSY LM1 discovery possible with < 10 pb-1 of data at √s= 14 TeV.

600GeV 520GeV,g qm m= =

SUSY LM1

Acceptance ~13%

Signal QCD Top pair Z(νν) W/Z Single top6319 107 54 48 33 3

Search for SUSY using leptonic decays

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 14

Concentrating on muons: Relatively easy to identify Good momentum resolution

Small SM backgrounds Low susy cross section but still

good S/B and S/√B

0 02 1g qq qq qqll qqllχ χ→ → → →

Sources of muons:(no particular order)•b/c quarks (semi-leptonic decays)•W/Z/top quark decays•decay-in-flight•τ-leptons• cosmic rays• mis-identified hadrons (punch through)• new physics

Muon identification:•Good track •Good matching muon stub•No additional track around the tracks•No large energy deposit around muontrack in calorimeter.

Inclusive muon+MET+Jets analysis

≥ 1 muon with pT >30 GeV MET>130 GeV At least three jets:

• 440, 440, 50 GeV• optimized η cuts

cos[Δφ(jet1,jet2)]<0.2 Cuts on Δφ(jet1,2,3,MET)

Excellent sensitivity to all LM benchmark points.

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 15

5σ reach contour

Event Yield (1 fb-1)LM1: 300 events (efficiency=0.07%)BG: 3 events with uncertianty~20%

Same-sign Di-muon+MET+Jets Analysis

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 16

LEPTevatron

A0=0, tan(β)=10, sign(µ)=+1

5σ reach contours

Sample Events mSUGRA LM1 17

≤0.3

W/Z+jets ≤0.05

QCD ≤0.01

Expected events for 1 fb-1 @14 TeV

tt

♦ Same sign muon (PT>10 GeV) ≥ 2

♦ ETmiss ≥ 200 GeV

♦ 1st Jet ET ≥ 175 GeV♦ 2nd Jet ET ≥ 130 GeV♦ 3rd Jet ET ≥ 55 GeV

1 2

2 3, 13,10 GeV, 6 GeVIso Iso

nHitsE Eµ µ

µ µ

χ ≤ ≥≤ ≤

01

01

g qqq q

µ νχµ νχ

±

±

→→

Dilepton+Jets+MET, mass edge

Observation and measurement of

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 17

SUSY di-lepton analysis in the talk by Oliver GutscheFriday January 9th 2:00 pm

0 02 1 llχ χ→

MET Independent Search

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 18

Dijet Search

Signal: If , no cascade decays through gluinos• Two jets uncorrelated in Pt and direction• .

BackgroundsQCD dijets: back-to-back, similar pT’s

Z→νν: Irreducible background

W→(e,µ) ν: MET when (e,µ) out of acceptance.

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 19

01

02, 0

1Mainly sensitive to ,... q q q q qχ χ χνν→→ →

q gm m<

MET-independent Dijet Search

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 20

α(Randal-Tucker-Smith) arXiv:0806.1049

Analysis does not rely on missing transverse energy.

Transverse αT

QCD : back to back jets α(αT)≤0.5.SUSY: α (αT ) can be > 0.5.

Event Selection

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 21

HT>500 GeV Third jet veto pt< 50 GeVΔφ(MHT,jet1,2,3)>0.3 rad Lepton veto |η(jet1)|<2.5

Δφ(jet1,jet2)

α αT

1 2 1 2 | |JT

J J JT T TTH p p MHT pp= + = +

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 22

Expected Event Yield

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 23

0 2.5 4

Data-Driven Background Estimation

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 24

Pre-selection (no η cut) + HT>500 GeV

B(Signal+BG) C(BG)

A(BG+some signal)~2 million QCD+~2K

tt/W/Z+~1k SUSYD(BG)

α T

0.55

|η| of lead jet

• Signal is in the central region.• Use the ratio

in forward region to predict thebackground in the central region.

• R= C/D (assumed to be independent of η.)

( 0.55)/ ( 0.55)T TR N Nα α= > <

background( )B A R= ×

Background Estimation

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 25

▲ Simulated signal+BG αT>0.55■ Estimated background αT>0.55

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 26

Signal Topology: Two squarks decaying into two quarks+ two LPS.• Squark-squark dominates• Squark-gluino contributes when

Background: QCD :0, Z→νν 57, W/Z 19, Total 86

g qq→

Summary

The results from global run data look very good. • Rate of events with large MET is low.

The dijet analysis looks very promising. Ready to analyze the real collider data.

Anwar Bhatti LHCDM@MCTPJan 6-10, 2009 27

BACKUP

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 29

CMS: Compact Muon Solenoid

Muons: • |η|<2.4 momentum resolution: dpT/pT~1% (pT~25 GeV) Calorimetry:• Hadron |η|<5.0, δE/E ~ 70% / √E + 8%• EM |η|<3.0, δE/E ~ 2.8% / √E + 0.3% + 12% / E

MUON BARRELDrift TubeChambers ( DT )

Resistive PlateChambers ( RPC )

SUPERCONDUCTINGCOIL

IRON YOKE

Silicon MicrostripsPixels

TRACKER

Cathode Strip Chambers (CSC )Resistive Plate Chambers (RPC)

MUONENDCAPS

CALORIMETERSECAL

ScintillatingPbWO4 crystals

HCALPlastic scintillator/brasssandwichTotal weight: 12,500 t

Diameter: 15 mOverall Length: 22 mMagnetic field: 3.8 Tesla

30Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP

CDF limits

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 31

tan(β)=5, A0=0 and μ<0.

Signal: If , no cascade decays through gluinos• Two jets uncorrelated in Pt and direction• .

BackgroundsQCD dijets: back-to-back, similar pT’s

Z→νν: Irreducible background

W→(e,µ) ν: MET when (e,µ) out of acceptance.

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 32

01

02, 0

1Mainly sensitive to ,... q q q q qχ χ χνν→→ →

Fraction of events passing selection cuts

Process Pt of third jet (GeV)<30 <50 <70

80 61 51

18 34 44

1 3 5

qq qg gg

q gm m<

Selection cuts described later.

Backgrounds

Sources of background QCD:

Seems to be under control but huge cross section Uncertainties due to higher order QCD effects missing in Monte Carlo

Third jet veto should minimize higher order effects.

W, Z, tops Zνν:

Represents an irreducible background Two jets+ real missing ET

Ideally, Zµµ events can be used but not enough statistics. Other control samples W+jets Photon+jets

Jan 6-10, 2009 Anwar Bhatti LHCDM@MCTP 33