Inclusive SUSY Searches at CMS with Emphasis on Detector Systematics R. Cavanaugh (on behalf of CMS) University of Florida SUSY06 • CMS Detector • Background Calibrations • Inclusive Search Strategies • Detector Systematics • Results
Inclusive SUSY Searches at CMS with Emphasis on Detector Systematics
R. Cavanaugh(on behalf of CMS)
University of FloridaSUSY06
• CMS Detector• Background Calibrations• Inclusive Search Strategies• Detector Systematics• Results
14 June, 2006 R. Cavanaugh, Florida, SUSY06 2
CMS Detector
MUON BARRELDrift TubeChambers ( DT )
Resistive PlateChambers ( RPC )
SUPERCONDUCTINGCOIL
IRON YOKE
Silicon MicrostripsPixels
TRACKER
Cathode Strip Chambers (CSC )
Resistive Plate Chambers (RPC)
MUONENDCAPS
Total weight : 12,500 tOverall diameter : 15 mOverall length : 21.6 mMagnetic field : 4 Tesla
CALORIMETERSECAL
Scintillating PbWO4 crystals
HCALPlastic scintillator/brasssandwich
14 June, 2006 R. Cavanaugh, Florida, SUSY06 3
SUSY Signature: MET + Jets + …
• Squark gluinoproduction
• Full Geant4 Detector Simulation
• 6 hard jets• leptons• 2 LSPs + 4 ν’s
14 June, 2006 R. Cavanaugh, Florida, SUSY06 4
Jet/MET Reconstruction Performance• Jets
• Low luminosity Pileup included• ET Resolution
• Stochastic term ≈ 125% / √ET• Constant term ≈ 3%
• Angular Resolution• High ET Jets: better than calo cell size
(∆φ x ∆η = 0.087 x 0.087)
• Missing Transverse Energy• Low luminosity Pileup included• from QCD
• Stochastic term ≈ 123% / √ΣET• ≈1700 GeV ΣET ≈700 GeV PT dijets
≈50 GeV observed MET• MET φ Resolution
• Low MET : approaches Jet size• High MET : approaches calo cell size
QCD
MET
Jets
CMS
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 5
MET Cleaning from Tevatron• MET is very powerful
SUSY discriminator• Difficult part is to convince
yourself that there is a real excess!
• Tevatron teaches us • MET is not easily
understood!• Non-collisional
backgrounds• Beam halo• Cosmic muons
• Detector Effects• Instrumental Noise• Hot/dead channels (DQM) D. Tsybychev, Fermilab-thesis-2004-58
Run IIV. Shary CALOR04
Run IIjunk
jets
e/γ
14 June, 2006 R. Cavanaugh, Florida, SUSY06 6
Early Study of MET Cleaning in CMS(of course, Real Data will be different!)
• Apply clean up cuts to remove fake high MET events (inspired by CDF & D0)
• ≥ 1 central jet (|η| 0.1 (Event Electromagnetic Frac.)• Fch > 0.175 (Event Charged Fraction)
• Affect on SUSY Signal
CMS Response to Beam Halo Simulationof LHC Point 5
tt full sim.
Pileup not included
Pileup included
CMS
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 7
QCD Multijet Background• Dijets typically back to back
• MET from jet E mismeasurement• Suppress by requiring
• Well separated Jet & MET objects• Typically ≥ 3 jets• Cut on HT (~ 2 pHatT)
• Muon triggers (include isol.)• helps…a lot!
• Prescaled jet triggers • extract the low ET shape and
normalisation directly from data
No cuts
“µ Trigger”
“µ Trigger” + ETJet1>900 GeV
“µ Trigger”+ ETJet1>900 + MET>200 GeV
Level 1
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 8
Electroweak Multijet Backgrounds: Z→µµ Standard Candle
• Large MET and ≥ 3 Jets expected from
• Z(→νν) + ≥ 3 jets• W(→µ(e)ν) + ≥ 3 jets • W(→τν) + ≥ 2 jets
• Z + n-Jets x-sect ∝ αsN• Measure from ≥ 2 Jets Data
• Z(→µµ) + ≥2 jets• Z(→ee) + ≥2 jets
• Normalise MC to Data for ≥ 3 Jets• Assume lepton universality • For W + n-jets, use
• Reduces / Avoids Systematics due to• QCD Scale, PDFs (possibly), ISR/FSR,
Jet Energy Scale, etc• Major Syst. Become
• Luminosity, Measurement of R, Uncertainty on ρ(Njet)
• Still requires tuning MC to Data for kinematic dists.
• 5% precision (~lumi) expected to be achieved with 1.5 fb-1
CMS
Z(→νν) + ≥ 2 jets
Z(→µµ) + ≥ 2 jets (Z peak normalised)Z(→µµ) + ≥ 2 jets (Z peak)
CMS
See Marc Buehler’s talk from D0, yesterday
14 June, 2006 R. Cavanaugh, Florida, SUSY06 9
Benchmark Test Points• Basis of detailed studies in
soon to be released CMS Physics TDR Vol. 2
• Low mass points for early LHC running but outside Tevatron reach
• High mass points for ultimate LHC reach
• Indirect constraints from WMAP for strict mSUGRAexclude most except LM1, 2, 6, 9
BenchmarkOptimisation Point
14 June, 2006 R. Cavanaugh, Florida, SUSY06 10
Inclusive Jet + MET Search• Selection Criteria
• MET>200 GeV + Clean-up• ≥ 3 jets:
• ET> 180, 110, 30 GeV• Indirect lepton veto • Cuts on ∆φ between jets and MET • HT/Meff=ET1+ET2+ET3+MET>500 GeV
• Results:• LM1 efficiency is 13%, S/B ~ 26 :
Number of events (below) for 1 fb-1
• ~6 pb-1 for 5σ discovery
• Lower jet multiplicity requirement reduces sensitivity to higher-order QCD corrections
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 11
Inclusive Muon + Jet + MET Search• Add muon ⇒ clean trigger• Cuts (optimize @ LM1)
• ≥1 isolated muon• pT > 30 GeV
• MET > 130 GeV• ≥3 jets:
• ET> 440, 440, and 50 GeV
• |η|< 1.9, 1.5, and 3
• Cuts on ∆φ between jets and MET
• Backgrounds (10 fb-1)
• LM1 Signal (10 fb-1)• 311 events
Single-µ“OR”Di-µ
pT of leading muon (GeV)
Trig
ger E
ffici
ency
mSUGRALM-1
No Cuts
HLT +Pre-selection
SUSY LM1
SM Backgrounds
CMS
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 12
Inclusive SS Dimuon + Jet + MET• Even cleaner signature
• Low background due to same sign requirement
• Concentrate here on• Identifying the SUSY diagrams
giving prompt muons• Strong muon isolation & tight
quality cuts • Selection Critera• Muon trigger• Muon isolation• Muon track parameters• High PT jets• Large Missing Transverse Energy
• Background (10 fb-1)• 1.5 (ttbar) events
• LM1 Signal (10 fb-1)• 341 events
• 65% efficient at identifying SUSY diagrams, 90% pure
q% µ±
µ±p+
Xp+
q
q µν
q′
1χ±%
ν%g%
q%
01χ%
01χ%
1χ±%
ν%
µν
14 June, 2006 R. Cavanaugh, Florida, SUSY06 13
Jet Energy Calibration/Systematics• Direct photon production:
• qg → qγ (90%) qqbar → gγ (10%) • pT(Jet) = pT(γ)
• use peak position to eliminate effect of tail from ISR
• Estimated Jet Energy Scale Uncertainty:
• Between 3% and 10% for PT ∈ [20, 50] GeV
• ttbar WWbb jjlνbb• Rescale jet with relative energy shift ∆C
• Fit resulting W mass spectrum & constrain to world avg.
• mW(∆C|data) = MWPDG• Compare with Monte Carlo
• For ~6 fb-1: ∆Cmeas = -14.96 ± 0.26 % (∆Ctrue = -14.53 %)
• Requires well understood b-tag (tracker)• Limited by pileup syst. uncertainty: 3%
measured shift ∆Cmeas
∆C(%)
t t
JES Systematic Uncertainty for PT > 50 GeV
γ jet
CMS
CMS
14 June, 2006 R. Cavanaugh, Florida, SUSY06 14
MET Shape Systematics
CMS
• Study effect of non-Gaussian tails in jet ET resolution contributing to fake MET• Approx. 15% of all
jets are mismeasured• Exaggerate
non-Gaussian Tails • Weight each jet (up to 3)
in event• Jet in Non-Gaussian tail: 1.15 • Jet in Gaussian peak : 1.00
• Combine into one event weight• Three different scenarios
• 3 jets under measured• 2 jets under measured• 1 jet under measured
• Overall Systematic Effect : ≈7%
t t
14 June, 2006 R. Cavanaugh, Florida, SUSY06 15
Expected CMS Reach for 1 fb-1
14 June, 2006 R. Cavanaugh, Florida, SUSY06 16
Expected CMS Reach for 10 fb-1
14 June, 2006 R. Cavanaugh, Florida, SUSY06 17
Conclusion• CMS has recently completed several inclusive SUSY
analyses for potential discovery• Full detector simulation, reconstruction• All backgrounds included
• Estimate low PT QCD from pre-scaled jet triggers• Estimate EW from Z→µµ Standard Candle
• Systematic uncertainties• Jet Energy Scale, MET Shape, Misalignment, etc
• Results to be published in CMS Physics Technical Design Report Vol. 2
• With 1fb-1, CMS can discover (or exclude) all of the low mass benchmark points• Including expected systematic effects
• Low mass SUSY visible almost immediately• Provided systematic effects are under control
• Current CMS focus is now on commissioning and startupscenarios
Backup Slides
14 June, 2006 R. Cavanaugh, Florida, SUSY06 19
14 June, 2006 R. Cavanaugh, Florida, SUSY06 20
The CMS CalorimetersEM calorimeter |η| < 3 :PbW04 crystals1 longitudinal section/preshower 1.1 λ∆η×∆ϕ = 0.0174 × 0.0174Central Hadronic |η| < 1.7 :Brass/scintillator2 + 1 Hadronic Outer – long. sections 5.9 + 3.9 λ (|η| =0) ∆η×∆ϕ = 0.087 × 0.087
Forward calorimeter 2.9 < η < 5:Fe/quartz fibers ∆η×∆ϕ = ~0.175× 0.17
Hcal barrel and EndCap
EM barrel and EndCap
Very Forward Calorimeter
Endcap Hadronic 1.3< |η| < 3 :Brass/scintillator +WLS2/3 longitudinal sections 10λ∆η×∆ϕ = ~0.15 × 0.17
Hadronic Outer
14 June, 2006 R. Cavanaugh, Florida, SUSY06 21
Muon System
1.6
ME4/1 restored
MB1
MB2
MB3
MB4
ME1ME2 ME3
14 June, 2006 R. Cavanaugh, Florida, SUSY06 22
Early Jet Energy Calibration• Require at least one of the two leading
jets to have |η|
14 June, 2006 R. Cavanaugh, Florida, SUSY06 23
Use Track and Muon System (Z→µµ) to Calibrate Calorimeter (MET)
• Variation on a Z→µµCandle theme• Derive MET corrections
from Z→µµ Sample• Apply to SUSY Sample
(to test)
• Some fine tuning required• But basically works
CMS
CMS
SUSYLM1
14 June, 2006 R. Cavanaugh, Florida, SUSY06 24
Inclusive Search Strategies• Use Missing Transverse Energy (MET) as the key signature for
SUSY in analyses presented here• R-parity conservation, neutral LSP
• SUSY benchmark points studied in detail using GEANT-based detector simulation and full reconstruction algorithms
• Consider all backgrounds as well as lepton fakes• QCD multi-jets, W/Z+jets, t-tbar, diboson
• Optimize significance to determine cuts at a particular benchmark point(s)
• Anticipate systematic effects and estimate uncertainties
• Determine 5σ reach in mSUGRA space using fast simulation
14 June, 2006 R. Cavanaugh, Florida, SUSY06 25
Inclusive OS Dilepton + Jet + MET• Cuts (optimize @ LM1):
• 2 OS SF isolated leptons (e,µ)• pT > 10 GeV
• MET > 200 GeV• ≥2 jets:
• ET1>100 GeV• ET2>60 GeV• |η| < 3
• Background (1 fb-1)• 200 events, mostly t-tbar• Systematic uncertainty 20%
• LM1 Signal (1 fb-1)• 850 events
Subtract different favor leptons
mllmax = 80.4 ± 0.5 (stat) ± 1.0 (misalign) GeV
14 June, 2006 R. Cavanaugh, Florida, SUSY06 26
Effects of Misalignment
• Misalign Tracker and MuonSystem separately• Evaluate the impact on the
dilepton end point
• Two scenarios:• First Data, 6 months, 100 pb-1 to 1 fb-1
di-muon efficiency decreased by ~30%di-electron efficiency decreased by ~10%
• Long Term, >6 months, >1fb-1di-muon efficiency decreased by ~13%di-electron efficiency decreased by ~2%
CMS
CMS