LHC SUSY discovery potential

1
LHC SUSY discovery potential
V.Zhukov University of Karlsruhe and SINP
MSU (for CMS and ATLAS collaborations)
CMS
ATLAS
SC Magnet coil
ECAL PbW04
HCAL scint./brass
Tracker Si strips Pixels
Muon barrel DT endcups CSC
2
Test model mSUGRA
Five universal (m, A at the GUT scale)
parameters describe masses and couplings
mo , m1/2 , A0 , tan? , sgn?

• New particles
• squarks (uL,R , dL,R , b1,2, t1,2) ,
  sleptons (l, v L,R , ?1,2 )
• gluino (g)
• gauginos ( ??1,2,3,4 , ??1,2 ) , ??1
  is a Dark Matter candidate
• SUSY Higgses (ho, A, H???

(?gtMz)
3
MSUGRA constraints

• Theoretical
• ElectroWeakSymmetryBreaking(EWSB)
• excludes small m1/2 and large m0 region
• depending on mtop (Tevatron)

• Experimental
• Higgs mass (LEP) mhoSM gt114.4 GeV
• Chargino mass (LEP) m???gt103 GeV
• Br(bs?) (CLOE,BELL,BaBar) (3.42????????????
• ????????????????????????????Muon
  g-2????????????????????????? excludes low mass
  region

• Dark Matter (if DM is supersymmetric)
• stau is LSP (exclude low m0)
• Relic Density(WMAP) ??h2 0.113?0.009??????
• RD define a narrow band in m0-m1/2 plane at
  fixed tan????
• (at large m0, tan?? depends on EWSB parameters)
  
• Indirect DM (EGRET) m1/2lt250 , tan?gt50

tan?50 A00
At tan? ? 20 only focus point and coannihilation
regions are complaint with the RD
4
Search strategy
MeffMETSPTj
ATLAS
1. Inclusive search Check any deviations from
SM predictions already at low statistics (small
?? low Lint ) Try to define SUSY mass scale
Meff min(g,q) Counting like experiments.
Uncertainties are very important. Difficult to
constrain model parameters.
METJets
Meff
2. Exclusive search With enough statistics
reconstruct kinematic endpoints in invariant
mass. Constrains mass spectrum and SUSY
parameters. Works in limited regions , at low
m0. Requires large statistics.
ATLAS
Contours where dilepton edge is working
5
SUSY cross sections
Contributions of different production channels
Main production channels
? tot pb
tan?50 A00
Total mSUGRA cross section(LO)
LO ISAJET PYTHIA NLO KNLO1.3-1.8
(m0100-2000) (PROSPINO)
6
MSUGRA Signatures
Different regions in mSUGRA mass plane
(high)
2 body decays
Main final state in the cascade
decays METJetsLeptons
3 body decays
Leptons OppositeSignSameFlavor(OSSF) from
c20 SameSignSameFlavor(SSSF) from different c?
test points
7
Signal topologies
MSUGRA averaged observables in m0-m1/2 plane
CMS fast simulations
tan?50 A00
8
Backgrounds
Most important SM backgrounds to the inclusive
SUSY search
MET, Nj, Nm are from CMS fast simulations
LO PYTHIA CTEQ5L (for ttbar, wt -gtTopRex)
W,Z leptonic decays in bosons production
9
Background signatures
Normalized SM bkg observables
CMS fast simulations
Njets ETgt30GeV
MET
HeffMET?ETj?PTl
Nmuons
ET highest Jet
PT highest muon
10
Uncertainties theory
Kinematic variable (jets ET, MET) Parton Shower
(PYTHIA ) or Matrix Ele ment (Alpgen) jets
give different jets ET spectrum and affect both
Njets and MET measurements (fake MET) Can
increases bkg by 3-5 times (ATLAS) .
ATLAS
PYTHIA PS 6.2 ALPGEN ME
Cross sections LO,NLO,NNLO factors 10
uncertainties . PDF LHPDF uncertainties
10 depending on channel.
PT of highest emitted jet in ttbar production
ISR/FSR QCD and QED radiations can increase jets
and leptons multiplicity. QCD ?s related
uncertainties contribute 5
PDF reweighting
PileUp, Underlying events, beam remnants tend
to increase MET. Require careful tunning of MC
generators to experiment
Most of this model dependent uncertainties will
be tuned to experiment in first few months of LHC
LHPDF uncertainties for different bkgs in the
SUSY trilepton search.
11
Uncertainties detector
Jets ET resolution

• Jets
• ET Resolution stochastic ?ET 1.25/?ET GeV
  
• Angular resolution
  ltcalo cell size (?? x ?? 0.087 x
  0.087)
• Corrections
  parton level 10
  MC
  particle level simulations Erec/EMC up to 30
• Fake Jets ( noise, PileUp, UE )
  tracker/calo matching???????ptracksT
  /Etj gt 0.2

?/E 5.6/ET? 1.25 /? ?ET ? 0.033

• Jets Energy Scale
• data-driven calibrations
• Balancing technique dijets, gjets, Zjets
• pT dijet(pTprobepT hlt1 )/2 , calibrate for
  each h bin
• W mass reconstruction in ttbar events
• Ecor(1C) Emeas , C12 dC 3 (due to PU
  uncert.)
• Energy calibration improves with statistics
• JeS 3 at 1 fb-1 PTgt50 GeV/c

1Hz
0.1Hz
12
Uncertainties detector
MET resolution
?MET2 (3.8)2 (0.97 ? ?ET )2 (0.012?ET)2
13
Inclusive searches
CMS Physics TDR set of realistic analysis using
full CMS simulation Most of analysis are tuned
to the bulk region (LM1 m060 m1/2250 tanb10)
) some other constraints on angular variables
?(j1j2),??(j1MET), h, etc. optimized to get
maximum significance
Main selection cuts MET gt130-250 GeV, bkg.
rejection gt103-105 Jets ETgt 100 GeV Leptons
PTgt10 GeV/c Kinematical fit on masses (top, Z, ho
)
Remaining bkg ttbar, Wjets, QCD, Zjets
14
CMS 5? discovery reaches
'First days' (2008) discovery reach m(q,g)lt
1000 GeV
1 year of Low Luminosity L1032
cm-2s-1 m(q,g)lt 2000 GeV
Full simulations
ATLAS is very similar
Systematic uncertainties can contribute up to 50
to significance. but will be reduced with the
first data.
15
MSUGRA topology selection
Can we identify the mSUGRA topology? L.Kane et
al Need a selector sensitive to a particular
signal topology.
Train Neural Network(NN) to a selected region
against others, using different observables
MET, Njets , Nlep , ETj , PTlep ,cosFj1j2 , ...
Example
Selection Efficiency in m0-m1/2 plane of two NNs
trained for (m0, m1/2, tanb)
( 1100, 300, 50)
( 500, 450, 50)
m1/2
preliminary
m0
Can also estimate relative contributions from
different production channels
Use NNs trained for the decays of g,q or
?? Simulated (R MC) and reconstructed (R rec)
ratios of ??(gg)?(qq)?(?,?) , R(gg) is
set to 1
16
Exclusive search
Always two decay chains with two LSPs at the end.
Reconstruct kinematic end points in invariant
mass distribution of final jets and leptons in
the long (longer is better) cascade decays.
2 body decays in low mass region mggtmq at large
mass (m0) mostly 3 body decays
Kinematic end points (MC) in Dalitz plot of Mll
and Mllq
Different combinations of invariant masses l1l2
, l1q1, l2q1 , l1l2q1 , llqq.. but combinatorial
bkg. from different chains.
How to separate hemispheres? - use different
decay channels (reduces statistics) -
subtraction of combinatorial bkg (mixed event
technique) - group particles in each event
around 2 axes using f.ex minimum of scalar
product pa. 80 efficiency for jets.
17
Kinematic end points analysis
m0100 m1/2300 tanb2
ATLAS TDR
Example
ATLFAST
Kinematic end point in Minv are related to the
masses mq, mq, ml , m??????m????... Can
solve equations assuming cascade shape and
reconstruct all masses.
Mll OSSF leptons
ll edge
30 fb-1
Reconstructed gluino mass in g-gtbb-gtc20bb
cascades
OSSF leptons jet Mllqmin(mllq1, mllq2)
llq edge
m0100 m1/2250 tanb10
100 fb-1
CMS
Ml1q or Ml2q depending on Mll cut
CMSJET
10 fb-1
lq edge
100 fb-1
more Minv distributions, see Allanach et al
18
SUSY parameters

• Reconstruct mass spectrum from the end points

• Use cross sections information

Bulk region SPS1a (m0100 m1/2250 tan?10)
Solve ambiguities by constraining kinematic edges
with cross sections.
300 fb-1
C.Lester et al.
ATLFAST
Second solution -gtambiguities

• Fit mSUGRA parameters

19
Beyond mSUGRA
Non universal SUSY(NUM) non universal scalars
or gaugino . many possibilities but final
signatures are the same..
Split SUSY squarks are heavy, gluino is light
and long lived ( Rhadrons). Can be measured in
muon system as a sign flipping heavy ionizing
slow object.
GMSB gravitino G (lt1KeV) is LSP . Charged NLSP
(?,l) can be long lived, producing signal in
muon system. Search for hard gamma from
??0-gt??G
NMSSM new particles, extra Higgses and
neutralino ?05 can be observed
UED all masses are almost degenerate but final
state are similar to SUSY. Small difference in
?ll in the direct sleptons production
And more .....
Can we separate all these possibilities? Need
very precise (lt10) mass measurements! Only in
limited parameter regions can separate different
models.
20
Summary

• After understanding the SM at LHC energy, can
  start searching for SUSY.

• Main signal signature MET high PT jets
  leptons can be distinguished from the SM channels
  (ttbar, wjets, zjets, QCD)

• LHC is expected to see mSUGRA sparticles
  production up to 2 TeV range at Lint gt10
  fb-1 in inclusive searches.
• The lower mass region lt 1000 GeV can be seen
  at Lint lt1 fb-1

• The identification of mSUGRA topologies will
  be possible already in inclusive searches and the
  model parameters can be precisely reconstructed
• in exclusive channels using kinematic end points
  at Lintgt30 fb-1
• for low mass region.

• The detector reconstruction uncertainties of
  the missing transverse energy and jets energy
  together with the theoretical uncertainties of
  the SM backgrounds are the main focus for future
  studies.