tt spin correlation

1
tt spin correlation
P. Pralavorio (CPPM)

1. Motivations and theoretical framework
2. Event simulation
3. tt decay channels
4. Extraction of spin correlation parameters
5. Systematics Results
6. Conclusions and perspectives

Prague V. Simak, K. Smolek CPPM F. Hubaut, E.
Monnier, P. Pralavorio, B. Resende
2
1. Motivations

• Top decays before hadronisation (t3x10-25 s)
• Top production and decay perturbative QCD and
  NLO computation
• No spin flip between production and decay
• and direct transmition to decay products

• Study top spin observables
• Measurement of SM parameters (only upper limit at
  Tevatron)
• Look for non standard couplings in the top quark
  decay (V-A, )
• Sensitive to new physics new production
  mechanisms, discret symmetry tests (CP), extra
  dimension, ...

Test Standard Model and possible deviations
3
1. Theoretical framework (1)
In the top rest frame, polarisation effects (S)
are observed by measuring angular distributions
of daughter particles
s
qi angle between decay particle of the top and
top spin quantization axis s ai degree to which
its direction is correlated with the top spin
(spin analysing power)
W b l,d,s v,u,c lej
? (NLO) 0.40 -0.40 1. -0.32 0.47
4
1. Theoretical framework (2)

• Studies at LHC
• Use naturally polarized single top (300 pb)
• tt pairs (800 pb) not polarised (lt 1) but

Correlations between spins of t et t
s (u.a.)
Apply Mtt lt 550 GeV, to increase the asymmetry
and reduce systematics coming from high pT tails
Mass of tt system, Mtt (GeV)
5
1. Theoretical framework (3)

• tt spin correlation Angular distribution
  of decay particles (predicted by SM)
  (i.e. spin analyzer) on each side (t, t)

• Maximum spin correlation effect at LHC using
  helicity basis
• T1 (T2) angle between t(t) direction in the tt
  rest frame and spin analysers direction in
  the t(t) rest frame.
• F angle between spin analysers direction
  in the t(t) rest frame

6
1. Spin correlation parameters
Two spin correlation observables C and D
(hep-ph/0403035)
C and D unbiased estimators
Standard Model
C0
(No spin correlation)
D0
(No spin correlation)
7
2. Event simulation
Event simulation is performed using

• TopReX 4.05 or AcerMC 2.2 or AlpGen 1.33 LO
  spin density matrix for the production and decay
  of tt pairs
• Pythia 6.2 or Herwig 6.5 hadronisation,
  fragmentation and decay
• Tauola Photos t lepton decay and radiative
  corrections
• Atlfast 2.60 fast ATLAS simulation
  reconstruction
• using CTEQ5L structure function, ISR-FSR, no
  pile-up.

Remark most of the generators (Pythia, Herwig,
MC_at_NLO) do not have the spin correlation
8
3. tt decay channels

• tt pairs possible decays (LHC 10 fb-1)
• Dileptonic (0.4M) ee, em, mm Prague
• spin analysing power
• - statistics, reconstruction (2 v)
• Semileptonic (2.5M) ejets, m jets CPPM
• statistics, reconstruction
• - analysing power
• Full hadronic (3.7M) jets
• statistics
• - trigger, background, analysing power
• Intrinsic background (1.7M) t X

9
3. Dileptonic channel

• pT and h cuts
• 2 opposite sign leptons pT gt 20 GeV (?lt2.5)
• PTmiss gt 40GeV
• 2 b-jets pT gt 20GeV (?lt2.5)
• Event topology reconstruction
• Solve 6 non-linear equations/6 unknows (n).
• 87 of events reconstructed.
• Other cut
• Cut on Mttlt 550 GeV

Cuts tuned to reduce their sensibility on C and
D e(sig)4, main background tt? t X (S/B6)
10
3. Semileptonic channel

• pT and h cuts
• 1 lepton pT gt 20GeV (?lt2.5) PTmiss gt 20GeV
• 1 b-jet pT gt 30GeV (?lt2.5)
• 2 non b-jets pT gt 30GeV (?lt2.5)
• 1 b-jet pT gt 30GeV (?lt2.5)
• Event topology reconstruction
• Other cuts
• lMwREC - Mwl lt 20GeV and lMtREC - Mtl lt 35 GeV
• Cut on Mttlt 550 GeV

LEPT.
HAD.
Cuts tuned to reduce their sensibility on C and
D e(sig)3, main background tt? t X (S/B15)
11
4. C and D extraction (1)
Use a Non Correlated MC generator (PYTHIA) to
parametrize reconstruction cuts effects
C0
Reconstruction and cuts
apply weights event by event
C0
12
4. C and D extraction (2)
Use this unique weight parametrization
everywhere in the following to unfold
reconstruction cuts effects ? example on
Correlated MC
C0.22
applying weights event by event
Reconstruction and cuts
C0.19
13
4. Monte Carlo Generators

• Generate full statistics in semileptonic channel
  with
• different Monte-Carlo generators including spin
  correlation
• Pythia for hadronisation, fragmentation and
  decays (Tauola, Photos)

Generator Q2 gg / qq C (0.01) D (0.006)
TopReX 4.05 mt2pT(t)2 86/14 0.19 -0.142
AcerMC 2.2 s 85/15 0.20 -0.135
AlpGen 1.33 mt2 86/14 0.18 -0.132
tt0jet
Good agreement between generators
14
4. Hadronization Pythia vs Herwig

• Generate full statistics in semileptonic channel
  with
• ACERMC 2.2 to generate partons
• Pythia or Herwig for hadronization

Hadronization scheme C (0.01) D (0.006)
Pythia 6.2 0.20 -0.135
Herwig 6.5 0.19 -0.137
small impact of hadronization scheme
15
5. Systematics
MC level
Analysis level

1. Spin correlation simulation, Q2
2. tt production
3. Structure function
4. Input top mass
5. ISR, FSR
6. b-fragmentation Peterson
7. Fragmentation, hadronization, decay

1. b-tagging efficiency
2. Light jet mis-calibration
3. b-jet mis-calibration
4. Event selection

Main contributions dileptonic semileptonic
16
5. Results
C and D results for SB at 10 fb-1 (stat syst)
Channel Dileptonic Semileptonic
C (Theory) 0.46 ?0.04 0.22 ?0.02
C (exp.) 0.22 ?0.02 ?0.09 0.19 ?0.01 ?0.04
Significance 2s 5s
D (Theory) -0.31 ?0.03 -0.15 ?0.01
D (exp.) -0.23 ?0.01 ?0.05 -0.14 ?0.006 ?0.02
Significance 4s 7s
17
6. Conclusions and perspectives

• Solid theoritical framework to study spin
  correlation at LHC in semileptonic and dileptonic
  channel. Wait for spin correlation in MC_at_NLO
• Measurements dominated by systematics.
• LHC will produce a high top statistics during
  the first year ? gt 5 s sensitivity on SM tt spin
  correlation after 1 LHC year (10 fb-1) is
  feasible.

• Perspectives
• Scientific note under preparation
• Analysis with the full simulation started
• Generalisation to top polarization studies
  (single top, ...)

18
Spare Mttbar cut

• Mtt cut allows to
• increase the C and D values
• get rid of uncertainties at high pT
• decrease significantly the reducible background

-C
19
Spare Systematics at MC level

1. Uncertainties on spin correlation simulation, Q2
   (TopReX)? AcerMC, AlpGen
2. tt production (85 gg vs 15 qq) ? 82/18, 84/16,
   86/14, 88/12, 90/10
3. Structure function (CTEQ5L) ? MRST98, GRV98
4. Input top mass (175 GeV) ? 170, 172.5, 177.5, 180
   GeV
5. (ISR on, FSR off) (ISR off, FSR on) ? 20 of
   the effect
6. b-fragmentation Peterson (-0.006) ? -0.0035 (-1
   s), -0.0085 (1 s)
7. Fragmentation, hadronization, decay (Pythia) ?
   Herwig

20
Spare Systematics at analysis level

1. b-tagging efficiency (60) ? 55, 65
2. Light jet mis-calibration (0) ? -2, -1, 1, 2
3. B-jet mis-calibration (0) ? -5, -3, -1, 1,
   3, 5
4. Event selection (see June 2004)

21
Spare Semileptonic systematics
Theo.Q2 gg/qq Struc. Funct. Mtop (2.5 GeV) FSR (20) b-frag (1s) Hadr. frag. b-jet miscalib (3) TOT
C(lej) 0.015 ltltstat 0.017 0.011 0.014 ltltstat 0.025 0.040
C(whad) 0.024 0.020 0.023 0.013 0.021 0.029 0.012 0.055
D(lej) 0.010 ltltstat ltltstat 0.010 0.008 ltlt stat ltltstat 0.020
D(whad) ltltstat 0.010 0.009 0.014 0.012 0.008 ltltstat 0.025

• Systematics not dominated by one contribution
• Systematics greater for whad than lej

22
Spare Dileptonic systematics
Effective observable MRST GRV b-frag ISR off FSR off Mtop btag e0.55 btag e0.65
C ltltstat 0.030 0.034 0.011 0.002 0.006 0.032 0.077
D 0.015 0.018 0.030 0.008 0.014 0.021 0.032 0.022
E 0.024 0.005 0.006 0.004 0.003 0.010 0.012 0.010
F 0.003 0.014 0.005 0.005 0.006 0.007 0.003 0.011
23
Spare pT comparison between MC
Other jet
lej
Kinematical cuts (pT and h cuts) applied
top
lepton

• TopRexAlpGen
• AcerMC harder pT (top) spectrum compared to
  TopReX, AlpGen

24
Spare Lab angles
E lt cos F gt
(At parton level SM 0.14, NC 0.10)
- doesnt need event topology reconstruction
F lt cos FTgt
(At parton level SM -0.14, NC -0.22)

• doesnt need event topology reconstruction
• (almost) the same in the Lab. and CM system