ATLAS sensitivity to W and top polarization in tt events

1
ATLAS sensitivity to W and top polarization in tt
events
ATLAS Physics Workshop, Rome, June 06-12, 2005
Fabrice Hubaut
CPPM/IN2P3/CNRS Univ. de la Méditerranée Marseil
le, FRANCE

1. Introduction, Motivations
2. W polarization in top decay
3. Spin correlation in top pairs
4. Conclusions and perspectives

Prague V. Simak, K. Smolek - Marseille F.
Hubaut, E. Monnier, P. Pralavorio, B. Resende
2
Motivations test top production and decay
Top production
Top decay
Weak Interaction single top
tWb vertex

• BR(t?Wb)100
• No top hadronization

LHC s 300 pb
Tevatron s 3 pb
LHC s 850 pb
Tevatron s 7 pb
Test SM, Search for anomalous couplings
Top polarization
W polarization
tt final states (10 fb-1)

• Full hadronic (3.7M) jets
• Dileptonic (0.4M) 2 l 2 b 2 n
• Semileptonic (2.5M) l n 2b jets

• Prague
• Marseille

3
tt event simulation
Event simulation is performed using

• TopReX LO density matrix for production and
  decay of tt including spin effects
• Pythia 6.2 hadronisation, fragmentation, decay
• Tauola Photos t lepton decay and radiative
  corrections
• Default Mtop175 GeV, CTEQ5L structure
  function, ISR-FSR

Fast Sim
Full Sim

• Method background systematics
• Atlfast 2.6.0
• Jet cone size 0.4 (0.7 for full sim compar.)
• 10 fb-1 of statistics (semilep dilep)

• Method fine detector effects
• Rome CBNT 004520 - 004521
• Jet cone size 0.7
• 0.8 fb-1 of statistics (semilep only)

4
tt event selection and reconstruction

• Event selection

Semileptonic

• e(sig)3.3, 85 000 events per 10 fb-1
• S/B12 main background tt ? ? X

Standard cuts

• e(sig)5.3, 21 000 events per 10 fb-1
• S/B6 main background tt ? ? l

Dileptonic

• Rome

• Event reconstruction

174.3 GeV s9.7 GeV
176.60.8 GeV s11.61.2 GeV

• Semileptonic quite easy (1 n)
• Dileptonic 2 n
• Solve 6 non-linear equations / 6 unknowns
• pT of tops and n to compute solutions proba.
• ? 80 efficiency, 65 of right solutions

Hadronic Top Mass (GeV)
5
W polarization in top decay

• Top weak decay ? V-A coupling as for all
  fermions

Standard Model (Mtop175 GeV) 0.703 0.297 0.000
Longitudinal W (F0)
Left-handed W (FL)
Right-handed W (FR)
NLO
0.695
0.304
0.001
Sensitive to EWSB
Test of V-A structure

• W polarization is measured through angular
  distribution of charged lepton

1/N dN/dcos?
angle between lepton in W rest frame and W
direction in top rest frame
?
spin1/2
1
1/2
cos?
6
Measurement method

• Selection cuts and reconstruction distort the
  parton level distribution
• Use an independent sample to parametrize these
  effects

parton level
Correction function
1/N dN/dcos?
1/weight
Reconstruction and cuts
cos?
Fit with 3rd order polynom in range -0.90.9
1/N dN/dcos?
cos?
apply weights event by event
cos?
Use this unique weight parametrization on every
sample
7
Comparison full/fast simulation (1)

• Compare correction function Full Sim. / Atlfast
  with semileptonic events

Muons only
Electrons muons
1/weight
1/weight
cos?
cos?

• Main differences Full Sim. / Atlfast come from
  electron reconstruction

8
Comparison full/fast simulation (2)

• Results after selection, reconstruction and
  correction

Fit with 2 parameters (constraint F0FLFR1)
F00.699 0.005 FL0.299 0.003 FR0.002
0.003
F00.683 0.028 FL0.308 0.014 FR0.009
0.014
semileptonic
semileptonic
1/N dN/dcos?
1/N dN/dcos?
(Atlfast) 10 fb-1
(Rome) 0.8 fb-1
Fast Sim
Full Sim
cos?
cos?

• Analysis limited by Rome statistics

9
Systematics b-jet calib. and top mass

• dileptonic

• semileptonic

• b-jet miscalibration
• expected behaviour
• positive miscalibration ? Mlb? ? cos ? ? ?
  FR? FL ?

• Top mass uncertainty
• expected behaviour

10
Systematics summary
(semileptonic channel)
Q-scale
PDFs
ISR/FSR
b-frag.
Hadronization
Dominant contributions
b-tag. (5)
b-miscalib. (3)
Top mass (2 GeV)
S/B (10)
Pile-up (2.3)
11
W polarization results
Combined results of semileptonic and dileptonic
channels for SB at 10 fb-1 (stat syst)
Results (stat syst) Standard Model
F0 0.698 ? 0.004 ? 0.016 0.703 0.002?(Mtop-175)
FL 0.301 ? 0.003 ? 0.023 0.297 - 0.002?(Mtop-175)
FR 0.001 ? 0.003 ? 0.012 0.000

• In 1 LHC year (10 fb-1), ATLAS can measure F0
  with an accuracy 2 and FR with a precision
  1.2
• Measurements largely dominated by systematic
  uncertainties
• Tevatron expectations with 2 fb-1 dF0stat0.09
  and dFRstat0.03

12
Sensitivity to new physics
From W polarization, deduce sensitivity to tWb
anomalous couplings ? model independent approach,
i.e. effective Lagrangian
)
and 4 couplings (in SM LO
F0

• 2s limit (stat?syst) on 0.04
• 3 times better than indirect limits
  (B-factories, LEP)
• Less sensitive to and already
  severely constrained by B-factories

1s
Anomalous coupling
13
Top polarization
Production

• Cross Section
• Top Spin Polarization
• Anomalous couplings
• Resonance Production

b
p
l
W
n
t
Decay

• Branching Ratios
• Anomalous couplings
• Rare / non-SM Decays

b
t
q
W-
q
p
Top quantum numbers

• W polarization

• Mass, Spin, Charge

14
tt spin correlation
(in helicity basis)

• tt pairs not polarised (lt 1 at NLO) but
  correlations between spins of t and t

LHC
s (a.u.)
Tevatron
(projections on lab. axes)
AD-0.24
Mass of tt system, Mtt (GeV)

• As for W polarization, measure angular
  distributions of daughter particles
• Same events selection and reconstruction as for
  W polarization
• Add cut Mttlt 550 GeV to increase asymmetry ?
  A0.42, AD-0.29

15
Measurement method

• Selection cuts and reconstruction distort the
  parton level distribution

Correction function for A
Correction function for AD
semileptonic
semileptonic
1/weight
1/weight
9cos?1cos ?2
3cosF

• Results after selection, reconstruction and
  correction

(stat) (stat) Standard Model
A 0.40 ? 0.02 0.50 ? 0.10 0.42
AD -0.28 ? 0.01 -0.29 ? 0.06 -0.29
Fast Sim
Full Sim

• Good agreement Full Sim. / Atlfast within Rome
  statistics

16
Spin correlation results
Combined results of semileptonic and dileptonic
channels for SB at 10 fb-1 (stat syst)
Results (stat syst) Precision Standard Model
A 0.40 ? 0.02 ? 0.04 11 0.42
AD -0.281 ? 0.009 ? 0.012 5 -0.290
(Mtt cut included)

• In 1 LHC year (10 fb-1), ATLAS can measure spin
  correlation 5
• Measurements dominated by systematic
  uncertainties
• AD observable can be measured more precisely
  than A
• Tevatron expectations with 2 fb-1 dAstat/A40

17
Conclusions-Perspectives

• With 10 fb-1, ATLAS can measure precisely the W
  polarization 1-2 and the top spin correlation
  5 combining semilepdilep tt events
• Dominated by systematics, complete study
  performed with Fast sim.
• First Full sim. study confirms robustness of
  measurement method
• This allows to test top production and decay
  with unprecedent precision (3 to 5 times better
  than Tevatron statistical expectations with 2
  fb-1 )
• Sensitivity to new physics search for anomalous
  couplings in a model independent approach
• 3 ATLAS notes written ATL-PHYS-PUB-2005-001,
  ATL-COM-PHYS-2005-015 (semilep) and
  ATL-PHYS-2003-012 (dilep)
• Scientific note first draft ready!

18
Scientific note

19
SPARE SLIDES
20
Event selection (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)
• ? double b-tag
• Event topology reconstruction
• Quality cuts
• lMwREC - Mwl lt 20GeV and lMtREC - Mtl lt 35 GeV

le,µ
LEPT.
HAD.

• e(sig)3.3, 85 000 events per 10 fb-1
• S/B12 main background tt ? ? X

21
Event reconstruction
(semileptonic channel)

• Jet calibration, b-tagging
• Select the 2 non-b jets with Mjj closest to MW
• Select the b jet with Mjjb closest to Mt
• PTmiss for PT? and Pz? by constraining Ml? to Mw
• Select the b jet closest to the lepton
• Select the ? with Ml?b closest to Mt

22
Background (semileptonic channel)
1 LHC year of statistics (10 fb-1) simulated for
each background, except
Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1)
Expected events (x106) Events after selection recons.
W(?l?) 4jets (AlpGen) 20 (pTjetsgt10 GeV) 400,1000
QCD (bb) 6000 (vsgt120 GeV) 200
Z(?ll-) jets (Pythia) 50 12
ZZZWWW 1 4
W(?l?) bb (AcerMC) 0.7 3
Single top 1.0 350
tt ? ?X 1.3 6200
tt ? all had (TopReX) 3.7 70
SIGNAL 2.5 85000
Poisson stat. rescaled by 63 Stat. rescaled
by 8
Non tt 50 lt S/B lt 90
tt ? ?X S/B13
S/B12 main background tt ? ? X
23
Event selection and reconstruction (dileptonic
channel)

• pT and h cuts
• 2 opposite sign leptons pT gt 20GeV (?lt2.5)
• PTmiss gt 40GeV
• 2 b-jets pT gt 20GeV (?lt2.5)
  ? double b-tag
• Event topology reconstruction
• Solve 6 non-linear equations/6 unknowns (n)
• 80 efficiency, 65 of right solutions

• e(sig)5.3, 21 000 events per 10 fb-1
• S/B6 main background tt ? ? l

Event reconstruction for selection of the
correct solution of the kinematical equations,
the pT of tops and n was used for computation of
the solutions probability
24
Background (dileptonic channel)
1 LHC year of statistics (10 fb-1) simulated for
each background
Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1) Signal and background for 1 LHC year (10 fb-1)
Expected events (x106) Events after selection recons.
QCD (bb) 30000 (PTgt20 GeV) lt200
Zjets, Wjets (Pythia) 4500 lt100
ZZZWWW 4500 lt100
W(?l?) bb 4500 lt100
Single top 1.0 7
tt ? ?l 0.5 3700
tt ? ljet (TopReX) 3.8 40
SIGNAL 0.4 21000
S/B6 main background tt ? ? l
25
Lepton reconstruction
Electron PT
Electron ?
Matter effects, brem.
Muon PT
Muon ?
26
Selection efficiency ()
Rome (0.19M) Atlfast (2.5M)
1 electron (pTgt20 GeV) 1 muon (pTgt20 GeV) 54.5 0.2 63.2
1 electron (pTgt20 GeV) 1 muon (pTgt20 GeV) 62.0 0.2 57.6
gt4 jets (pTgt30 GeV) 39.4 0.1 41.0
Ptmiss gt20 GeV 91.1 0.1 90.2
All kinematic cuts 6.5 0.1 6.5
Dmass cuts D(MWhad) lt 20 GeV D(Mtop) lt 35 GeV 2.3 0.1 3.3
el90
ej95
27
Actual meas./limits of W polarization
Top Standard Model weak decay ? V-A coupling as
for all fermions

Standard Model (Mtop175 GeV) 0.703 0.297 0.000
Measurement or actual limit F00.89 ? 0.30 ? 0.17 (stat) (syst) (Tev. run II, 162 pb-1) FR lt 0.18 _at_ 95 CL (Tev. run I , 109 pb-1) FR lt 0.01 from b s ? (CLEO, BELLE, BABAR) ? indirect limit, SM dependent FR lt 0.18 _at_ 95 CL (Tev. run I , 109 pb-1) FR lt 0.01 from b s ? (CLEO, BELLE, BABAR) ? indirect limit, SM dependent
Longitudinal W Fraction F0
Left-handed W Fraction FL
Right-handed W Fraction FR
NLO
0.695
0.304
0.001
28
Sensitivity to tWb anomalous couplings (1)
Assume a variation of each coupling
independently
1s
1s
1s
Coupling
2s limit (stat?syst) 0.30 0.13 0.04
Uncertainties on FR and F0 with 10 fb-1 ?
Best sensitivity to (linear behavior)
29
Sensitivity to tWb anomalous couplings (2)
Comparison with other expectations (2s limit)

Our study (low lumi,tt) 0.30 0.13 0.04
Tev (Run II, tt) 0.5 0.3 0.3
LHC (High Lumi, single top) 0.06 0.07 0.13
Indirect (B-factories) 0.002 0.005 0.4
Indirect (LEP) - - 0.1
100 fb-1, assuming a 5 systematic uncertainty
Sensitivity to largely better than other
ones (even indirect!)
30
Measurement of aW, AFB
Information which can be derived from the W
polarization
1. Spin analyzing power of the W (aW) in the
polarized top decay
with qW angle between W and top spin polarization
(12 accuracy)
2. Forward Backward Assymetry (AFB), related to
the angle between the charged lepton and the
b-jet in W rest-frame
(7 accuracy)
31
Beyond SM - VA component
The correction used is extracted from a V-A
hypothesis (70 FO - 30 FL)
Correction changes in case of VA component ?
bias expected
Iterative process
32
Top spin

• Top decays before hadronisation (t3x10-25 s)
  due to high mass bare quark
• Top production and decay perturbative QCD and
  NLO computation
• No spin flip between production and decay
• direct transmition to decay products
• In top rest frame, polarisation effects (S)
  observed by measuring
  angular distributions of daughter particles

• 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 analyzing power)

W b l,d,s v,u,c lej
? (NLO) 0.40 -0.40 1. -0.31 0.47
lej least energetic jet in top rest frame
33
tt spin correlation observables
Tevatron

• An optimal basis (i.e. A(qq)1 at LO) exists
  off-diagonal basis A0.8 NLO
• First measurement by D0 (run I, 125 pb-1)
  Agt-0.25 at 68 C.L.

LHC

• No optimal basis, asymmetry in helicity basis
  A0.33
• Smaller QCD corrections (a few ), theoretical
  uncertainties under control

angle bwn spin analyzers direction in the t(t)
rest frame
34
tt spin correlation measurement

• Event selection and reconstruction exactly the
  same as for W polarization
• except, add cut on tt invariant mass to enhance
  spin correlation Mttlt 550 GeV
• Measure A and AD with simple unbiased estimators
• Selection cuts distort the parton level
  distribution
• Unique correction function, apply weights event
  by event
• Redo the complete study of systematic
  uncertainties

(e 75)
35
Systematics jet calibration and top mass

• semileptonic

• dileptonic

36
Systematics summary
Q-scale
PDFs
ISR/FSR
b-frag.
Hadronization
Dominant contributions
b-tag. (5)
b-miscalib. (3)
light-miscalib. (1)
Top mass (2 GeV)
S/B (10)
Pile-up (2.3)