top physics via dilepton final stateskehoe/d0/talks/topdileptons.pdf3/31/05 robert kehoe (smu.): top...

Top Physics via Dilepton Final StatesTop Physics via Dilepton Final States

Robert KehoeRobert Kehoe

10/18/0410/18/04Southern Methodist U.Southern Methodist U.

Dept. of Physics SeminarDept. of Physics Seminar

3/31/05 Robert Kehoe (SMU.): Top Quark Dilepton Physics 2

Top in the Electroweak Model...Top in the Electroweak Model...

ÿ top firmly predicted to complete 3rd quark doublet– no anomalies– absence of flavor changing neutral current interactions– weak isospin measurement of b-quark

T3b = -1/2

ÿ coupling to Higgs near unity– why?– fermion masses

• related to couplings to Higgs boson• couplings not predicted• haven’t explained fermion masses

ÿ high top mass– strongly correlated with Higgs mass


Particle ID

- electrons- photons- muons- taus- jet from gluonsand light quarks-b-quark jets- missing Etfrom neutrinos


(eg(eg. pp+/-+/-))

neutrinos

outward from beam-pipe

charge

A Generic HighA Generic HighEnergy DetectorEnergy Detector

ÿ Tracking:– charged particle directions

and momentaÿ Calorimeter:

– main energy depositionsin event

ÿ Muon Detection:– identify penetrating

muons


Jets and TopJets and Top

ÿ jets are a prominent element of many interactions

ÿ however, top special among quarks– jets used as fundamental objects to reconstruct events

• analogous to tracks for lighter quarks

ÿ Once mtop > 130 GeV– b-jets very energetic– signature to permit descrimination from background

• thru jet multiplicity, kinematics, and flavor identification

ÿ So, jets have become a crucial element of top analysis– jet systematics have strong impact on ability to do top physics

• jet energy scale– even more critical when trying to measure top’s properties

• eg. initial and final state radiation complications for kinematic fitting

†

pp


‘hadronization’or

‘fragmentation’

ÿ strength of strong interaction– increases with increasing distance

ÿ when two colored particles interact– nature does not permit ‘naked’ color– energy in strong interaction

• grows as particles move apart

– when energy greater than masses of fundamental particles• a ‘jet’ of particles ‘pulled’ out of the virtual sea or vacuum• extremely messy and poorly understood process

Jet ProductionJet Productionq

q’ 2 `jets’ of particles


FragmentationFragmentation

ÿ Field-Feynman– independent fragmentation– particle multiplicity near

jets independent of rest ofevent

ÿ Lund String Model– color connections in event

dictate where particles go– better reflects experimental

observations

Lund String

Field-Feynman


Jet AlgorithmsJet Algorithms

ÿ ‘fixed-cone’– begin with significant energy cells or towers

• draw cone and calculate Et and position in eta,phi• iterate until stable• merge-split criteria

– standard algorithm at hadron colliders• likely to be important at ATLAS (F. Merritt, S. Rajagopalan)

ÿ Kt or Durham algorithms– cluster without fixed size– group particles which have small Pt w/respect to each other (‘kt’)– does not seem to work well at hadron colliders due to physics ‘noise’

ÿ energy flow– holy grail to some to get optimal energy resolution for dijet resonances– identify individual clusters and group them

• replace caloriemter clusters with track momentum when possible• correct remaining clusters by known non-linearites PER-PARTICLE

– difficult at hadron colliders, worse with many multiple interactions• confusion of overlapping clusters, unreconstructed energy, tracking

ambiguities


Identifying Identifying Neutrinos: Missing Transverse EnergyNeutrinos: Missing Transverse Energy

ÿ Neutrinos do not interact with the detector– we infer presence from imbalance in

transverse momentum

- total energy: unknown- total longitudinal momentum: unknown- total transverse momentum: zero

ÿneutrinos: missing ET

–from calorimeter, plusmuons and jet energy scalecorrections

hard scattering

p

p_

e-

nb

q’q

b

electron

inferredneutrino


Jet Energy ScaleJet Energy Scale

ÿ methods currently used– O: underlying event, noise

• minimum bias events– R: non-linearities, dead material

• direct photon candidate events• statistics up to 200 GeV energy

– S: particle showers• jet transverse shapes in data

ÿ uncertainties– large statistical errors– substantial systematic errors

• increase with energy due toextrapolation

Ecorr = (Euncorr - O)/ R*S


The Top Quark: ProductionThe Top Quark: Production

ÿ in proton-antiproton collisions– top-antitop pair strong production

• mainly quark-antiquark annihilation• Run I measurement (DØ @ 1.8 TeV):

• Run II cross section (@ 1.96 TeV)– expected to be ~35% higher– i.e. ~ 7 pb

pbtt 7.17.5)( ±=s

~85% @ TeV

~15%dominant @ LHC


The Top Quark: Decays and Final StatesThe Top Quark: Decays and Final States

ÿ BR(t Ÿ Wb) ~ 100%ÿ W and b-quark decays

– specify final states

– W decays to• quark-antiquark• OR lepton-neutrino

– b-quark gives a jet• b may decay semileptonically• b decay displaces jet vertex

– isolated high PT leptons (from W’s)– soft leptons in jets (from b’s)– detached vertices in jets (from b’s) ---x---b-tag

xx---SLT

---xxtopo

qqqqlnqqlnln

tb

W-

l,q

n,q’

_

_


Dilepton ChannelsDilepton Channels

ÿ event selection– 2 high PT isolated charged leptons (e, m)– large missing ET (from 2 neutrinos)– > 2 jets (from b-quarks)– a `topological’ analysis:

• i.e. large total transverse energy (HT)

ÿ few physics processes leave energetic lepton pairs– WW, Ztt: estimated from Monte Carlo– Z/g* is most common, but doesn’t produce neutrinos

• fake missing ET: determine from data

ÿ leptons can, rarely, be faked by detector imperfections– due to noise, or similarities of photon and electron showers– fake leptons in W+jets– fake missing ET and leptons from QCD– determined from data

backgrounds

l

b

b

n

p p

E T

ljet

jet

n


Dielectron ChannelDielectron Channel

ÿ both Ws decay to electron-neutrinoÿ primary backgrounds

– Z-> ee + fake missing Et– WW -> ee+2nu– Z->tautau->ee+4nu– W->enu + 1 fake electron– QCD 4jets -> 2 fake ‘e’s + fake missing Et


InstrumentalInstrumentalBackgrounds - Backgrounds - eeee

ÿ fake electrons– b->c+enu negligible

• different than in muon case– pi0 jet passing all ID cuts

• pi0 gives 2 photons– can convert or overlap other tracks

– fake rates ~ 10-5

ÿ fake missing Et– calorimeter noise– multiple interactions– jet resolution

expected tobe true atATLAS aswell


eeee Results Results


Electron-Muon ChannelElectron-Muon Channel

ÿ one W decays to electron, one to muonÿ backgrounds

– Z -> tautau -> emu + 4 nus– WW -> emu + 2 nus– WZ -> emu + 1 nu– Wg + Zg -> 1 or 2 mu + g fakes ‘e’– W+jets -> mu + 1 nu + fake ‘e’– W+jets -> e + 1 nu + fake isolated ‘mu’


eemm - Results - Results


Lifetime b-tagging in Lifetime b-tagging in eemm

ÿ secondary vertex tag (SVT)– look for displaced vertices (> 2 tracks): 10x reduction in background– jet tagged as b-jet if

• signed decay length significance > 5


Tag RatesTag Rates

ÿ b-tag efficiency– tracking efficiency

in jets– position resolution

of tracker• dictates how well

discriminate latedecaying hadrons

ÿ mis-tag rate– tracks in jets

mismeasured– look like don’t

come from eventprimary vertex


beforetagging

final sample


Dielectron: s = 12.7 +8.0- 6.5 (stat) ± 0.8 (sys) ± 0.8 (lumi) pb

Electron-Muon: s = 9.2 +4.8- 3.5 (stat) ± 0.6 (sys) ± 0.6 (lumi) pb

em + b-tag: s = 11.1+5.8- 4.3 (stat) ± 1.4 (sys) ± 0.6 (lumi) pb


The Top Quark MassThe Top Quark Mass

ÿ factor 2 improvement inmeasurement comes at LHC

– e(Tev.) = 1.5 GeV– e(LHC) <~1 GeV

ÿ window on EWSB– enters calculation of SM

parameters as radiativecorrection

– mH ~ 115 GeV/c2?

Run I measurements

r) Ä (1 G

∂á)

M

M(1M

F2Z

2W2

W +¥=-¥

Higgs mechanismRadiative

Corrections


Measurement of mMeasurement of mtt in Dilepton Events in Dilepton Events

ÿ dilepton channels– provide ~ 1/5 the # of candidate events as lepton+jets channels

• higher efficiency, but much lower Branching fraction• analysis statistically dominated, unlike single lepton channels

– jets• always b-quarks• higher Pt than b’s from W’s

– fewer jets, lower background: substantially lower systematicuncertainties than l+jets

~ 1 GeV3.512.8Total

~ 1 GeV2.03.6systematic

< 0.5 GeV2.812.3statistical

LHC (10 fb-1)Run IIa (2 fb-1)Run I

Error (GeV)

comparableto l+jets

better thanl+jets


Dilepton Event CharacterizationDilepton Event Characterization

ÿ final state defined by 4-vectors of 6 decay products– 18 independent quantities– measure 14 directly: lepton

and jet momenta, missingEx and Ey

– 3 constraints:• l-nu pairs give Mw• mt = mtbar

– still have -1C fit!

ÿ sample expected neutrinoeta-distribution


Event WeightEvent Weight

ÿ use constraints to solve for nu momentum

– obtain 4 solutions, only 1 correct– can calculate solved missing Et

ÿ calculate weight from missing Et agreement

ÿ integrate weights for many configurations for each top mass

template for mtop = 175 GeVw/ and w/o parton corrections


Mass SensitivityMass Sensitivity

ÿ for different input top masses– does method give correct result?

– depends on how reconstruct event• 3 jet events confuse things

– which are b’s– can sum up pairs of nearby jets

– peak position µ mtop


Dilepton Mass: Steps to Complete AnalysisDilepton Mass: Steps to Complete Analysis

ÿ create templates from weight distributions– weight distribution characteristics

• peak, width, whole shape, secondary peaks...• what gives best senstivity to mtop?

– current example: 40 GeV bins– obtain for many different top masses– also for background

ÿ extracting a mass from a data sample– maximum likelihood fit to S and B

distributions– ensemble tests of 10 event samples

• average of many experiments: 173 GeV (175input)

• large statistical errors

very preliminary!


Prospects at LHCProspects at LHC

ÿ cross section ~100x at Tevatron– 10 fb-1 gives 100s to 1000s more tops

• i.e. O(10k) dilepton events identifiedand O(100k) single lepton events

ÿ top cross section and branching ratio– constrain non-SM decays

• eg. t->H+b would produce anomalous ratioof ee to mumu (to tautau) events

• t-> Zq...

ÿ top spin correlations– can be indication of CP-violation in top

sector– leptons from top provide best indication

of top spin direction– dileptons excellent place to look

ÿ mass measurement– might use Mlb

• also has information about W helicity

– two leading jets methods– event recosntruction as earlier

†

cosq* =2• Mlb

2

M t2 - MW

2 -1

top physics via dilepton final stateskehoe/d0/talks/topdileptons.pdf3/31/05 robert kehoe (smu.): top...

Documents