Top quark mass

Top quark massTop quark mass

For DØ collaborationRegina Demina

University of RochesterWine and Cheese seminar at FNAL, 07/22/05

07/22/05 Regina Demina, Joint Theoretical and Experimental Seminar at FNAL 2

OutlineOutline• Introduction• Top quark mass measurement in Run II

– Matrix element method description– In situ jet energy scale calibration on hadronic W-mass– Sample composition – Result – Systematics

• Tevatron combined top mass• Top quark production

– Update on cross section in l+jets channel– Search for resonance production


Top Quark Mass: Motivation

• Fundamental parameter of the Standard Model.

• Important ingredient for EW precision analyses at the quantum level:

which were initially used to indirectly determine mt.

After the top quark discovery, use precision measurements of MW and mt to constrain MH.

W Wt

b

W W

H

MW mt2 MW ln(MH)

CDF&D0RUNII


Top production Top production

At √s=1.96 TeV top is produced in pairs via quark-antiquark annihilation 85% of the time, gluon fusion accounts for 15% of ttbar production


Top Lifetime and DecayTop Lifetime and Decay

• Since the top lifetime top ~ 1/ M3

top~10 -24 secqcd ~ -1 ~10 -23 sec

• BR(tWb) – Both W’s decay via W l

• final state: llbb - DILEPTON

– One W decays via Wl

• final state: lqq bb - LEPTON+JETS

– Both W’s decay via Wqq• final state: qqqq bb ALL HADRONIC

e-e (1/81)

mu-mu (1/81)

tau-tau (1/81)

e -mu (2/81)

e -tau (2/81)

mu-tau (2/81)

e+jets (12/81)

mu+jets(12/81)

tau+jets(12/81)

jets (36/81)

the top quark does not hadronize. It decays as a free quark!

Lepton provides a good trigger, all jets are tough


Top ID in “lepton+jets” Top ID in “lepton+jets” channelchannel

• 2 b-jets • Lepton: electron or muon• Neutrino (from energy

imbalance)• 2 q’s – transform to jets of

particles• Note that these two jets

come from a decay of a particle with well measured mass – W-boson – built-in thermometer for jet energies

lWorqqW

bWt

ttpp

'


DDØ detectorØ detector• Electrons are identified as

clusters of energy in EM section of the calorimeter with tracks pointing to them

• Muons are identified as particles passing through entire detector volume and leaving track stubs in muon chambers. Track in the central tracking system (silicon+SciFi) is matched to track in muon system

• Jets are reconstructed as clusters of energy in calorimeter using cone algorithm DR<0.5


Top mass using matrix Top mass using matrix element method in Run Ielement method in Run I

• Method developed by DØ (F. Canelli, J. Estrada, G. Gutierrez) in Run I

Systematic error dominated by JES 3.3 GeV/c2

With more statistics it is possible to use additional constraint on JES based on hadronic W mass in top events – in situ calibration

Single most precise measurement of top mass in Run IMt =180.1±3.6(stat) ±4.0(syst) GeV/c2


Matrix element methodMatrix element method

• Goal: measure top quark mass• Observables: measured momenta of jets and leptons • Question: for an observed set of kinematic variables x

what is the most probable top mass • Method: start with an observed set of events of given

kinematics and find maximum of the likelihood, which provides the best measurement of top quark mass

• Our sample is a mixture of signal and background

)()1(),(),( sgn xPfmxPfmxP bkgtopttoptevt


Matrix Element MethodMatrix Element Method

Normalization depends on mt

Includes acceptance effects

probability to observe a set of kinematic variables x for a given top mass

Integrate over unknown q1,q2, y

f(q) is the probability distribution than a parton will have a momentum q

dnσ is the differential cross sectionContains matrix element squared

t

t

W(x,y) is the probability that a parton level set of variables y will be measured as a set of variables x

bq’

q

),()()();()(

1);( 2121sgn yxWqfqfdqdqmyd

mmxP t

n

tt


Transfer functions Transfer functions (parton(partonjet)jet)

• Partons (quarks produced as a result of hard collision) realize themselves as jets seen by detectors– Due to strong interaction partons turn into

parton jets– Each quark hardonizes into particles

(mostly and K’s)– Energy of these particles is absorbed by

calorimeter – Clustered into calorimeter jet using cone

algorithm• Jet energy is not exactly equal to parton

energy– Particles can get out of cone– Some energy due to underlying event (and

detector noise) can get added– Detector response has its resolution

• Transfer functions W(x,y) are used to relate parton energy y to observed jet energy x


Dependence of JESDependence of JES

dependence of JES is derived on +jet data, but the overall scale is allowed to move to optimize MW


• All jets are corrected by standard DØ Jet energy scale (pT, )

• Overall JES is a free parameter in the fit – it is constrained in situ by mass of W decaying hadronically

• JES enters into transfer functions:

JES in Matrix ElementJES in Matrix Element

JES

EJES

EW

JESEEWp

j

pj

),(),,(


Normalization

e+jets μ+jets

),()()();()(

1);( 2121sgn yxWqfqfdqdqmyd

mmxP t

n

tt


Signal IntegrationSignal Integration• Set of observables – momenta of jets and leptons: x• Integrate over unknown

– Kinematic variables of initial (q1,q2) and final state partons (y: 6 x3 p) = 20 variables

– Integral contains 15 (14) -functions for e()+jets• total energy-momentum conservation: 4• angles are considered to be measured perfectly: 2x4 jet +2 lepton • Electron momentum is also considered perfectly measured, not true for muon

momentum: 1(0)– 5(6) dimensional integration is carried out by Vegas– The correspondence between parton level variables and jets is established

by transfer functions W(x,y) derived on MC• for light jets (from hadronic W decay)• for b-jets with b-hadron decaying semi-muonically• for other b-jets

• Approximations– LO matrix element– qqtt process only (no gluon fusion – 15%)


Background integrationBackground integration

• W+jets is the dominant background process• Kinematics of W+jets is used as a representation

for overall background (admixture of multijet background is a source of systematic uncertainty)– Contribution of a large number of diagrams makes

analytical calculation prohibitively complex– Use Vecbos

• Evaluate MEwjjjj in N points selected according to the transfer functions over phase space

• Pbkg- average over points


Sample compositionSample compositionLepton+jets sample

– Isolated e (PT>20GeV/c, ||<1.1)

– Isolated (PT>20GeV/c, ||<2.0) – Missing ET>20 GeV– Exactly four jets PT>20GeV/c, ||

<2.5 (jet energies corrected to particle level)

Use “low-bias” discriminant to fit sample composition – Used for ensemble testing and

normalization of the background probability.

– Final fraction of ttbar events is fit together with masse+jets +jets

# of events 70 80

Signal fraction 45±12% 29±10%


Calibration on Full MClepton+jets


calibrated calibrated

expected: 36.4%

DØ RunII Preliminary

Mt=169.5±4.4 GeV/c2

JES=1.034±0.034


Systematics summarySystematics summary

Source of uncertainty Effect on top mass (GeV/c2)

B-jet energy scale +1.32-1.25

Signal modeling (gluons rad)

0.34

Background modeling 0.32

Signal fraction +0.5-0.17

QCD contribution 0.67

MC calibration 0.38

trigger 0.08

PDF’s 0.07

Total +1.7-1.6


B-jet energy scale● Relative data/MC b/light jet energy scale ratio

•fragmentation: +-0.71 GeV/c2

different amounts of 0, different + momentum spectrum fragmentation uncertainties lead to uncertainty in b/light JES ratio

compare MC samples with different fragmentation models: Peterson fragmentation with eb=0.00191 Bowler fragmentation with rt=0.69

•calorimeter response: +0.85 -0.75 GeV/c2

uncertainties in the h/e response ratio + charged hadron energy fraction of b jets > that of light jets corresponding uncertainty in the b/light JES ratio

•Difference in pT spectrum of b-jets and jets from W-decay: 0.7 GeV/c2


Gluon radiationGluon radiation

• The effect is reduced by – Requiring four and only four jets in the final state

– High PT cut on jets

• Yet in ~20% of the events there is at least one jet that is not matched (DR(parton-jet)<0.5) to top decay products– These events are interpreted as background by ME method

• We study this systematic by examining ALPGEN ttj sample and varying its relative fraction between 0 and 30% (verified on our data by examining the fraction of events with the 5th jet)

• Final effect on top mass 0.34 GeV/c2


● W+jets modeling: +-0.32 GeV/c2

study effect of a different factorization scale for W+jets events (<pT,j>

2 instead of mW2 + SpT,j

2)

● PDF uncertainty: +-0.07 GeV/c2

CTEQ6M provides systematic variations of the PDFs reweight ensembles to compare CTEQ6M with its systematic variations (by default the measurement uses CTEQ5L throughout: use a LO matrix element, and for consistency with simulation)

Signal/Background Modeling

● QCD background: +-0.67 GeV/c2

Rederive calibration including QCD events from data (lepton anti-isolation) (note: sample statistics limited) can be reduced in the future


● Signal fraction: +0.50 -0.17 GeV/c2

Fitted top mass depends slightly on true signal fraction (if signal fraction is smaller than expected): => Vary signal fraction within uncertainties from topological likelihood fit - Note: ftop fit yields identical

result with factor √2 smaller uncertainties

Signal fraction

Cross check on data: cut on log10(pbkg)<-13 Ftop=31%46±6%Mtop=170.2±4.1 GeV/c2


Systematics summarySystematics summary

Source of uncertainty Effect on top mass (GeV/c2)

B-jet energy scale +1.32-1.25

Signal modeling (gluons rad)

0.34

Background modeling 0.32

Signal fraction +0.5-0.17

QCD contribution 0.67

MC calibration 0.38

trigger 0.08

PDF’s 0.07

Total +1.7-1.6


Result and cross checksResult and cross checks

• Run II top quark mass based on lepton+jets sample: Mt=169.5 ±4.4(stat+JES) +1.7

-1.6 (syst) GeV/c2

• JES contribution to (stat+JES) 3.3 GeV/c2

• Break down by lepton flavor– Mt(e+jets)=168.8 ±6.0(stat+JES) GeV/c2

– Mt(+jets)=172.3 ±9.6(stat+JES)GeV/c2

• Cross check W-mass


Summary of DSummary of DØØ M Mtt

measurementsmeasurements• Statistical

uncertainties are partially correlated for all l+jets Run II results

DØ Run II preliminary


Projection for uncertainty on Projection for uncertainty on top quark masstop quark mass

Assumptions:• only lepton+jets channel considered • statistical uncertainty normalized at

L=318 pb-1 to performance of current analyses.

• dominant JES systematic is handled ONLY via in-situ calibration making use of MW in ttbar events.

• remaining systematic uncertainties: include b-JES, signal and background modeling, etc (fully correlated between experiments) Normalized to 1.7 GeV at L=318 pb-1.

• Since most of these systematic uncertainties are of theoretical nature, assume that we can use the large data sets to constrain some of the model parameters and ultimately reduce it to 1

GeV after 8 fb-1.


Combination of Tevatron Combination of Tevatron resultsresults

JES is treated as a part of systematic uncertainty, taken out of stat error


CombinationCombination

• Mt=172.7±2.9 GeV/c2

• Stat uncertainty: 1.7GeV/c2

• Syst uncertainty: 2.4GeV/c2

• hep-ex/0507091

• Top quark Yukawa coupling to Higgs boson

• gt=Mt√2/vev=0.993±0.017


What does it do to Higgs?What does it do to Higgs?

• MH=91+45-32GeV/c2

• MH<186 GeV/c2 @95%CL

MH,GeV/c2 Mt,GeV/c2

MW

,GeV

/c2

68% CL



ttbar cross section in ttbar cross section in l+jetsl+jets with b-tagwith b-tag

• Isolated lepton – pT>20 GeV/c, |e|<1.1, ||<2.0

• Missing ET>20GeV

• Four or more jets – pT>15 GeV/c, |

=8.1+1.3-1.2(stat+syst)±0.5(lumi) pb

DØ RunII Preliminary, 363pb-1

≥4j, 1t ≥4j, 2t

Expect bkg 21.8±3.0 1.9±0.5

Observe 88 21


Cross section summaryCross section summaryDØ RunII Preliminary

pbttpp ),(

Submitted for publication Updates


ttbar resonances in ttbar resonances in l+jetsl+jets with with b-tagb-tag

• Check ttbar invariant mass for possible resonance production


• Events are kinematically constrained – mT=175GeV/c2

– Leptonic and hadronic W masses

NNLOtt)=6.77±0.42


ttbar resonances in ttbar resonances in l+jetsl+jets with with b-tagb-tag

• Limit M(Z’)>680 GeV/c2 with =1.2%MZ’ at 95%CL

*R. Harris, C. Hill, S. Parke hep-ph/9911288


*

Run I limit 560 GeV/c2

Run II limit 680 GeV/c2


ConclusionConclusion

• First DØ RunII top mass measurement in l+jets channel to surpass Run I precision– Mt=169.5 ±4.4(stat+JES) +1.7

-1.6 (syst) GeV/c2

• Developed method for in situ jet energy scale calibration using hadronic W-mass constraint

• Combined Tevatron top mass measurement reaches a precision of 1.7%

• ttbar production cross sections updated for l+jets channel• Invariant mass of ttbar system probed for resonance

production, exclusion limit for M(Z’)>680 GeV/c2 at 95%CL

Backup slidesBackup slides


Parton Level TestsText


L+jets sample composition L+jets sample composition


Kinematics in l+jets sampleKinematics in l+jets sampleDØ RunII Preliminary, 363pb-1

Documents

Top quark mass