High Pt Physics:High Pt Physics:from the Tevatron to LHCfrom the Tevatron to LHC

Tommaso DorigoUniversity of Padova and INFN

• Introduction: the Tevatron, CDF and D0 in Run II• Tools for high-Pt physics: jets, leptons, b-tags, and all that • Higgs boson searches and prospects• Top quark physics searches and prospects• Electroweak physics searches and prospects

• Not discussing today: • Precision QCD measurements• Searches for new Physics / BSM / SUSY

• Conclusions and perspectives

Corfù, September 6th, 2005

What this talk is not• Not a showroom

– skipping / forgetting / ignoring many interesting new results

– some analyses only briefly mentioned– not giving a complete panorama

• Not a fair balance between CDF and D0– actually totally unfair– mostly focusing on CDF

• Not a snapshot of where we stand– rather, a view of the issues we are facing in high-Pt

physics in preparation for the LHC

The present trenches of The present trenches of high-Pt physics: high-Pt physics:

the Tevatronthe Tevatron

The Tevatron in Run II• Massive upgrade with respect to Run I,

to increase L by 1.5 orders of magnitude– Main injector, pbar recycler– crossing time from 3.5 s to 396 ns– increased antiproton yield and transfer

efficiency

• From a endured start in 2001-2002, the Tevatron is now working excellently

– So far collected more than 800 pb-1 /exp– Peak instantaneous luminosity by now

regularly above 1032

– less downtime, fewer stops for beam studies needed – just fine smooth running

• In 2005-2006 crucial upgrades are being worked at to complete the picture

• electron cooling• stacktail bandwidth upgrade

• Two foreseen plans for data accumulation • Base plan: the minimal objective• Design plan: if everything works great

Run II: where we are right now

Integrated Weekly Luminosity (pb-1)

0

10

20

30

40

50

60

10/1/03 9/30/04 9/30/05 9/30/06 9/30/07 9/29/08 9/29/09

electron cooling

stacktail bandwidth upgradeDesign

Base

Total Luminosity (fb-1)

0

1

2

3

4

5

6

7

8

9

10/1/03 9/30/04 9/30/05 9/30/06 9/30/07 9/29/08 9/29/09

Design

Base

Have been following design curve!Upgrades continuing – electron cooling of antiprotons is critical. As L increases, CDF and D0 catching up by modifying trigger tables, improving DAQ Design curve means 8 fb-1 by 2009!

WE ARE HERE

CDF and D0 in Run 2CDF and D0 in Run 2

The CDF DetectorCDF significantly upgraded from Run 1:

• New L00+SVX+ISL silicon detector• New central tracker • Extended muon coverage to ||<1.5• New end-plug calorimeters• SVT measures IP to 45 m at Level 2!

The challenge is now a smooth operation for many years of running…

The D0 DetectorMassively upgraded from Run 1 to include:• 77,000 ch scintillating fiber tracking • 2.0 Tesla solenoid• 800,000 channel silicon detector (4 barrel layers, 2-sided disks)• Extended muon coverage (MDT)Tracker working well despite lowvolume (R=1/3 RCDF)High performance b-tag to ||<2.0

Tools for high-Pt Tools for high-Pt PhysicsPhysics

The most common animals: JetsIn hadronic interactions, jets of hadrons are the most common things one can observe

They are common, but are they obvious to define ?

“Obvious: something you may think about for 20 years and maybe understand”

After 20 years of studies of pQCD, we think we understand what is going on…

What we measure in our detectors is the combination of a multitude of effects

Disentangling them is the keyto understanding each of them better

Identification and measurement of hadronic jets

Both CDF and D0 mainly use a cone algorithm (R=0.4 or 0.5) to identify localized depositions of energy in their calorimeters and measure hard partonsOther algorithms (midpoint, Kt) are mainly used in QCD studies

When faced with the measurement of the kinematics of hardparton emissions, one has to deal with two distinct issues:

- SCALE: to calibrate the energy response, to minimize the averagemeasurement error on a sample of jets

- RESOLUTION: to improve the precision of the energy measurement, decreasing the measurement error on an individual jet

The first issue is fundamental for precision mass measurementsof hadronically decaying objects (e.g. top quarks)The second issue is critical for the successful identification oflow S/N signals (e.g. Higgs bosons)

The JetClu Algorithm• Was initially designed to meet specifications from the Snowmass

Accord (1992)– A seeded, iterative cone algorithm, with R=0.7– custom prescription for splitting and merging

• Several drawbacks – not best option for QCD measurements– pQCD uses larger cone (Rsep=1.3) to emulate experimental procedure– not collinear safe, not IR safe (see next slide)

• But also strong points– conceptually simple– sensible choice for 2 TeV physics– makes it easier to compute corrections and systematics

• Start from Et-ordered list of seed towers (Et>1 GeV)– do preclustering by creating list of cones centered on seed towers,

removing seeds as they are englobed in cones– then add to cones Et of towers within, recompute baricenter, move

cone, to convergence– if two cones share too much energy (>75%) they are merged

Shortcomings of standard cone algorithmsInfrared safetyThe jet multiplicity changesif an arbitrarily soft emissionis detected between two partons the cone algorithm does not give a stable answer in the IR limit

Collinear safetyReplacing a massless partonby the sum of two collinearparticles a jet may fail detection due to lack of a seed, and the jet multiplicity changes

Fixed-order pQCD calculations contain uncanceled divergencies…

The Midpoint Algorithm

• Conceived to remove some of the problems of JetClu when compared to theoretical calculations

• IR safety mended by introducing imaginary seeds at midpoint of each pair of jets close in angle, and iterating to convergence

The Kt algorithmQCD appears to separatepartons into different jets according to their relativetransverse momentumThe Kt algorithm is thereforepreferred by theory, and comparisons between experimentalmeasurements and theoreticalcalculations are more straightforward

One event, three algorithms

Calibration of Jet Energy• To calibrate the energy measurement in CDF we use a detector-dependent

correction, a scale correction, and a treatment of additional small physical effects

– eta-dependent correction dijet balancing– multiple interaction correction f(Nvtx)– absolute scale correction: E/p of single tracks are used to tune the MC, which is

then used to derive “calhad” corrections.– last, out of cone and underlying event corrections are made– Systematic errors reduced to 3% (data/MC comparisons, -jet balancing)

• Calorimeter stability, MC (fragmentation, simulation of single particle resp.) • Understanding of out-of-cone radiation and UE• Simulation of response function versus jet rapidity

• D0 has an almost-compensated calorimeter (e/ <1.05, linear with energy ); disuniformities and gaps among cryostats need to be corrected

– EM part is calibrated with Zee decays – U noise measured in situ; other offset corrections address pile-up (energy from

previous interactions) and underlying event– Response is measured as a function of rapidity and Et with gamma-jet events – showering correction: Et flux vs R off jet cones

CDF Jet Energy CorrectionsPt

corr = (Ptraw frel – MI) fabs – UE + OOC

The b-Jet Energy Scale Issue• b-jets are different from generic jets

– large mass of leading hadron– semileptonic decay– hard fragmentation

• Originally thought the most pressing issuefor precision Mtop measurements

– after demonstration of auto-calibration with Wjj the picture is brighter

– residual systematics of b-JES to top mass estimated at less than percent level

– but that is MC extrapolation… Need to measure b-JES anyhow!

• To calibrate b-jets, CDF exploits the SVT to trigger on Zbb events in Run II

– extract signal, fit, get scale from Z mass (more on that later)– But this technique is unfeasible at the LHC

• background cross section is huge• rate of any b-jet trigger impossible to handle

• Another possibility is searching for gamma-b events– balancing the photon in the transverse plane with the jet, one obtains

a calibration– but b-fraction of jets is typically 40-50% even after a tight b-tagging

by secondary vertex identification– D0 and CDF currently studying this technique – expect results soon

b-jet calibration with B events• Use the MPF method:

– select back-to-back j events– determine Rhad from missing Et

projection – apply b-tagging, separate into different

samples for more handles• resulting sample is a mixture of b and

charm, light quarks• use also tighter b-tag by exploiting mass

of tracks in secondary vertex– can fit for Rb

What to do at the LHC ?

• Zbb signal extraction is unfeasible• gamma-B balancing techniques might work – studies are ongoing• Calibrations using top quark decays are possible, but one would

prefer an independent determination• I have a suggestion: use Zbb eventsAdvantages:

– Automatically selects qq initial state, boosting the S/N by an order of magnitude at typical TeVatron energy, surely more at LHC, with respect to inclusive Zbb vs gluonbb ratio

• Typical initial state of gluonbb does not produce photons!– Can fully exploit dedicated detectors for H – Resolution on E is so good, one can determine b-jet scale by just

looking at jet-jet ANGLE!Disadvantages:

– Statistical power is limited by small cross section

Improving the jet energy resolution

• Calibrating the calorimeter response to streams of hadrons is one of the foundations of mass measurements– It is a correction to the average systematic offset to the

measurement

• But the precision of an individual jet’s energy determination is no less a foundation– Separation of reconstructed hadronic resonances

(W,Z,top,Higgs, other fancier animals) critically depends on it– Even continuous-Q2 distributions benefit from a more precise

measurement– Less known is that top mass measurements do benefit greatly

from improved resolutions even in high S/N samples

Tools for the improvement of the Et resolution

• CDF has taken seriously the challenge to improve the jet Et resolution– Triggered by HSWG studies (more later)– Issue is complex: resolution can be improved in different ways depending

on event characteristics, jet rapidity, flavor of parton…– focus is improvement of dijet mass resolution through more precise jet Et

measurement– also focusing on b-jets

Three candidate algorithms identified and studied:

- H1 algorithm: use tracker for central charged hadrons- Track+Cal algorithm: categorize cal towers, disentangle photon response, use tracker for charged tracks- Hyperball algorithm: olistic approach to the problem. Use ALL information on jet measurement, exploit intercorrelation between jet observables and Et measurement error

The Hyperball algorithm: statement of the problem

• From an idea developed for the HSWG, 2003• Imagine one measures a scalar quantity (say the Et of a jet), which is

subject to all sorts of biases• Alongside with Et, one measures heaps of other characteristics of the jet

– several quantities in the calorimeter– track Pt information– photon clusters in Strip Chambers– b-tagging information– etcetera

• Many of the latter carry information about the biases of the Et measurement– for instance, a charged fraction larger than one speaks of a undermeasurement

of calorimeter Et• Simple minded approaches to remove biases neglect cross-correlations

– First I correct for the charged fraction, then the presence of muons, then the missing Et along the jet direction… In the end the computed biases wash out each other to some extent

• How to correct for these biases all at once ?

Basics of the algorithm• Main hypothesis: A scalar field Et:RNR exists and is continuous• Its value is the average error (pos or neg) in the jet Et measurement

performed in the calorimeter, as a function of all thinkable jet observables: Et = Etmeas- Ettrue

• Cannot determine Et with infinite precision• How to best measure it ?• Hyperball method:

– Fill RN with MC b-jets (we know Et for them!)– Need to average locally the value of Et – What does locality mean ?

• close to point to be estimated• similar value of most important variables• smaller correlation variables are less important for averaging

– Generalized distance in RN: D2(x,y) = iwi(xi-yi)2

– Use D to find MC points closest to point where scalar field is needed– Need to determine W vector such that closest MC points provide best estimate of

<Et> – Geometrically it means determining shape of hyperellipsoids

Applied the algorithm toB-jets (QCD direct prod.)Resolution improvesby about 30% throughoutthe Et spectrum studied

That means we can reallyget back the 10% relativeresolution we promised forHbb decay searches

Still lots to improve:-refine list of variables used-use more MC for Et estimates-optimize everything

Promising results… Work in progress!

Identification of High Pt LeptonsMost high Pt final states studied at the Tevatron involve the detection of leptons

- easy to trigger on- high signal purity- easy to calibrate using standard candles (W,Z

bosons)

Tevatron experiments are exploiting to the fullest these signatures, producing lots of precision Electroweak physics measurements with them

Tau leptons are also beginning to contribute appreciably, especially to new physics searches which may be generation-dependent

CDF

D0

CDF

Tagging b-jetsIdentifying b-jets is of paramount importance for low-mass Higgs boson searches.Three methods are well-tested and used:– Soft lepton tagging– Secondary vertex tagging– Jet Probability tagging

For double tag searches, efficiency factors get squared! To retain signal, both CDF/D0 have loose and tight tagging optionsEfficiency drops at low jet Et and high rapidity but is 45-50% for central b-jets from Higgs decayMistag rates are kept typically at 0.5%

I.P.B

SV tagging: tracks with significant IP are used in a iterative fit to identify the secondary vertex inside the jet

D0

CDF

Tight/loose SV tag eff.

Secondary vertex tagging

This event display shows how charged tracks are used to fit for secondary vertices in jets from a ttbar candidate (single lepton decay)

Decay lengths for 50 GeV b-jets are typically of the order of a few millimeters and they can be easily reconstructed with tracks having at least 3 associated hits in the silicon detectors (d is around 20 microns)

The last resort – or the main one?The Monte Carlo Simulation

• The technology of reproducing the known behavior of high energy interactions has reached exquisite heights

– Now available several choices which model QCD (let alone EW interactions) very successfully

• full matrix element computations and parton shower modeling agree better by the day– But tuned with the data… Will they stand the test of a x7 jump in CM energy ?– Unfortunately we are still critically dependent on PDF fits

• even larger extrapolation in the unknown at LHC (more on that later)

• Almost every analysis of high Pt processes now relies heavily on Monte Carlo simulations

• Let’s not forget that gross mistakes are brought by relying too much on MC to extrapolate into the unknown

• Need to keep a cool head– Lesson for the future from the past: use of data is fundamental at the start of a

new endeavour, as was in CDF and D0 in the early days of top searches• method 1 vs method 2• likelihood methods vs pure counting experiments

“Simulation”

From the Latin “simulacrum”…

My Webster’s offers the following:

1) The act of simulating; pretense; feigning.

2) A simulated resemblance

3) An imitation or counterfeit

4) The use of a computer to calculate, by means of extrapolation, the effect of a given physical process

Higgs Boson SearchesHiggs Boson Searches

SM Higgs: Production and Decay

mH (GeV/ c2)

Exclud

ed

mH (GeV/ c2)

Exclud

ed

At the Tevatron, about five 120 GeV Higgs bosons are produced in a typical day of running (will be 15/day in two years).

Direct production occurs mostly via gluon-gluon fusion diagrams.Associated production through a virtual W or Z boson provides sensitivity in the region where LHC will have more trouble. At higher mass, the WW(*) final state becomes dominant.

e

W*H

W

q

qb

b

l

Even the WHWWW(*) process is promising despite the low yield, due to the striking signature of missing Et plus three leptons, two of which may be of the same charge but different flavor.

What we know about the Higgs

• Although they did not directly observe it, the LEP experiments have collected a wealth of information on the Higgs boson through comparisons of EW observables to EW theory + radiative corrections

• From theory we know its couplings, its decay modes, and how its mass impacts the W and top masses.

• If it exists, then we know its mass with about 60 GeV accuracy, and the direct search limit already cuts away a large part of the allowed mass region

• Latest LEP results: MH=126+73-48 GeV,

MH<280 GeV @ 95% CL (Winter ‘05) now being updated for new Mtop…

Higgs Sensitivity WG PredictionsIn 2003 the Tevatron chances for Higgs discovery were re-evaluated

Idea: with available data and operating detectors, can better assess Tevatron reachSurprisingly, the new results meet or exceed 1998 Susy/Higgs WG ones.

CDF

BASEDESIGN Keys to success:

-mass resolution improvements;- optimized b-tagging;- shape information vs counting.

Lu

m (

fb-1)

Can we see dijet resonances if they are there?

The S/N is not higher than 1/5 at the most in the signal region

– good testing ground for H!– can use to test/improve dijet mass resolution with

advanced algorithms

We barely saw it in Run 1…

Can we use it in Run 2 ??

A low mass Higgs search entails believing that we can:- appropriately reconstruct hadronically-decaying objects

- accurately understand our background shapes

All of that can be proven if we see the Zbb decay in our data.

CDF sees Zbb decays in Run 2!

Double b-tagged events with no extra jets and a back-to-back topology are the signal-enriched sample: Et

3<10 GeV, 12>3

Among 85,784 selected events CDF finds 3400±500 Zbb decays

- signal size ok- resolution as expected- jet energy scale ok!

This is a proof that we are in business with small S/N jet resonances!

CDF expects to stringently constrain the b-jet energy scale with this dataset

A few additional notes

• In Run II we are demonstrating that by measuring with precision the JES of light-quark jets using Wjj, the part of (Mt) due to modeling of b jets (decays, fragmentation, color connection) can be reduced to below 1% (more later).

• The Zbb signal becomes important mainly as a testing ground of algorithms targeting the jet resolution improvements

• Anyway Zbb decays may contribute appreciably to b-JES determinations: already with 300 pb-1 one gets a statistical error well below 2%

b-jet Et scale = dominant systematics in Run I top mass measurements • top decay is a two-body one• very nearly linear relationship between Eb and Mt

• At Tevatron, b Jet Energy Scale syst. is approx. (Mt) (GeV) ~ (Eb) (%)• At LHC, typical top quark boost softens the dependence: (Mt) (GeV) ~ 0.7 (Eb) (%)

- for light-quark jets (Mt) (GeV) ~ 0.3 (Eb) (%)

Search for WH in Run 2To search for WHlbb eventsa detailed understanding of the composition of the W+jets sample is mandatory. In the 2-jet bin CDF finds 187 events with a b-tag,where 175±26 are expected, mostly from Wbb production and mistags.

A fit to the dijet mass distribution allows to extract a 95% CL limit of 5 pb to SM WH production.The obtained limitis consistent bothwith a priori predictions and with expectations based on HSWG results.

Results with double tagged events

When two jets are required to be b-tagged, backgrounds arestrongly reduced and mostly Wbb, ttbar remainThe data is still in good agreement with expectations The extracted limit of WH production is 3-10 pb for MH=110-150 GeV

WH Search in D0D0 also study their W+2jet bin with b-tagging in 384 pb-1 of high-Pt leptons from Run 2 data. The dijet mass distribution shows no anomaly with 1 b-tag. The 2-tag distribution is divided in search windows to set limits to Higgs production. They find 4 events with two b-tags in the mass window centered on 115 GeV (exp. 2.4±0.6)

95%CL limits on WH*B(Hbb) are set at 7 to 9 pb for MH=105-135 GeV

By-product: a 95% CL limit is set to Wbb production (R>0.75, Pt>20 GeV) at 4.6 pb.

High Mass Searches: HWW(*)The SM production of WW pairs has been measured by CDF in Run 1 and by both CDF and D0 in Run 2: excellent agreement with NLO.

To search for Higgs boson decays, events with two high-Pt leptons (e,) and large missing Et are selected; the tt background is rejected with a jet veto.

Then both experiments use the helicity-preferred alignment of charged leptons in to discriminate known backgrounds.

W+ e+

W- e-

CDF results on HWW

8 events are observed in 184 pb-1 of Run 2 data with the Mll <80 GeV cut, with an expected background of 8.9±1.0.

A likelihood fit to the ll distribution is performed to extract a limit on the HWW cross section as a function of its mass.

The result is WW*B(WWll)>5.6 pb for MH=160 GeV.

CDF searches for HWW events by selecting two tight leptons (ee,e) with Et

e(Pt)>20 GeV and

missing Et>25 GeV (50 GeV if ll<20°). A strict jet veto (Et<15 GeV if ||<2.5) rejects top candidates.Finally, a small dilepton mass is required (Mll<55-80 GeV for MH=140-180 GeV).

Putting it all together…

Higgs Physics: perspectives The Higgs boson is being hunted at the Tevatron in all advantageous search channels. D0 and CDF are competing – that’s good! – but will soon start to also combine their results.

No surprises with the analyzed 200 pb-1 samples, but we have already three times more data on tape to look at!

We are on track to supersede the LEP2 lower limit on MH by 2007

By the end of 2009, the Tevatron might be able to see a MH=115 GeV Higgs at 5, or exclude it all the way to 180 GeV.

…but that will require both cunning and the Tevatron delivering according to the design plan!

What I feel I can promise at 95% CL: exclusion up to 135 GeV, 3evidence at 115 GeV.

Implications for LHC

• Scenario A: Tevatron design, LHC delays firs hints from CDF e D0 (3, early 2008) allow LHC to put their chips in the right place confirmation, common discovery (as did Adone for J/? Seems improbable…

• Scenario B: Tevatron design, LHC in time Tevatron “confirms” the first signal from LHC

• Scenario C: Tevatron base plan (or killed), LHC whatever you know the story.

LHC starts collecting physics data in April 2008 if everything works as it should, the Higgs is discovered by CMS and ATLAS in 2009 (a few fb-1 should suffice)However, fits prefer a Higgs mass in the region favoring Tevatron and hampering LHC…. Let’s hypothesize MH=115 GeV. Three possible scenarios:

Top Quark PhysicsTop Quark Physics

The Top Quark at the Tevatron• The top quark just turned 10! Run I results:

– (tt) =5.7±1.6 pb (D0), 6.5±1.4 pb (CDF) (@1.8TeV)– Mt = 178.0±2.7±3.3 GeV (D0+CDF) – many other measurements – but still imprecise – of Vtb, BR, spin; limits to

single production, non-SM production and decays.

• From the “discovery” mode the Tevatron soon adapted to using top quarks as a perfect pQCD laboratory

• As new data pours in, the plan is the same: first, cross section measurements are performed; then the mass, then the kinematics and the search of anomalies, and lastly, the measurement of intrinsic phhysical properties

• That modus operandi allows to optimize the output of physics results as analysis tools get perfected and more sophisticated: – high- Pt lepton identification – b-tagging– precise measurement of jet energy scale

A bit of history of Top Quark Quest

Production of top at Tevatron• At Tevatron production of ttbar pairs occurs by qq annihilation (85%) or

gluon fusion (15%) proportions inverted WRT LHC!• Theoretical cross section (NNLO) is 6.1 pb 1/1010 collisions 2 events

per hour • Single top production is not irrelevant (3 pb), but its signature is way less

characteristic so far obtained only upper limits to single top production

Top Quark DecaysSince Vtb=1 and Mt>Mb+MW, the decay tWb is dominant. Final states of ttpairs are classified according to the decay of the two W bosons:

Mt large t =f(Mt3) ~ 1.5 GeV >> QCD ~

0.2 GeV and thus:• t is produced and decays free;• polarization can be studied in the decay, since the depolarization time d ~ Mt/2 is much longer

Measurements exploit mainly three decay channels • Dileptonic: B=4/81, S/N ~ 3-10 • Single lepton: B=8/27, S/N ~2-5• All hadronic: B=4/9, S/N ~1/6

Top quark decays also constitute an excellent laboratory to study weak interactions of quarks freed from any QCD effects:

Cross section measurements

• The most precise tt measurements come from the analysis of single lepton decays compromise between S/N and yield

• “standard” recipe: – trigger on electrons or with Et (Pt)>15 GeV– offline selection: Et (Pt)>20 GeV– Missing Et>25 GeV– 3 or 4 jets with Et>15 GeV– B tagging and/or cut on Ht (sum of transverse energy or all objects,

including missing Et)• Both CDF and D0 use MC to estimate physical backgrounds and

data to handle false b-tags • “W+1 jet” and “W+2 jet” events are used to verify sample

composition and yield

Example: D0 lepton+jets with b-tag

The measurement from D0 uses events with a lepton (e, ), missing Et>30 GeV, and 3 or 4 jets with Et>20 GeV, and one or two b-tags. The eight determinations of the signal (e or mu, 3 or 4 jets, 1 or 2 b-tags) are used in a likelihood fit to obtain tt = 8.6 ±1.6 ±0.6 (lum.) pb

Single tags Double tags

tt in the all-hadronic Channel • The “least fortunate” of measured

channels • First measurement in Run I, now

repeated by CDF

• Selection: multi-jet trigger, and– >=6 jets (Et>20 GeV) – >=1 b-tag– kinematical selection (centrality, Et,

Aplanarity, subleading Et)

• Background (essentially QCD HF prod. with radiation, and fake HF+rad.) estimate by parametrizing the probability P(Et,Ntrk,…) to get a b-tag – works well pre-kinematic selection

• With 311 pb-1:

tt= 7.5 ± 1.7 (stat.) +3.3-2.2 (syst.) pb

Summary of measurements of tt

CDF

The ttbar production is studied at the Tevatron in all significant final states andwith different methodologies. The experimental error is still larger than the theoreticalone (NNLO, Cacciari, Kidonakis – 15%) but is reaching it quickly

Top mass measurements

27348 /126 cGeVM H

..%95@/280 2 LCcGeVMH

The final Run I results, recently updated with the latest measurement by D0, reach a combined precision of 2.5% In Run II the goal is to reach a precision close to 1% per experiment That would mean getting very close to LHC promises and not far to the physical limit of precision with direct reconstruction techniques

CDF Template Analysis Overview

Wbb MCData

tt MC

Datasets

Templates

2 mass fitter:Finds best top mass and jet-parton assignmentOne number per eventAdditional selection cut on resulting 2

Result

Likelihood fit:Best signal + bkgd templates to fit dataCompare to paramiz’n, not directlyConstraint on background normalization

Likelihoodfit

Massfitter

Result with 318 pb-1

Mtop = 173.2 +2.9-2.8 (stat) ± 3.4 (syst) GeV/c2

Green histos: data distributions

Curves: expected signal andbackground from global best fit

Systematics Summary

Jet Systematic Source Uncertainty(GeV/c2)

Relative to Central 0.6

Hadronic energy (Absolute Scale) 2.2

Parton energy (Out-of-Cone) 2.1

Total 3.1

Systematic Source Uncertainty(GeV/c2)

Jet Energy Scale 3.1

B-jet energy 0.6

Initial State Radiation 0.4

Final State Radiation 0.4

Parton Distribution Functions

0.4

Generators 0.3

Background Shape 1.0

MC statistics 0.4

B-tagging 0.2

Total 3.4

Systematics dominatedby jet energy scale.

Was 6.8 !!- Reduced double counting- Tuned simulation to data

W mass resonance in tt events!

• Can we use Wjj mass resonance to constrain JES?

• Mtop measurement sensitive primarily to energy scale of b jets. (W mass constraint in 2.)– But studies show most

uncertainty is shared by light quark, b jets.

– Only 0.6 GeV/c2 additional uncertainty on Mtop due to b-jet-specific systematics.

So use Wjj to improveunderstanding of q jets,therefore b jets, thereforeMtop.

This constraint will onlyimprove with statistics!

Measure JES using dijet massBuild templates using invariant mass mjj of allnon-tagged jet pairs.

• Rather than assuming JES and measuring MW...

• Assume MW and measure JES

• Parameterize P(mjj;JES) same as P(mt

reco;Mtop)

The “2D” measurement

• Too many correlations to treat this as an independent measurement of JES.

• Take the plunge and fit for Mtop and JES simultaneously…

– Need “2D” templates: P(mtreco;Mtop,JES) and P(mjj;Mtop,JES).

– More complex, but still tractable.– Constrain to prior knowledge: JES = 0 ± 1.

• Advantages:– Improve uncertainty on JES (dominant systematic)improve

uncertainty on Mtop.

– With this method, JES uncertainty begins to scale directly with statistics!

Apply 2D fit to the data

• Reported error includes both “pure statistics” and (reduced) JES systematic.

• Breaks down to+2.7

-2.6 (stat) ± 2.5 (JES)

• 20% improvement in uncertainty due to JES!

JES = -0.10 +0.78-0.80

(stat only) 27.3

6.3top GeV/ (syst.) 7.1JES)(stat. 5.173 cM

29.28.2top GeV/ (syst.) 5.1(JES) 1.3(stat.) 2.173 cM

1D resultwas:

Current Top mass MeasurementsCDF took full advantage of the new jet energy corrections to surpass D0’s precision recenlty; D0 still has a systematic uncertainty close to 6 GeV on jet energy scale, but will soon improve both statistics and systematics of its measurements

Other top physics results

Besides cross section and mass, CDF and D0 have begun to study in detail production and decay characteristics of top quarks

See talk by Gervasio G.this afternoon

Electroweak PhysicsElectroweak Physics

Production of W and Z bosons

At the Tevatron W and Z bosons are fundamental tools for calibrations and checks - EM energy scale - Momentum scale in tracker - Studies of the resolution in missing transverse energy - Studies of the calorimeter response to low Et hadronic activity (boson recoil) - input to PDF at low x from W charge asymmetry measurementsTheir cross section, known at NNLO (2% precision), may provide an important normalization point to sidestep the luminosity error

Assorted signalsW bosons collected by CDF and D0 are now close to a million, and 105 the Z bosons(but analyzed signals are still catching up WRT these numbers)

W with high rapidity electrons

• CDF measured the We process with forward electrons (1.1<<2.8) using the ISL and the new plug calorimeter

• Important measurement for the determination of the PDF’s (d/u a x piccolo) thanks to the asymmetry in production and decay (PDF e V-A)

• Important also to detect effect from soft gluon resummation in low-x production impact in all LHC cross section measurements

W = 2.874 ± 0.034(stat) ± 0.167(syst) ± 0.172 (lum.) nb.

These new CDF W asymmetry data points will be included in the MRST04 fits soon

Zooming out: 22 years of EW measurements

Wl as a luminosity monitorAll cross section measurements refer to the inelastic cross section, which is known with a 4% precision Given the high precision reached by NNLO calculations for (W), one may think of using the latter as a normalization point of integrated luminosity of a given dataset HEP-PH/0405130 (Frixione, Mangano) investigates the uncertainties from Tevatron and LHC acceptance and other effects It appears possible to measure (W) to within1-2% both at Tevatron and LHC – it is critical however to study the PDF’s at high rapidity

However, it is necessary to known with precision the stability of data taking and the collection efficiencies…

A possible important contribution of Tevatron to LHC: the PDFs ?

At LHC all cross section measurements will strongly depend on the knowledge of PDF at low x

- would benefit from high-rapidity W/Z measurements at Tevatron- Heavily relying on HERA data

LHC goal on W mass: 15 MeV requires an accurate knowledge of production mechanisms at low x (5x10-4 : 10-2)

- Top mass measurement systematics-the search of new physics signals with low cross section is often based on an accurate knowledge of the rate of production of physics backgrounds

How to measure PDFs at low x ?

• In order to determine the PDFs at low x, one needs to study light things going forward …

• An attempt: quarkonia in D0…Still lacking enough statistics.

0 1 2 3 4200

300

400

500

600

700

800

900

1000

MRST2002

y

data: D0theory: gg normalised to data

B

d/

dy

(p

b)

CTEQ610-7

10-6

10-5

10-4

10-3

10-2

10-1

100

100

101

102

103

104

105

106

107

108

109

fixedtarget

HERA

x1,2

= (M/1.96 TeV) exp(y)Q = M

Tevatron parton kinematics

M = 10 GeV

M = 100 GeV

M = 1 TeV

422 04y =

Q2

(GeV

2 )

x0 1 2 3 4

2000

3000

4000

5000

6000

7000

8000

9000

10000

MRST2002

y

Tevatrontheory: gg arbitrary normalisation

B

d/

dy

CTEQ6

x=10-3, Q2=100 GeV2 x=10-4, Q2=10 GeV2

W mass measurements

Pt scale Calibrations

Calibration of EM Et scale

Measurement of hadronic recoil

Hadronic recoil, cont’d

Systematic uncertainties on MW

Summary of MW measurement

• One of the most complex measurements at the Tevatron – requires a perfect understanding of tracker and calorimeter, and an optimization of calibrations

• Analyses are reaching conclusion • The CDF systematic uncertainty is already defined

and slightly better than Run I one (76 vs 79 MeV)• D0 is finalizing calorimeter calibrations… Results are due

soon • The total MW error per experiment from Run II is

expected to be about 40 MeV (compare to best single measurement to date: Aleph, MW=58 MeV)– LHC will be able to improve these measurements by a factor 2,

but accurate PDF measurements are required at low x

Conclusions and Conclusions and PerspectivesPerspectives

Conclusions and perspectives

• Run II at the Tevatron has entered a mature stage, and precision measurements are coming out – Mt= 173.5+3.7

-3.6 ±1.7 (syst) GeV(CDF)– Mt= 172.7±2.9 GeV (CDF+D0)– First improvements in MW determinations expected soon – Improved Higgs mass limits foreseen for 2007

• Many analysis tools are being perfected even in view of a possible use in LHC – tools for energy calibration – b-jet energy scale may not be necessary – Luminosity monitoring with W production– PDF at low x important both for measurements and searches!

• The Higgs boson might be at reach of CDF and D0 before LHC starts playing the game

• Many a precision QCD measurement and new limits to supersymmetric processes and other new physics are public . but did not fit here…

The inclusive jet cross sectionNel Run 1 si osserva un disagreement con NLO QCD a CDFD0 non conferma né smentisceUna modifica delle PDF (CTEQ4HJ) con tweaking di g(x) ad alto x permette di riaggiustare le coseNel Run II, misura ripetuta con sezione d’urto x1.6 nei bins di maggior Et sempre un eccesso, ma minore di prima…

Single top production

0

50

100

150

200

250

300

350

400

10/01/03 12/31/03 03/31/04 06/30/04 09/29/04

Date

Inte

gra

ted

Lu

min

os

ity

(p

b-1

)

Design

Base

Actual

2005 : so far along Design

2004

2003

2002

Design for 2005 Base for 2005

Jetclu e Midpoint

L’algoritmo Kt a CDF

CDF: aWWee candidate

CDF: a WWecandidate

Method checks

• Prove to ourselves that parameterizations and likelihood machinery work: measure the top quark mass in MC samples.

• Mtop fit unbiased across input top mass and jet energy scales.

• Reported uncertainty scaled by ~1.03 as shown (effect of non-Gaussian likelihood).

Misura di tt nel canale +jets

• Il decadimento ttbjjb’ non è mai stato messo in evidenza finora– mancanza di un trigger per selezionare leptoni ad alta

efficienza– difficoltà di identificare i – Tuttavia, il BR è del 15%...

• E’ in realtà possibile selezionare eventi di top con la sola richiesta di significativa Et mancante nell’evento:– si selezionano eventi con soli jets adronici (veto e, ) con un

trigger multijet– si usa la (Et(miss))=Et(miss)/sqrt(Et) – si eliminano eventi in cui la missing Et è prossima ina un jet

adronico– in questo modo si raccolgono eventi di tipo ttbjjb’ ma anche

eventi in cui un elettrone o muone fallisce i tagli di identificazione• Si usano insomma gli scarti delle altre analisi!

Selezione eventi e stima fondi

• Eventi con >=4 jets, e missing Et significance>4, (j,Et(miss))>0.4

• >=1 b-tag• Il fondo viene stimato con eventi

senza b-tags usando una parametrizzazione della probabilità di fake– tiene conto delle caratteristiche dei

jets e della contaminazione da veri b-quarks di eventi con Et mancante puntante lungo il jet taggato

– verificata su diversi campioni di controllo

– buon accordo prima della selezione cinematica

– sistematica conservativamente stimata al 10%

RisultatiLa procedura permette di isolare un campione di 108eventi, contenente un segnale di oltre 60 eventi di top.La presenza del segnale è confermata dalla distribuzionedi variabili cinematiche discriminanti.

Si trova tt=5.8±1.1(stat) +1.7 -1.1(syst) pb

Diboson production• Diverse misure di produzione associata: WW,

Wg, Zg, WZ, ZZ… studiate da CDF e D0 con leptoni di alto Pt e fotoni

• Tests stringenti del MS: limiti a accoppiamenti anomali, e finestra su possibile nuova fisica

• CDF: (WW)=14.6+5.8-5.1

+1.8-3.0±0.9 pb

(th: 12.4 +-0.8)• D0: (WZ)=4.5+3.8

-2.6 pb (th:3.7±0.1pb)

• Con Et(>7 GeV, R>0.7(CDF): (W)B(Wl)=18.1±3.1 pb(th:19.3±1.4 pb);

(Z)B(Zll)=4.6±0.6 pb (th:4.5±0.3 pb).

• Con Et(>8 GeV, R>0.7 (D0): (W)B(Wl)= 14.8±2.3 pb

(th:16.0±0.4 pb); (Z)B(Zll)= 4.7±0.5 pb

(th: 3.9±0.2 pb).

CDF

D0

CDF

Improving the dijet mass resolution• One of the keys to a successful

extraction of the H signal is to increase the dijet mass resolution for pairs of b-jets

• Standard CDF/D0 jet correction algorithms tuned for best scale determination, not for best resolution

• H1 algorithm: use tracker to measure Pt of charged component

• in HSWG studied prototype of b-specific correction using identified muons, Et dependence of had response on top of H1

• Also developed advanced algorithm to account for subtle correlation among satellite observables and jet Et measurement

• Global result is that M/M=10% is achievable in central calorimeter

Study of V-A in the decay of quarks free from strong interactions

W decays and lepton Pt

Results compared

D0: upper limits to f+

D0 Results on HWWD0 also searches for HWW decays by selecting events with two oppositely charged leptons (ee,e: Pt>12,8 GeV; : Pt>20,10 GeV), missing Et>20 GeV (30 for ), and imposing a loose jet veto (Et<90 GeV, or Et

1,Et2 <50,30 GeV).

Combining the three channels they find 9 events, when 11.2±3.2 are expected from background sources in 177 pb-1 of Run 2 data.

They can thus exclude *B>5.7 pbat 95% C.L. for MH=160 GeV.

The azimuthal angle ll between the two leptons is then required to be less than 1.5 for electron pairs (2.0 for the e, combinations).

WHWWW(*) SearchCDF also searches for the striking signature of three W bosons in 193.5 pb-1 of Run 2 data.First, the dataset with a lepton with Pt>20 GeV and a second with Pt>6 GeV of same charge is analyzed and found in agreement with expectations.

Then, optimized cuts are applied to the second lepton (e.g. Pt>18 GeV for MH>160 GeV) and on the vector sum of leptons transverse momenta (Pt

ll>35 GeV).Zero events are observed, when 0.95±0.61± 0.18 are expected from known sources.95% CL limits are thus set at 12 (8) pb for MH=110 (160) GeV.

Low Mass H SearchesThe only chance to see Hbb at the Tevatron is through associated production with bosons

ZHllbb is the cleanest signature, but it yields too few eventsW/ZHjjbb has the lowest S/N but the high BR helps at larger Higgs massWHlbb is next-to-best, but CDF was “unlucky” in Run I The best channel is ZHbb

CDF has a new combination of Run 1 results with ZHllbb, bb channels. They search events with two jets with <2.6, missing Et>40 GeV, no isolated track with Pt>10 GeV. The limit is obtained by a fit to the mass distribution of b-tagged events.

The Run 1 CDF limit is now at 7.2 to 6.6 pb for MH=110 to 130 GeV.

Reconstructed masses

mjj mtreco

Green histos: data distributions

Curves: expected signal andbackground from global best fit

If this were the only Mtop result…Electroweak fit usingonly this result as topquark mass.

A full combination ofCDF/D0, Run I/II ismonths away…

Most direct implication: theHiggs boson mass moves back down into the excluded region, and away from the “theoretically loathed” M>135 GeV region……And back well into the LHC-problematic region!Also note: restored hope for MSSM-aficionados

Summary of ResultsNice measurements, although I can’t help feeling they sound more and moreas “cross-checks”…

2

2

2

2

2

2

2

2

,2

2,,

jets 4,2

2,,2

)()()()(

)()(

t

tb

t

tbjj

W

W

W

Wjj

yxj j

measUEj

fitUEj

i i

measiT

fitiT

MMMMMMMM

pppp

Event-by-event Mass Fitter• Distill all event information

into one number (called reconstructed mass).

• Select most probable jet-parton assgnmt based on 2, after requiring b-tagged jets assigned to b partons.

Reconstructedtop mass isfree parameter

ν

W+

W-

t

t

b-jet

b-jet

XConstraints

l

PT balance

mlν=mW

mjj=mW

mt1=mt2

Signal templatesSelected templates (GeV)

140 150 160

190180170

200 210 220

Parameterization:Build signal p.d.f. as a functionof generated mass.

Reconstructed Mass

Background template

Major contributions:

W+heavy flavor

W+jets (mistag)

QCD

Mass Template Source

Background Source # of events

W+jets (mistags) Mistags, QCD 4.4 ± 1.0

Wbb Wbb, Wcc, Wc, WW/WZ

2.1 ± 0.7

Single Top Single Top 0.33 ± 0.04

Total 6.8 ± 1.2

CDF Run II Preliminary (162 pb-1)

Constraint usedin likelihood fit.

Unbinned likelihood fit

• Free parameters are Mtop, ns, and nb.

– Profile likelihood: minimize w.r.t. ns,nb, no integration

• Fluctuations of nb are a systematic effect. Allowing nb to float in the fit means information in data is used to reduce the systematic uncertainty.

2

2exp

2

)(

bg

1

top)(shape

bgshape

)();()(

bn

bfitb

bs

nn

N

i bs

ibgbisigsNbs

nn

eL

nn

mPnMmPnnneL

LLL

bkgd (mean)constraint

InterestingParameter!

Templates: subdivide sample• Use 4 categories of

events with different background content and reconstructed mass shape.

• More b tags are better– Increases S:B– More “golden” events,

where correct jet-parton assignment is found.

Category 2-tag 1-tag(T) 1-tag(L) 0-tag

j1-j3 ET>15 ET>15 ET>15 ET>21

j4 ET>8 ET>15 15>ET>8 ET>21

S:B 18:1 4.2:1 1.2:1 0.9:1

Result with 318 pb-1

• Subdivision improves statistical uncertainty.– Pure and well

reconstructed events contribute more to result.

– Adds 0-tag events.• Subdivision does not

improve systematic uncertainty.– Most systematics,

including jet energy scale, are highly correlated among the samples.

2-tag 1-tag(T)

1-tag(L)0-tag

Expected Fraction of Sensitivity

2-tag 1-tag(T) 1-tag(L) 0-tag

35% 45% 9% 11%