Nobel symposium – Stockholm – 13-17 May 2013

Roberto Chierici CNRS (Lyon)

Unveiling the top secrets (with the LHC)

2

The top secretsWhy doing top physics is more exciting than ever.

Especially after the discovery of a neutral scalar boson

Selected(*) experimental resultsAnalyses at the LHC and the experimental challenge on top physics

(focus on) precision physics for constraining the Standard Model(some) searches for new physics coupled to top

Conclusions and outlookOpen questions in the exciting way forward

(*) Disclaimer: this seminar will not (can not !) give a complete review of results on top physics. The choice made is personal and, by definition, biased. This talk will mostly cover LHC results. Tevatron results are flashed when relevant. For the state of the art of experimental results please go here:

• https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsTOP• https://twiki.cern.ch/twiki/bin/view/AtlasPublic/TopPublicResults• http://www-cdf.fnal.gov/physics/new/top/top.html• http://www-d0.fnal.gov/Run2Physics/top/

Outline

3

The top secretsExperimental

challenges

INTRODUCTION

e μ τ bscdu tνe ντνμ WZH

4

• Unique because of its enormous mass: it weights like a tungsten atom ! Still a point-like particle in our understanding The top and the Higgs are “strongly” coupled The top mass dramatically affects the stability of the Higgs mass

o If we consider the SM valid up to a certain scale

A particle with unique characteristics

• It is the only quark that does not hadroniseo τ(had)~h/QCD~2 10-24 so τ(top)~h/top~5 10-25 so Compare with τ(b)~10-12 s Decays before forming a “dressed” top quarks No bound tq states, its spin properties are directly passed to its decay

products Flavour and EWK physics at their best !

2vtt ym 1ty

t W, Z H

mt2 ln(mH)

5

mW=mW(mt2,

log(mH))

• Can use the fact that mt, mW, mH are linked at loop level to constrain the SM

Constraining the SM

• The top also provide other direct constraints to the model Direct access to parameters of the SM (mt, Vtb) Other stringent tests of SM (QCD in d/dX, couplings, CPT invariance,

…)

Now known at NNLO QCD. Vacuum meta-stability when the minimum of V() is just local

The Higgs/symmetry breaking sector can be explored with more insights coming from top physics

arXiv:1205.6497

6

• Top physics is now essential to study the properties of the Higgs boson (and contribute to the measurement of the top Yukawa coupling)

• Top is also the best gateway to the searches for new physics: many new models involve the top sector Many may concern the top sector exclusively Top partners are present in SUSY, UED, little

Higgs, 4th generation models

• Final states from new physics involving top partners may be very similar to top-pair production, typically with the associated production of other particles→ Interesting to study total cross sections and

the “environment” of top production→ Precise measurements of differential cross

sections are important probes as well (charge asymmetries, spin structure, anomalous couplings)

• Do not forget that top physics may be an important background for searches !

Direct probe of EWSB and new physics

t~t~→tt+Xg~ g~→tt(tt)+Xt’t’→bbWW, ttZZb’b’→ttWWFour tops

(single) top production at the LHC

7

• Top is produced in pairs (QCD) or singly (EWK)• Single top EWK production happens via three main contributions

(7 TeV)~64 pb (7 TeV)~15.6 pb(7 TeV)~4.6 pb, ,

,

,

Vtb~1

• Backgrounds coming from W/Z+jets, top pair production, QCD

(NNLO+NNLL) scales PDFs [pb]

Czakon, Fiedler, Mitov (arXiv:1303.6254)

7 TeV 8 TeV

172.0+4.4-5.8 +4.7

-4.8 245.8+6.2-8.4 +6.2

-6.4

Top (pair) production at the LHC

8

,

,

,

• BR~10% • BR~44% • BR~46%

• Top pair QCD production happens mainly via gluon fusion

• Final states depend on the decay of the W bosons

• Backgrounds coming from W/Z+jets, single top (tW), QCD

…but still a needle in a haystack

9

process

Events/s8 TeV, peak

lumi

Events/y8 TeV, 25/fb

bb ~106 ~3 1012

Wℓ ~70 ~2.5 108

Zℓℓ ~6 ~25 106

tt ~1.5 ~6 106

Experimentally challenging

10

light jets (energy scale)

b-tagging e, μ, τ

Missing ET

• Top pair studies use all parts of HEP detectors… Charged lepton reconstruction Jet reconstruction Missing transverse energy

b-tagging

• Optimal use of the detectors… Particle Flow reconstruction in CMS

o Combine all sub-detector information to reconstruct and identify particles

Exploit excellent calorimetry in ATLAS• … and sophisticated analysis tools:

B-tagging, τ reconstruction, kinematic fitting

Muon resolution

MET resolution in Z→μμ events

ε(b-tag, data)/ε(b-tag, MC)

11

Production (diff.)cross-section

Top pair environment

W, t polarisations

Decay modes

~100%Spin, couplings,

mass, …

PDFs

Top pair cross sectionsSingle topTop propertiesLooking for new physics

EXPERIMENTAL RESULTS

Collected data

12

• Impressive performance of the LHC in 2011(@7TeV)/2012(@8 TeV) About ~6/fb collected in total at 7 TeV About ~23/fb collected at 8 TeV

2011

2010

2012• Statistics important for top

physics LHC is the first top factory ever !

o O(1M) tt @7TeV, O(10M) @8 TeV While precision measurements

soon limited by systematic errors, many possibilities for other studies open upo Rare processeso Searches for new physicso Constrain of systematic errors and

backgrounds by using data

• Results presented in this seminar: Mainly at 7TeV

o Not necessarily at full statistics A few 8TeV result

Example: how the single top signal improves

13

tt→qqbμνb

tt→eνbμνb

tW→μνbμν

tt→eνbqqb

Top pair (differential) production cross

sections

Total cross section measurements

14

• Monitoring the total production cross section is the first fundamental step for understanding top physics at the LHC Test the presence of new production mechanisms In the frame of the SM, test QCD predictions and help constraining the PDFs (especially gluons)

o Important for Higgs production, for instance

Indirect determination of mt or S. Constrain a very important background for many searches at the LHC

• Almost all decay modes are investigated at the LHC

• The measurements are performed at different level of complexity: Counting experiment in acceptance Fit to data in several portions of phase space

with in situ constraining of various backgrounds

Multivariate analyses

Top pair cross section• Di-lepton final states (e, μ) background

free Likelihood fits to the number of

reconstructed (b-tagged) jets. DY background data-driven

15

ATLAS-CONF-2012-131

CMS-PAS-TOP-11-003

CMS-PAS-TOP-11-004

• ℓ+jets final states represent a good compromise between statistics and purity Multidimensional ML fit to data Use data themselves

to constrain the backgrounds by including regions where they dominate

• Hadronic channels (all-jets, τ+jets) are very difficult Entirely dominated by

QCD, need to estimate it directly from data

Use NN to separate signals from backgrounds

Cross section combination• All top pair final states are (being) investigated

ℓ(e,μ,)+jets, (all but ℓℓ )+jets and fully hadronic final states in the combination.

Best measurements in the di-lepton channels

• Combinations performed taking into account correlations between errors Experimental uncertainty close to 5% Already challenging the newly available NNLO computations

o Current TH errors about 4%

16

Czakon, Fiedler, Mitov, arXiv:1303.6254

TOPLHCWG

Tevatron

Top pair differential cross sections

17

• Test top physics in different portions of the phase space Important test of pQCD, constrain of MC models and systematic effects, sensitive to new physics Use unfolding techniques on background-subtracted reconstructed

distributions for a direct comparison to theory predictions Propagation of the systematic errors (only shape errors important)

o Most relevant coming from background knowledge, radiation and hadronization

• Look at lepton, jets, but also at more complex variables involving top quarks Need a full reconstruction of top kinematics Compare to reference generators and predictions on differential

distribution from theory

m(tt), di-lepton

arXiv:1207.5644

pT(t), l+jetspT(ℓ), ℓ+jets

arXiv:1211:2220

Data start to challenge NLO predictions?

A special case: radiation in top pair

18

• At the LHC top quark are often produced with extra jets from initial (or final) state radiation Higher energy and high scale of the process Initial state preferentially from gluons (more

colour)

• Impact in the ability to reconstruct top pair ~half of the event with a jet with pT >50 GeV! Jet pairing may be difficult (see following) Systematic errors due to radiation description in

MC can be dominant

• Important to monitor and describe jet production Inclusive jet multiplicities, extra jet pTs, s,… Constrain generator parameters on radiation Aim also to look for new physics production in

tt+jets

pT(tt), ℓ+jetst

t

t

tg

Njets, ℓ+jets

ATLAS-CONF-2012-155

arXiv:1211:2220

Associated production of top and bosons

• tt+bb. Important also for SM physics Higgs in association to top. Top

Yukawa.

• Study N(b-jet) in di-lepton events

MadGraph: 1.2%; POWHEG: 1.3%

19

Trileptons Dileptons

• tt+W/Z are rare processes in the SM Monitor couplings between t and

Z Investigate top pair in association

with extra leptons

CMS-PAS-TOP-12-014

CMS-PAS-TOP-12-024

ATLAS-CONF-2011-153

• tt+ important for the direct measurement of the top charge Look for isolated photons Need more statistics, results with

5/fb at 7TeV in agreement with NLO predictions

• Not yet sensitive to ttH, both ATLAS and CMS ready for new data

Single top: why is that important?• The production cross section

gives direct access to the CKM matrix element |V|tb May also test the presence of

a possible 4th generation quark

Check for presence of FCNC Important background for

Higgs searches in associated production W/ZH→qqbb

20

• Investigate t-channel and tW production s-channel still out of range for an observation t-channel: 1 isolated e or μ, one b-tagged jet, one forward jet,

missing ET

tW channel: 2 isolated charged leptons (e, μ), one b-tagged jet, missing ET

• Main backgrounds from top-pair production, W+jets, QCD Use data whenever possible to constrain the backgrounds

Single top cross sections

21

CMS-PAS-TOP-12-011

ℓ>2

Phys. Lett. B716 (2012) 142

arXiv:1209.4533

• Typically use multivariate techniques (NN, BDT) Optimize S/B separation using full event

properties, constrain systematic effects by simultaneously analyzing signal and background dominated regions

• Cross sections in agreement with the SM expectations, |Vtb| can be derived by assuming

22

op properties

Constraining the SM with the top mass

23

• Remember: the top mass, the W mass and the Higgs mass depend on each other

Determination of the top mass

24

t

t

t

tb-tag

b-tag

• Indirect methods (most of them still in the works at the LHC) Use the dependence on the top mass on other variables

o Top pair cross sectiono Lepton pT and end-point methods o Invariant mass of the system J/Ψ+lepton from Wo Decay length of the b hadron

Main issue: need of a lot of statistics

• Direct reconstruction methods Full reconstruction by resolving the pairing ambiguities (all

channels studied) Use kinematic constrained fitting to improve the mass resolution

o Constrain the light jet energy scale in situ by using the W mass constraint Fit the mass with MC template fits or event by event likelihood fits

o Methods very sensitive to the description of radiation and JES uncertainties

Top mass direct reconstruction

• Plenty of statistics to reconstruct the resonance

• Top mass determination is calibrated by using MCs

• Dominated by systematic errors Dominant sources are JES

and TH uncertainties Improvements from fitting

JES in situ and constraining radiation from data

25

Eur. Phys. C72 (2012) 2046

arXiv:1209.2319

CMS-PAS-TOP-11-017

arXiv:1207.6758

CDF, ℓ+jets ATLAS, μ+jets CMS, hadronic

CMS, ℓ+jets

CMS

Top mass combinations• Work ongoing in the TOPLHCWG and

TEVEWWG to properly perform combinations Need to properly assess correlation of

systematic sources in between experiments

• In both di-lepton and +jets channels, ℓthe measurements at the LHC are now competitive with the corresponding ones at the Tevatron

→ After 2 years of data the LHC providesan error on the top mass similar to thatof the Tevatron after 20 years of run→ Both nail down the top mass at O(1GeV),

and agree

• Towards the first world combination with Tevatron and the LHC Work ongoing in the TOPLHCWG In contact with the TEVEWWG→ will need a detailed understanding of the

correlated systematics errors, especially modelling

26(*) Best Linear Unbiased Estimator[Lyons, Gibaud, Clifford; Nucl. Instr. Meth. A270 (1988), 16.]

TEVEWWG

TOPLHCWG

Spin structure of top decays

27

• The spin structure of the top decay is transmitted to its daughters By investigating the helicity of Ws from

top we can test the V-A structure of the coupling o The experimental “analyzers” are typically

the decay product of the Ws

• Measure d/dcosθ*ℓ, the angle between the lepton and the b direction (in the W rest frame)

JHEP 1206 (2012) 088

Constraining anomalous couplings• The polarization fractions can be

extracted by a fit to data Fit performed with and without the

assumption of FR=0 Main systematic errors represented by

JES and theory uncertainties/W+jets normalization

Agreement with the expectations in both ATLAS, CMS and combined results

28

0 in the SM

TOPLHCWG

• The helicity fractions can be translated into constraints of anomalous couplings and NP operators The LHC combination is consistent with the

expectation of the SM

TOPLHCWG

Top polarization and spin correlations

29

• While top quarks are produced individually unpolarized in top pair production… Can be studied via the angular distributions of

the leptons from W decay Fully leptonic final states particularly well

suited

• …the spin of the two tops are correlated Strength depending on the spin quantization

axis Can be measured from angular distributions

of the top decay products

o A: correlation strength at productiono i: amount of spin information from each probe

Δ between leptons particularly well suited variable

Sensitive to NP in both production and decay !

Phys. Rev. Lett. 108, 212001 (2012)

CMS-PAS-TOP-12-016

5 observationSensitivity ~25%

30

Looking for signs of physics

beyond the SM with top

Charge asymmetries• Tevatron observes anomalous charge

asymmetries

• May be an indication of new physics mechanisms in the production of top pair, both in s- or t-channel?

31

arXiv:1101.0034arXiv:0712.0851

• At the LHC it is possible to be conclusive on this, but the asymmetry needs to be defined differently Initial state charge symmetric

• In the SM the asymmetry is not exactly zero Asymmetry introduced by interferences

between ISR and FSR

Phys. Lett. B717 (2012) 129

CMS PAS-TOP-12-010

Differential asymmetries at the LHC• AC is determined from the background-subtracted distribution of |

yt|-|ytbar| Detector effects are unfolded

• In many new physics scenarios the charge asymmetry depends on phase space High mass/pT regimes enhance the quark annihilation part of the

initial state Measure Ac differentially as a function of pT, y or mass of the top pair

system• Good agreement between data and SM expectations within

uncertainties Results also compared with EFT predictions

o Anomalous axial coupling of gluons to quarks: capable to explain the Tevatron anomaly 32

EFT: PRD84:054017,2011NLO: Rodrigo, Kuehn

Phys. Lett. B717 (2012) 129 Eur. Phys. J. C72 (2012) 2039

Eur. Phys. J. C72 (2012) 2039

Phys. Lett. B717 (2012) 129

ATLAS

CMS

Di-leptons, 5/fb

ATLAS-CONF-2012-057

ℓ+jets, 1/fb

ℓ+jets, 5/fb

Di-leptons, 5/fb

Resonant top-pair production

33

• Several new physics models predict the presence of new heavy particles decaying into top-pairs

• Event’s kinematics strongly depend on the mass of the intermediate state Low mass analyses: standard tt

reconstruction High mass analyses: no lepton isolation, top

tagging techniques are requiredo Analyze substructure of jetso Fake rates directly determined via QCD data

QCD

tt1.3 TeV < mZ′ < 1.5 TeV

excluded ranges at 95% C.L.

mZ’(gKK) > 1.7(1.9) TeV

ℓ+jets

JHEP09 (2012) 029

hadronics

ATLAS-CONF-2012-136

34

Semi-leptonic candidate with m(tt)=1.6TeVmZ’(gKK) > 1.7(1.9) TeV

ATLAS-CONF-2012-136

arXiv:1209.4397

mZ’(gKK) > 1.5(1.8)TeV

Top-like searches

35

• Final states from top production, or top-like

• The studies presented so far are only a part of the rich program in ATLAS+CMS But no sign of new physics yet

mWR’>1.13 TeV

Excited W’

Z’ exchange

SUSY

FCNC

mt’ > 557 GeV

Conclusions

36

• Top physics is a pillar of the current research program in HEP Ideal probe for constraining (directly+indirectly) the symmetry

breaking of the SMo The top is way heavy → the Higgs scalar mostly couples to tops

Ideal probe for looking for new physics beyond the model itselfo Via precision measurements o Via direct searches for new signals

• Tevatron has handed the baton to the LHC From a handful of events to a deep testing of the sector We are now starting to challenge the theory predictions in many

respects

• Still the Standard Model looks healthier than ever No hints of new physics (yet?)

• Exciting times ahead of us… what should we do next?

Remember: the fate of the universe depends on the Higgs boson and the top quark !

Q1: what is the future of top physics at the LHC?

37

Inclusive quantities

Differentialcross sections

Top-pair “environment”

Evidence

• Top physics will remain the “Swiss knife” of LHC physics Privileged window for a direct observation of new physics Best handles for constraining existing physics

• What should the priorities be in the next years?

+direct searches total cross sections Wtb vertex structuretop propertiesconstrain systematics, PDFs

+direct searches

+direct/indirect searches

top propertiestop couplingsmeasure systematics

+direct/indirect searches

2010 20122011

a. Keep looking for NP in top(like) eventsb. Look for tails and “top pair environment”

Access to new physics in association, to top couplings to bosons, ttH Useful to constrain hard and soft QCD in situ → back to precision

measurementsb’. Fight systematic sources directly

Q2: which systematic sources to fight for precision?

38

• In the absence of direct evidence of new physics, precision measurements are more vibrant than ever All precision QCD/EWK measurements in top physics are dominated

(by far) by systematic uncertainties. How much (and where) can we gain in the future?

• Diversify analyses ! Exploit different (smaller) region of acceptance,

much less sensitive to traditional systematic error sources

Use different techniques with independent systematic sources and combine measurements. Always room for new ideas….

• We have the possibility to use the enormous amount of data we (will) have to directly constrain them. What are the best ways to go? Modelling uncertainties particularly worrying May expand measurements as a function of

observables Use data (top) themselves to constrain

systematic errorso Radiation, CR/soft QCD effects…

Q3: what is the future of top physics at accelerators?

39

• Very precision physics at a HL LHC (14 TeV, 3000/fb) Establish ttH and measure top couplings to

bosonso Couplings of top to Z/ can be improved to 2-10%

precision Searches in top in the focus of the research

programo Impressive improvements expected in direct

searches like FCNC, resonances,…o Final word on asymmetries and their dependence

on kinematics (e.g. m(tt))

300/fb

500/fb, 500 GeV

(in the hypothesis of still being in the realm of the SM)

• What will be left beyond the LHC, the case of ILC The keys are the EWK couplings of the top

o tZ/ couplings at better than 1%. Left and right handed polarized couplings are accessible via beam polarizations

o Access to (g-2)top to 0.1%→constrain compositeness to 100 TeV

Δyt/yt ~10%(500 GeV)-6%(1TeV) Top mass by threshold scans, extremely precise

access to the top quark pole mass