Marina Cobal Marina Cobal

Università di Udine

1

Physics at Hadron CollidersPart II

The structure of an event

2

One incoming parton from each of the protonsenters the hard process, where then a number ofoutgoing particles are produced. It is the nature

of this process that determines the maincharacteristics of the event.

Hard subprocess: described by matrix elements

An event: resonances

3

The hard process may produce a set of short-lived resonances, like the Z0/W± gauge bosons.

Resonances

4

•In this range the momentum scale is known at the permill level.• it is a cross-check of the detectorperformance in particular for the lepton energy measurements

The structure of an event: ISR

5

One shower initiator parton from each beam may start off a sequence of branchings, such as q → qg,

which build up an initial-state shower.

Initial state radiation: spacelike parton shower

The structure of an event: FSR

6

The outgoing partons may branch, just like the incoming did, to build up final-state showers.

Final state radiation: timelike parton showers

An event: Underlying events

• Proton remnants ( in most cases coloured! ) interact: Underlying event,consist of low pT objects.

• There are events without a hard collision ( dependent on pT cutoff)

An event: Underlying events

Underlying event:•Multi-parton interaction•Beam-beam remnants•Initial/final state radiation

Underlying Event

9

• Studying underlying event is crucial for understanding high pT SM events at LHC.

• ingredient for many analyses. In fact they affect: the jet reconstructions and lepton isolation, jet tagging etc..

• One can look at charged track multiplicitiesNch in transverse regions which are little

affected by the high pT objects.

• Reasonably described by models

The structure of an event: Pile up

In addition to the hard process considered above, further semi-hard interactions may occur between the partons of two other incoming hadrons.

‘Pile-up’ is distinct from ‘underlying events’ in that it describes events coming from additional proton-proton interactions, rather than additional

interactions originating from the same proton collision.

Pile up

11

2012 ATLAS event; Z in with 25 primary vertices

Z in eventwith 25 vertices

• Multiple interactions between partons in other protons in the same bunch crossing– Consequence of high

rate (luminosity) and high proton-proton total cross-section (~75 mb)

• Statistically independent of hard scattering– Similar models used

for soft physics as in underlying event

Et ~ 58 GeV

Et ~ 81 GeV

without pile-up

Prog.Part.Nucl.Phys.60:484-551,2008

Pile up

Et ~ 58 GeV

Et ~ 81 GeV

with design luminosity pile-up

Prog.Part.Nucl.Phys.60:484-551,2008

Pile up• Multiple interactions

between partons in other protons in the same bunch crossing– Consequence of high

rate (luminosity) and high proton-proton total cross-section (~75 mb)

• Statistically independent of hard scattering– Similar models used

for soft physics as in underlying event

Challenge Pile up: example ETmiss

14

• Requirements on track vertexing

• Number of reconstructed vertices proportional to the pile-up

• Measure pile-up density event by event: Use it to subtract from the jets energy a pile-up term. do the same with isolation cones.

without PU suppression

with PU suppression

Important for quantities, affected by soft hadrons, for example;

ETmiss = -| Σ pT |

Use data!

• Inelastic hadron-hadron events selected with an experiment’s “minimum bias trigger”.

• Usually associated with inelastic non-single-diffractive events (e.g. UA5, E735, CDF … ATLAS?)

Minimum bias events

Need minimum bias data if want to:1) Study general characteristics

of proton-proton interactions2) Investigate multi-parton

interactions and the structure of the proton etc.

3) Understand the underlying event: impact on physics analyses?

In parton-parton scattering, the UE is usually defined to be everything except the two outgoing hard scattered jets: Beam-beam remnants.1) Additional parton-parton

interactions.2) ISR + FSR

Can we use “minimum bias” data to model the “underlying event”? At least for the beam-beam

remnant and multiple interactions?

The underlying event

The “soft part” associated with hard scatters

Minimum bias

16

• Non head-on collisions, with only low pT objects. Those are the majority of the events in which there is a small momentum transfer

Δp ~ h/Δx

• Distributed uniformly in η: dN/d = 6• On average the charged particles in the final

state have a pT~500 MeV

Not well described bymodels!Shape is sort of OKNormalisation is off

Minimum bias

17

• It is interesting by its own to study such events. Also an ingredient for many analyses you will see.

• A necessary first step for precision measurements (such as top-quark mass)

• A key ingredient to modelling pile-up

• As can be seen most of the events do have quite low pT

• Anyhow those events constitute a noise of few GeV per bunch crossing

18

Monte Carlo Simulations• Attempt to simulate all physics

and experimental aspects as well as possible in MC

• Examples shown here: – Pile-up– Jet response– Electron acceptance on

detector level– Corrections from quark to

jets

• Use data ('data-driven' techniques) to verify that MC is correct w.r.t all relevant aspects

• Apply corrections (a.k.a. scale factors) to MC where necessary

19

Monte Carlo Simulations• MC contains two aspects

– description of detector response → efficiency, resolutions– description of shapes (physics model) → acceptance

• This allows to translate the cross section measurement into a determination of a correction:

N.B. assuming good description of efficiency and acceptance by MC – uncertainty ?

Monte Carlo for Processes with jets

Parton shower

MC simulation of LHC event

Hard partonic scattering

Incoming parton distributions

QCD and QED radiation

HadronisationParticles

Additional partonic scatters

Detector simulation

A Monte Carlo Event

Initial and Final State parton showers resum the large QCD logs.

Hard Perturbative scattering:

Usually calculated at leading order in QCD, electroweak theory or some BSM model.

Perturbative Decays calculated in QCD, EW or some BSM theory.

Multiple perturbative scattering.

Non-perturbative modelling of the hadronization process.

Modelling of the soft underlying event

Finally the unstable hadrons are decayed.

24

Uncertainties

• Statistical uncertainties, due to finite number of events• Systematic uncertainties, due to errors and biases in the analysis

• Simplest, most-often-used approach: assume that systematic errors are mutually independent, i.e. uncorrelated– make list of all sources of systematic uncertainties– remove those that are correlated with others– repeat analysis for variation of each uncertainty separately– add variations up in quadrature

• More complex treatment of systematics not addressed today • Most analysis work goes into dedicated studies aiming to minimize the systematic

uncertainty

Table of uncertainties

Example: CMS top pair production in di-lepton channel

Experimentalaspects

Theoryuncertainties

backgrounds

SM processes

26

• No hope to observe light objects ( W,Z,H) in the fully hadronic final state!• We need to rely on the presence of an isolated lepton!

• Fully hadronic final states can be extracted from the backgrounds only with hardO(100 GeV) pT cuts-> works for heavy objects!

QCD Sector

Snapshot of QCD

QCD vertices

Colour factors

QCD Potential

Jets from quarks and gluons• Quarks and gluons cannot exist as free particles ->

hadronization• Collimated stream of charged and neutral hadrons -> QCD jets

Where do Jets come from at LHC?

1.8 TeVs

14 TeVs

inclusive jet cross-section

• Fragmentation of gluons and (light) quarks in QCD scattering

• Most often observed interaction at LHC

Multi-jet events at LHC

Jet multiplicity

35

• Another possible test of QCD isobtained by checking the jet

multiplicity

• Tests also the modelling of theradiation

Where do Jets come from at LHC?• Decay of heavy Standard

Model (SM) particles

Prominent example:

qq q q WW Hjj qq q q WW Hjj

top mass reconstruction

Where do Jets come from at LHC? Associated with particle production in Vector Boson Fusion (VBF) E.g., Higgs

Where do Jets come from at LHC?

•Decay of Beyond Standard Model (BSM) particles

– E.g., SUSY

electrons or muons jets

missing transverse

energy

,jets

,leptons

Te f Tjf T ppM p

What is a jet?

How to identify jets?Jet algorithm should collect all particles in the same way for:

•Leading order partons•Partons+gluon emission•Parton shower (soft)•Hadrons-> detector

Jets• Definition (experimental point of view):

bunch of particles generated by hadronisation of a common confined source– Quark, gluon fragmentation

• Signature– Energy deposit in EM and HAD

calorimeters – Several tracks in the inner detector

41

• Calorimeter energy measurement

- Gets more precise with increasing particle energy

- Gives good energy measure for all particles except ’s and ’s

-Does not work well for low energies

-Particles have to reach calorimeter, noise in readout

jet algorithms

Jet Reconstruction Task

Jet Reconstruction

• How to reconstruct the jet?

– Group together the particles from hadronization

– 2 main types

• Cone

• kT

44

Jet reconstruction algorithms: cone

Jet reconstruction algorithms: Kt

Di-jet quark flavours

arXiv:1210.0441v3

Jet physics: jet energy scaleBefore looking at jet physics be aware of few issues, first of all when we have steeplyfalling cross sections-> we have a sensitivity of its measurement from the energy scale-Jet energy determined from calorimeter(+tracking information)-Sophisticated calibration procedure

Different contributions to JES error.(jets reconstructed with the Anti-kTalogrithm cone 0.6 that isused in ATLAS)

Jet production

50

• NLO QCD works over ~9 orders of magnitude!• excellent exp. progress: jet energy scale

uncertainties at the 1-2% level• for central rapidities: similar exp. and theo.

uncertainties, 5 - 10%• inclusive jet data : starts to be important tool for

constraining PDFs, eg.also by using ratios at different c.o.m. energies