HASCO 2017 Introducon to QCD and Jet Physics Vincenzo Cavasinni Universita’ di Pisa and INFN

Summary

Introducon to the QCD -The color interacon -Asymptoc freedom -The QCD elementary cross secons

Jets -Jet formaon and reconstrucon -jet reconstrucon algorithms

-jet energy calibraon Rc -jet cross secons measurement p p -jets for searching new physics R c Evidence for a new quark interaction and a new quantum number: the color e+e− → q( jet)q( jet) Ratio : e+e− → µ+µ−

Several other evidences for color eg : Γ(π 0 →γγ)

BR(τ → qq 'ν)

V. Cavasinni University of Pisa and INFN Interaction between colored quarks: the Quantum Chromo Dynamics, QCD is mediated by a massless vector particle: the gluon « ex of exchange of colored gluon. q q q qb r qb r qb r

rb br

qr qb qr qb qr qb q q b r qr qr

br rb rr

« We would expect 9 gluons: SU(3) algebra: 3 * 3= 8 + 1 octet: singlet: « But color is confined: and a singlet gluon would be not confined: « -> long range force…. V. Cavasinni University of Pisa and INFN Implicaons of colored gluons • Gluon can interact with itself

αs→ 2 (not the case for photons) αs → • Asymptoc freedom q q a) As in QED there are effect from vacuum polarization:

αS the result into a shielding of the carge 2 the αS decreases as Q →0 (screening)

In QCD there are additional gluon loops: q q Their effecs are ooposite to the vacuum polarization: b) q q The color charge is propagated (tube of charge). As a result : αS 2 αS increases as Q →0 (antiscreening)

q q The antiscreening effects dominate

V. Cavasinni University of Pisa and INFN " 2 2 % 2 2 αS (µ ) Q David Gross, αS (Q ) = αS (µ )$1− ( 33− 2 f )ln +...' # 12π µ 2 & Frank Wilczek and David Politzer, 1973

The vacuum polarization depends on the number of 2 2 gluon-loop effect Active quarks f (Q >4mq ) By summing up the leading log terms (geometric series: Σlnn ~ 1/(1-ln)): α (µ 2 ) 2 S QCD “running coupling constant” α S (Q ) = ⇒ α (µ 2 ) Q2 1+ S (33− 2 f )ln 12π µ 2 2 Q2→∞ αS (Q ) ###→0 (if 33 > 2 f) :The asymptotic freedom

Perturbation theory can be applied. But αS=0.1-1 and, differently than α, 2 2 2 2 It varies appreciably with Q . Defining ln Λ = lnµ −12π /[(33− 2 f )αS (µ )] We have only 1 parameter : Λ∼100-200 ΜeV ~ 1 fm-1 12π to be extracted experimentally (QCD has not α (Q 2 ) = S 2 2 a natural scale as in QED (me)). (33− 2 f )ln Q / Λ 2 2 Conversely if Q ~ Λ , αS is large,>1: interaction becomes strong,confinement?? V. Cavasinni University of Pisa and INFN From αS we can extract Λ, but with poor sensivity

100<Λ<400 MeV

In the same Q interval α varies

αS ~ 0.32 (Q=2 GeV)From 1/137 to 1/128 ~0.12 (Q=100 GeV)

+ − e e ( pp) → 3 jets q ex. + + − ∝αS e αS e e ( pp) → 2 jets e- q N.B. Λ depends on the active quark number q + e - e q Other processes to extract αS

V. Cavasinni University of Pisa and INFN The Standard Model Lagrangian

Gauge boson (gluon) self interacons

Gauge boson (gluon)-quark interacons

V. Cavasinni University of Pisa and INFN 8 processes QCD tree level cross sections

Combridge, Kripfganz Ranft (1977) t,s,u of the incoming partons (quarks or gluons)

t = −s(1− cosθ * ) / 2, u = −s(1+ cosθ * ) / 2 There is also the triple gluon vertex : ggg They are in the ratio 64/9 (color) Absent in QED Values at θ*=90o

2 d 2 σ παS |M|2 , averaged on initial color and spin * = M d cosθ 2s and summed up on the final ones

V. Cavasinni University of Pisa and INFN The elementary cross sections have to be weighted by the initial parton pdf’s

5 10 30 years of “jet physics” 103

101

10-1 LHC 13 Partons, after the scattering, evolve towards the 10-3 “parton showering” and hadronisation (jets) ≈10 7 10-5 LHC 13/Tevatron Tevatron 10-7 SppC

1 2 3 4 pT,jet(TeV) V. Cavasinni University of Pisa and INFN Jet physics

What jets are? QCD Interacons between quarks and gluons (proton constuents) yield quarks and gluons that, however, are not directly accessible (color confinement): in fact their fragment into a bunch of collimated hadrons (pion, kaons, neutros, protons,..): a “jet” of parcles. So idenfying and measuring jets, allows the reconstrucon of the kinemacs of the elementary QCD interacons.

Credits to: C.Roda – Lectures on “Jet Physics”, Pisa University, 2017. A dijet event A dijet event at the p-pbar collider at LHC

Noce that the two jets have to be balanced in Pt also in the lab, but not necessarily in longitudinal momentum (opposite rapidity) because the two primary interacng proton partons can have unbalanced momentum (pdf’s) V. Cavasinni University of Pisa and INFN Why do we care about jets: QCD studies Background of many other physics processes

Jets will be the background for nearly all searches:

σ(ET>100 GeV) ~ µb Signal processes: • SUSY involving heavy particles: σ(squark-squark @ M~ 1 TeV)~ pb • SM Higgs:

σ(Higgs @ M=125 GeV)~ 20 pb

QCD background is 5-6 order of magnitude higher than signals. Control of details of jet reconstruction √s (TeV) and calibration is essential.

V. Cavasinni University of Pisa and INFN Part 1 – jet formation and reconstruction

V. Cavasinni University of Pisa and INFN How a jet is generated

V. Cavasinni University of Pisa and INFN How a jet is generated Probability to emit gluon

Integral is over energy of Gluon and angle of gluon

Integral diverges if E or theta are small… QCD has to be renormalised.

V. Cavasinni University of Pisa and INFN In QCD quarks and gluons can radiate (αS) Probability to αS emit a gluon

gluon αS

Integral is over energy of Gluon and angle of gluon

Integral diverges if E or theta are small… QCD has to be renormalised.

V. Cavasinni University of Pisa and INFN Jet generaon Probability to emit a gluon

V. Cavasinni University of Pisa and INFN Jet generaon

Gluon can turn into a quark-antiquark pair

V. Cavasinni University of Pisa and INFN Jet generaon

V. Cavasinni University of Pisa and INFN Jet generaon and reconstrucon Detector

V. Cavasinni University of Pisa and INFN Jet generaon and recontsrucon Detector

Parton jets Truth jets Reconstructed jets V. Cavasinni University of Pisa and INFN Jet idenficaon

Transverse view of a collision e+e- ->hadrons (jets) the cleanest situaon

How many jets in this event: It is clear they are 2!

V. Cavasinni University of Pisa and INFN Jet idenficaon

Vista Transverse view of a collision trasversale di una collisionee+e- -> hadrons (jets) the e+ e- cleanest situaon: only jets in final state

How many jets in this event: It is clear they are 2!

V. Cavasinni University of Pisa and INFN Jet idenficaon

How many jets in this event: And in this one? It is clear they are 2!

V. Cavasinni University of Pisa and INFN Jet idenficaon

How many jets in this event: And in this one? It is clear they are 2!

V. Cavasinni University of Pisa and INFN Jet idenficaon

How many jets in this event: And in this one? It is clear they are 2!

V. Cavasinni University of Pisa and INFN It is necessay to implement a “jet algorithm” to identify jets in the event and consequently to measure their 4-momenta

V. Cavasinni University of Pisa and INFN Jet Algorithms

Jet algorithm: a set of rules that define how to ”cluster” jets from a collection of input signals from calorimeters and tracking

O(1k) calorimetric signals (clusters, cells) à reconstructed jets O(100) hadrons produced from fragmentation à truth jets O(10) à parton jets

The different jet types give different views of the same event and allow to compare measurements (reconstructed jets) to predicons (truth jets, parton jets) à comparison between predicons and measurements. V. Cavasinni University of Pisa and INFN Jet algorithm • The physics results (new particle discoveries, coupling measurements, PDFs…)should not depend on the particular jet algorithm used • Testing a measurement with different jet algorithms, or with different values of parameter for a jet algorithms is a very important test that allows to verify that the measurement is robust with respect to QCD radiations, hadronizations … • Exception: the sensitivity of a measurement may depend on the “resolution” with which the event is investigated …

A dependence that we should avoid …. à

V. Cavasinni University of Pisa and INFN Jet algorithms • Two cathegories:

• Cone: a cone is build with Lorentz invariant quantities: RC and particles are collected within this cone: Rc 2 2 p p RC ≈ Δφ + Δη

Rc • Sequential: start from single elements: track or, calorimetric cluster and try to recontruct back the branching effects due to the QCD showering and recombining the various elements with a “distance” algorithm between them.

V. Cavasinni University of Pisa and INFN The kt algorithm family • A family of sequential algorithms characterized by a specific “distance” definition used to join two tracks or calorimeter cluster The general definition is`:

ΔR2 d = min(k 2α ,k 2α )⋅ ij ti tij R2 2α kti = tranverse momentum of the element i diB = kti ktij = relative transverse momentum between elements i,j

α = 1 à Inclusive kt algorithm α = 0 à only angular distances – Cambridge-Achen algorithm € α = -1 à Antikt algorithm

V. Cavasinni University of Pisa and INFN Antikt • At present the most popular at LHC

i

-Recombination of collinear elements are privileged

-Large pT elements lead the recombination sequence -Produces regular shape jets

V. Cavasinni University of Pisa and INFN Examples

Example:

V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN The kt algorithm builds jets of irregular shape driven by the clusterization of soft element

anti-kt Thew anti-kt is driven by the highest pT elements and produces regular (cone) jets:

-Easier to manage experimentally - corrections can be linear

V. Cavasinni University of Pisa and INFN Jet reconstrucon and matching efficiency

V. Cavasinni University of Pisa and INFN Where the jet energy is? (ATLAS) A.Gupta

Central calorimeter region

2/3 of the jet energy is deposited in EM compartments

V. Cavasinni University of Pisa and INFN Where the jet energy is? (ATLAS) A.Gupta Central calorimeter region Only part of the energy

2/3 of the jet energy is deposited in EM compartments

V. Cavasinni University of Pisa and INFN Jet calibration

Correction of the detector effects to be considered to obtain the “true jet” : • uncompensation of calorimeters (e/h) • effect of cracks, dead material, losses in material in fron of the calorimeters • longitudinal leakage • magnetic field effect. Truths contain the

particles swept off from B field (pT<350 MeV do not reach the calorimeter entrances) • Pile-up contributions

V. Cavasinni University of Pisa and INFN Jet energy resolution Jet energy resolution affects the unfolding/evaluation of the jet distributions: systematics “unfolding” is the technique to reconstruct back from the measured jet distribution the “parton” distribution (QCD)

V. Cavasinni University of Pisa and INFN Jet Energy Scale Uncertainty

• This is NOT the resolution: it is a systematic error on the absolute jet energy • Many physics measurements are largely affected by th JES (jet cross sections, jet- jet resonances, , limits on new physics )

V. Cavasinni University of Pisa and INFN Jet Energy Scale Uncertainty

CMS ATLAS

Eur. Phys. J. C (2015) 75:17

2011 JINST 6 P11002

JES uncertainty ranges between 2 and 5 % for p > 20 GeV V. Cavasinni University of Pisa T and INFN Part III – Physics results

V. Cavasinni University of Pisa and INFN Why bother ?

Understanding/testing the QCD theory of SM in a new kinematic range never explored before (LHC)

Our level of understanding and modeling of the QCD interactions has direct impact on the potential we have for -QCD jets are often the largest precision measurements and backgrounds in many Higgs discovery decay channels eg: ! H bb; H (had) (had),.. → → τ τ -W/Z+jet is often one of the largest background to top-quark, V. Cavasinni University of Pisa SUSY, Higgs and exotic searches. and INFN The ingredients in the predictions…

V. Cavasinni University of Pisa and INFN The ingredients in the predictions… Hard Scattering and parton shower: fixed order NLO partonic calculations (NLOJET+ +,MCFM,BlackHat), LO+PS (Pythia, Herwig), High multi. LO+PS (Alpgen, Sherpa, Madgraph,…). State of the art: NLO+PS (MC@NLO+PS, POWHEG+PS, MEPS@NLO)

V. Cavasinni University of Pisa and INFN The ingredients in the predictions… Hard Scattering and parton shower: fixed order NLO partonic calculations (NLOJET+ +,MCFM,BlackHat), LO+PS (Pythia, Herwig), High multi. LO+PS (Alpgen, Sherpa, Madgraph,…). State of the art: NLO+PS (MC@NLO+PS, POWHEG+PS, MEPS@NLO)

PDF: CTEQ/CT, NNPDF, MSTW, ABM, HERAPDF…

Differences: data used the fit, αs value, treatment of errors, parameterization…

V. Cavasinni University of Pisa and INFN The ingredients in the predictions… Hard Scattering and parton shower: fixed order NLO partonic calculations (NLOJET+ +,MCFM,BlackHat), LO+PS (Pythia, Herwig), High multi. LO+PS (Alpgen, Sherpa, Madgraph,…). State of the art: NLO+PS (MC@NLO+PS, POWHEG+PS, MEPS@NLO)

Mostly anti-kT – various cone sizes PDF: CTEQ/CT, NNPDF, MSTW, ABM, HERAPDF… Measurements are unfolded at

Differences: data used the fit, αs value, particle level to be compared with treatment of errors, parameterization… predictions

V. Cavasinni University of Pisa and INFN The ingredients in the predictions… Hard Scattering and parton shower: fixed order NLO partonic calculations (NLOJET+ +,MCFM,BlackHat), LO+PS (Pythia, Herwig), High multi. LO+PS (Alpgen, Sherpa, Madgraph,…). State of the art: NLO+PS (MC@NLO+PS, POWHEG+PS, MEPS@NLO)

A sizeable contribution in understanding our measurements also comes from the quality of modeling of soft hadronic physics… no results in this talk but the measurements are used for Mostly AntikT – various the tuning of MC used. cone sizes PDF: CTEQ/CT, NNPDF, MSTW, AMB, HERAPDF… Measurements are unfolded at

Differences: data used the fit, αs value, particle level to be compared with treatment of errors, parameterization… predictions

V. Cavasinni University of Pisa and INFN Inclusive Jet cross-section

V. Cavasinni University of Pisa and INFN Each point is a measurement

V. Cavasinni University of Pisa and INFN N = Number of jets in the ΔETΔη range

ΔET

Jet Energy (TeV

ΔET (Δη, Δy)

V. Cavasinni University of Pisa and INFN ΔET

Jet detecon efficiency. ε∼ 95-100% at pT> 20GeV

V. Cavasinni University of Pisa and INFN STDM-2013-11 CMS-PAS-FSQ-12-031 CMS-PAS-SMP-12-012 Inclusive cross-secon @ 8 TeV

8 TeV 11 orders of magnitude Low pile-up data to extend to the low pT range down to 20 GeV and |y|<4.7

Antikt 0.7

20 Gev – 2 TeV

LHC data allows pQCD tests in a new kinematic regime – extended in pT and y Covers 11 orders of magnitudeV. / Cavasinni two jet sizesUniversity of Pisa and INFN V. Cavasinni University of Pisa and INFN Measurements uncertaines

V. Cavasinni University of Pisa and INFN Inclusive jet measurement at 13 TeV (preliminary)

Quark compositness parametrised as a current-current contact interaction ( a la Fermi) with a cutoff "Λ": present limits: Λ >~ 10TeV → substructure dimension d <~10−18cm

The existence of substructures of quarks (preons??) would show up as a discrepancy between the data and the QCD predicons (“Rutherford effect”) V. Cavasinni University of Pisa and INFN Jet energy scale uncertainty effect

V. Cavasinni University of Pisa and INFN Jet energy scale uncertainty effect

Uncertainty 5% on ET à 30% 30% error on cross section

V. Cavasinni University of Pisa and INFN Rao predicons/data

V. Cavasinni University of Pisa and INFN Rao predicons/data

• interval between the black linesà total uncertainty of data(statistical + systematic) • colored band widthàtheoretical uncertainties from various PDF parametrizations. • Agreement between theory and experiment within ~ 10-20% at pT<1 TeV V. Cavasinni University of Pisa and INFN Jets for new physics..

V. Cavasinni University of Pisa and INFN qq,gg,qg (jet-jet) resonance search

Ex: W’à qq’ à jet jet

Limit on CrossSectionxBRxA compared to model predictions

V. Cavasinni University of Pisa and INFN Resonance search in qq,gg,qg

Generic Limits on a new signal CrossSectionxBRxA compared with model expectations

V. Cavasinni University of Pisa and INFN Summary • Jets are the experimental manifestations of parton (quark, gluon) production; • Jets are complex objects and need a definition of an algorithm and a recombination scheme; • In order to compare predictions to experimental measurements a complex procedure of calibration needs to be applied; • Often the systematic uncertainty on the energy scale of jets is the largest source of experimental uncertainty on a measurement; • Skipped lots of important items: calorimeter, energy calibration, how to calculate systematics, jet substructure techniques (more in Gerald Eigen lecture)

Jets are complicated objects but they are fundamental components of the standard model and represent also an important tool to search for new physics!

V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN V. Cavasinni University of Pisa and INFN y*=1/2(y1-y2) In ATLAS |y*| <1.7 à χ<30

V. Cavasinni University of Pisa and INFN Compositness search with χ variable X is compositness sensitive especially at low values. X is less sensitive to the JES uncertainty.

Present limit on the compositness scale: Λ > 8.1TeV @ 95%C.L. V. Cavasinni University of Pisa and INFN Resonance search at very high pT

W’ à qq’

V. Cavasinni University of Pisa and INFN Resonance search at very high pT

W’ à qq’

In the last years new algorithms have been developed to extract substructures inside a large R V. Cavasinni University of Pisa and INFN Invariant mass from jet sub-structure

One example: Z à jj

Red: Zà jj Black: multi-jet bkg Pre grooming

Post grooming

Jet sub-structure tecniques are in rapid evoluon and allow to extend the energy reach for V. Cavasinni University of Pisa new physics studies. and INFN