Jet Analysis in HeavyIon Collisions Elena Bruna INFN

Jet Analysis in Heavy-Ion Collisions Elena Bruna INFN Torino & Yale University 5 th International School on QGP. Torino, March 2011

“The same thrill, the same awe and mystery, come again and again when we look at any problem deeply enough. With more knowledge comes deeper, more wonderful mystery, luring one on to penetrate deeper still. With pleasure and confidence we turn over each stone to find unimagined strangeness. ” R. Feynman These (experimental) lectures won’t probably tell you everything you would ever wanted to know about jets…but I hope some of the young minds will be inspired and start/continue working on hard probes to turn over more stones… 5 th International School on QGP. Torino, March 2011

Past, Present, Future… Pb+Pb @ √s. NN=2. 76 Te. V LHC RHIC LEP e+ e- • Full jet reconstruction energy of the hard scattering but challenging in A+A • New jet-finding tools • Do physics with jets ! Elena Bruna (Yale&INFN Torino) 3

Outline Jet I: Intro & Motivations Jet II: Full Jet Reconstruction Jet III: Results Jet IV: The Present: from RHIC to LHC Elena Bruna (Yale&INFN Torino)

Jet I: Intro & Motivations

Jets in high-energy collisions p. QCD Factorization: Collins, Soper, Sterman Nucl. Phys. B 263 (1986) 37 PDF c, x c Partonic x-section a, xa σab d, x d p Elena Bruna (Yale&INFN Torino) b, xb p Fragmentation function Factorization: assumed between the perturbative hard part and the universal non-perturbative fragmentation (FF) and parton distribution functions (PDF) Universality: fragmentation functions and parton distribution functions are universal (i. e. FF from ee, PDF from ep, use for pp) 6

Jets in high-energy collisions p. QCD Factorization: Collins, Soper, Sterman Nucl. Phys. B 263 (1986) 37 PDF Partonic x-section Fragmentation function QCD factorization works! p + p p 0 p+p √s=200 Ge. V p + p p p+p √s=200 Ge. V Elena Bruna (Yale&INFN Torino) 7

Jets in high-energy collisions p. QCD Factorization: Collins, Soper, Sterman Nucl. Phys. B 263 (1986) 37 PDF Partonic x-section Fragmentation function c, x c PDFs: Probability for a parton a(b) to carry a fraction xa(xb) of the hadron momentum a, xa σab d, x d p b, xb p Universal can be measured with fit to experimantal data for one or more processes that can be calculated with perturbative QCD, i. e. deep inelastic scattering DIS (like e-p), Drell-Yan processes (qq l+l-) and others Many PDFs on the market (CTEQ, GRV, MRST, …) Elena Bruna (Yale&INFN Torino) 8

Jets in high-energy collisions p. QCD Factorization: Collins, Soper, Sterman Nucl. Phys. B 263 (1986) 37 PDF Partonic x-section Fragmentation function Hard scattering: dσ/dt = parton cross section calculable in powers of αS LO NLO Elena Bruna (Yale&INFN Torino) 9

Jets in high-energy collisions p. QCD Factorization: Collins, Soper, Sterman Nucl. Phys. B 263 (1986) 37 PDF Partonic x-section Fragmentation function Fragmentation Functions: probability to find, at scale Q, a hadron h with a fraction z of the parton c momentum universal and measured with fits to experimental data Many D on the market (KKP, AKK, …) p(hadron) p (parton) pa rto n z= z Elena Bruna (Yale&INFN Torino) 10

Jets in Nucleus-Nucleus collisions Detector Jet Tomography! Hard processes make perturbative QCD applicable high momentum transfer Q 2 Self-generated “hard” probes Hard processes scale as Nbin Calibrated LASER/x-ray Elena Bruna (Yale&INFN Torino) 11

Jets in Nucleus-Nucleus collisions jet energy loss in the medium Questions: 1) How does the parton lose energy? 2) What happens to the radiated energy? 3) Collisional energy loss? 4) Does the energy loss depend on the parton type? Interpretation: Gluon radiation DEloss ~ ρgluon (gluon density) DEloss ~ ΔL 2 (medium length) [~ ΔL with expansion] DEgluon > DEquark, m=0 > DEquark, m>0 Important to measure DE of gluons light heavy quarks… ^ = m 2 / L is the <p 2> transferred from the parton to a Transport coefficient: q T gluon per unit path length Elena Bruna (Yale&INFN Torino) 12

Jets in Nucleus-Nucleus collisions Eskola, Honkanen, Salgado, Wiedemann Nucl Phys A 747 (2005) 511 q^ = 5 – 15 Ge. V 2 / fm Elena Bruna (Yale&INFN Torino) from RHIC RAA Data 13

Some Predictions: FF Gyulassy et al. , nucl-th/0302077 Renk, Phys. Rev. C 79: 054906, 2009 Borghini and Wiedemann, hep-ph/0506218 z=ph/pjet ph pjet ξ stretches the low z part Elena Bruna (Yale&INFN Torino) Energy loss in the medium softer fragmentation 14

Some Predictions: Jet shapes If energy loss by gluon radiation broadening of the jet energy profile R = jet radius (on η-ϕ plane) =√(Δϕ 2+Δη 2) ωmin (p. Tmin) = minimum p. T on particles in the jet Energy loss ratio goes down with larger b. Energy loss ratio becomes smaller with smaller R and larger ωmin. Limit of large R and ωmin=0 no out-ofcone energy ΔEin~E I Vitev, S Wicks, B-W Zhang, JHEP 0811, 093 (2008); EPJC 62, 139 (2009). Elena Bruna (Yale&INFN Torino) 15

Some Predictions: Jet shapes I Vitev, S Wicks, B-W Zhang, JHEP 0811, 093 (2008); EPJC 62, 139 (2009). Vitev, Zhang, PRL 104 (2010) 132001, ar. Xiv: 0910. 1090 Limits: • small Rmax and large ωmin single particle suppression. • large Rmax and small ωmin all jet energy recovered RAAjet=1 ! (jet production is hard process, scales as Nbin) Elena Bruna (Yale&INFN Torino) 16

Some Predictions: Jet shapes I Vitev, S Wicks, B-W Zhang, JHEP 0811, 093 (2008); EPJC 62, 139 (2009). Vitev, Zhang, PRL 104 (2010) 132001, ar. Xiv: 0910. 1090 Limits: • small Rmax and large ωmin single particle suppression. • large Rmax and small ωmin all jet energy recovered RAAjet=1 ! (jet production is hard process, scales as Nbin) Elena Bruna (Yale&INFN Torino) 17

Jet quenching from single high-p. T hadrons Observations at RHIC: 1. Large suppression of high-p. T hadrons: factor ~ 5 2. Photons are not suppressed • They don’t interact with the medium (good!) • Nbin scaling works Elena Bruna (Yale&INFN Torino) 18

Jet quenching from single high-p. T hadrons Observations at RHIC: 1. Large suppression of high-p. T hadrons: factor ~ 5 2. Photons are not suppressed • They don’t interact with the medium (good!) • Nbin scaling works 3. Also Heavy Flavor is suppressed at RHIC • same as light quarks • role of bottom? • collisional energy loss/resonant elastic scattering? Elena Bruna (Yale&INFN Torino) 19

Jet quenching from single high-p. T hadrons ALICE, Phys. Lett. B 696 (2011) 30. RHIC suppression < LHC RHIC: high p. T hadrons hadronize from quarks LHC: from gluons (larger color charge!) Prediction: Vitev(hep-ph/050322 v 1) • GLV – p. QCD factorization • medium-induced gluon brems. Elena Bruna (Yale&INFN Torino) 20

Jet quenching from di-hadrons Start from a high-pt “trigger” particle and look on the away side (in f). Azimuthal correlation function shows ~complete absence of “away-side” jet Partner in hard scatter is absorbed in the dense medium not the case in d+Au final state effect Elena Bruna (Yale&INFN Torino) 21

Jet quenching from di-hadrons Start from a high-pt “trigger” particle make azimuthal correlation ~complete absence of “away-side” jet Partner in hard scatter is strongly interacting with the dense medium not the case in d+Au final state effect ! Path-length dependence of di-jet topologies Out-of-plane y in-plane x Back-to-back suppression out-of-plane stronger than in-plane Elena Bruna (Yale&INFN Torino) 22

Jet quenching from di-hadrons increasing p. Ttrig increasing p. Tassoc STAR, Phys. Rev. C 82 024912 (2010) Elena Bruna (Yale&INFN Torino) 23

Jet quenching from di-hadrons increasing p. Ttrig At low trigger p. T & low p. Tassoc: • Mach Cone – conical emission? • Cherenkov Radiation? • pure 3 D hydro? [won’t discuss this] At high trigger p. T: increasing p. Tassoc double bump: Phys. Rev. C 82 024912 (2010) • re-emergence. STAR, of away-side jet (punch thru)? or • tangential jets? Elena Bruna (Yale&INFN Torino) 24

�Trigger and Surface Biases Experiments online-trigger dependent: • Large p. T or energy deposition triggers bias towards hard fragmentation! • EM calorimetry bias towards large EM fraction Trigger particles biased toward the surface Surface bias, as seen in hydro models Elena Bruna (Yale&INFN Torino) 25

High p. T: towards jets ALICE, Phys. Lett. B 696 (2011) 30. What we have so far: • Suppression of high-p. T hadrons in A+A (at RHIC and LHC) w. r. t. p+p • Evidence for parton energy loss in the medium But: • Geometrical bias: dominated by surface jets • Jet energy not constrained • Limited kinematic reach Renk and Eskola, hepph/0610059 What we want: • Precise measurement of the parton energy loss • Measurement of the modified fragmentation function How? Elena Bruna (Yale&INFN Torino) 26

High p. T: towards jets How? 1. g-Jet • Jet energy well constrained • limited kinematic reach (x-sec scales as αSαem) • Difficult to have a clean measurement of photons Di-Hadron z = p(h)/pparton Eγ = pparton p ≠ pparton Leading Hadron Elena Bruna (Yale&INFN Torino) Courtesy Thomas Ullrich 27

Direct g-hadron fragmentation functions STAR, Phys. Rev. C 82 (2010) 34909 1. Good agreement w/ theory models 2. more assoc h± for p 0 than for g different parton energies for p 0 and g (p 0 come from fragmentation of higher energy parton) 1. Au+Au: different path-length for the recoil jet for p 0 and g and triggers Elena Bruna (Yale&INFN Torino) Trig particle= g or p 0 Assoc particle: h± 28

Direct g-hadron fragmentation functions IAA= ratio of associated yield per trigger in Au+Au to that in p+p Trig particle= g or p 0 Assoc particle: h± 8<Etrig<16 Ge. V/c 1. IAA < 1 for z. T>0. 3 2. data can distinguish between different theoretical models 3. low z. T: expected differences between p 0 and g IAA due to path-length dependence of the energy loss Measurements do not indicate path-length or color-charge dependence ! Elena Bruna (Yale&INFN Torino) 29

High p. T: towards jets How? g-Jet • Jet energy well constrained • limited kinematic reach (x-sec scales as αSαem) • Difficult to have a clean measurement of photons 2. Full Jet Reconstruction • Larger kinematic reach • large background complex and challenging ! 1. Ejet = pparton Eγ = pparton z = p(h)/pparton Courtesy Thomas Ullrich Elena Bruna (Yale&INFN Torino) 30

Jet II: Full Jet Reconstruction Elena Bruna (Yale&INFN Torino)

Jets: Theory vs Experiment Theory (p. QCD): jet = High-p. T parton produced in hard scatterings, or the closest object to a parton Experiment: jet = spray of collimated hadrons GOAL: measure the parton energy in experiments do jet physics! Tool: Full jet reconstruction with jet-finding algorithms • for both Theory and Experiment ! Elena Bruna (Yale&INFN Torino) 32

p. T Theoretical requirements p. T cone iteration Collinear safety � � � � � Jet 1 y y Jet 1 replaces one parton by two at the same place the algorithm should be insensitive to any collinear radiation. Jet 2 Infrared safety a soft emissions that add very soft gluon the jet-finding algorithm should not be sensitive to soft radiation Elena Bruna (Yale&INFN Torino) 33

Experimental requirements • Detector independence: the performance of the jet algorithm should not be dependent on detector segmentation, energy resolution, … • Stability with luminosity: jet finding should not be strongly affected by multiple hard scatterings at high beam luminosities. • Fast • Efficient: the jet algorithm should find as many physically interesting jets as possible, with good energy resolution Elena Bruna (Yale&INFN Torino) CDF 34

Jet Finding algorithms Review of CDF Jet Algorithms, ar. Xiv: hep-ex/0005012 v 2 Fast. Jet JHEP 0804: 005, ar. Xiv: 0802. 1188 Fast. Jet JHEP 0804, 063 (2008), ar. Xiv: 0802. 1189 v 2 Particles are combined into jets • the larger experimental coverage, the better Which particles? The measured ones: • charged tracks (TPC) • neutral towers (EMC) • charged energy (Hcal) Different ways of combining particles jet-finding algorithms Sequential Recombination Cone k. T CDF Jet. Clu, Mid. Point Anti-k. T D 0 Cone Cambridge - Aachen CMS Iterative Cone ATLAS Cone Py. Cell SISCone Elena Bruna (Yale&INFN Torino) 35

Sequential Recombination k. Ti, j= particle transverse momentum (p. T) k. T: p>0 (soft particles merged first) Anti-k. T: p<0 (hard particles merged first) R=resolution parameter Example: Anti-k. T Blue = highest p. T particle 2π If dij<k. Ti-2 merged ϕ 0 -1 Elena Bruna (Yale&INFN Torino) η +1 If dij>k. Ti-2 not merged call it a jet 36

k. T vs anti-k. T Fast. Jet M. Cacciari, G. Salam, G. Soyez 0802. 1188 • ALL particles are clustered into “jets” • k. T not bound to a circular structure • Anti-k. T circular shape, “cone” radius ~R parameter – Expected to be less sensitive to background/“back reaction” (it starts from high-p. T particles) ideal choice in heavy-ion collision • Recombination algorithms are collinear and infrared safe Elena Bruna (Yale&INFN Torino) 37

R matters! Elena Bruna (Yale&INFN Torino) 38

R matters! Elena Bruna (Yale&INFN Torino) 39

R matters! Elena Bruna (Yale&INFN Torino) 40

R matters! Elena Bruna (Yale&INFN Torino) 41

R matters! The choice of R depends on • The system we are looking at (e+e-, pp, Au. Au, Pb. Pb, …) • Tradeoff: don’t want to loose too much out-of-cone radiation (corrections for hadronization become difficult) but want to have a small background in the jet area In pp: ~80% of jet energy within R=0. 4 for 20 Ge. V jets Elena Bruna (Yale&INFN Torino) p+p 200 Ge. V STAR Preliminary 42

Jets in Heavy-Ion Collisions at RHIC and LHC Central Au+Au √s. NN=200 Ge. V STAR EMC + tracking data ETjet ~ 21 Ge. V Central Pb+Pb√s. NN=2. 76 Te. V ALICE tracking data STAR preliminary Why measure jets in heavy ion collisions? [inclusive, di-jets, jet-hadron, g-jet, . . ] • Access kinematics of the binary hard-scattering • Characterize the parton energy loss in the hot QCD medium − modified fragmentation, energy flow within jets, quark vs gluon jet difference − flavor and mass dependence • Study medium response to parton energy loss – establish properties of the medium Elena Bruna (Yale&INFN Torino) 43