Review of final-state structure of pp interactions at the LHC Paolo Gunnellini

Review of final-state structure of pp interactions at the LHC Paolo Gunnellini DESY (Deutsches Elektronen Synchrotron), Hamburg QCD at cosmic energies Kalchida (Greece) May 6 Paolo Gunnellini QCD at cosmic energies May 6

Outline Final states at the LHC Generalities on Monte Carlo event generators Generator tuning and machinery Available tunes Our understanding of QCD exp. data Current issues and incompatibilities Comparison between predictions and measurements Diffractive selections Hard QCD events and multijet scenarios Vector-boson final states op-antitop final states 5 Universality of tunes? 6 Summary and conclusions DISCLAIMER: Mainly results from ALAS and CMS! Paolo Gunnellini QCD at cosmic energies May 6

Hard scattering at the LHC o simulate proton-proton collisions, physicists generally use Monte Carlo event generators Ingredients of the hard scattering Factorization theorem Parton distribution functions Matrix element calculation..starting from lowest order in α S but might include additional real and virtual corrections σ ab F (Q ) = dx dx f a (x, Q ) f b (x, Q ) ˆσ ab (x, x, Q ) otal Cross Section Partonic Cross Section Parton Distribution Functions with x = longitudinal proton momentum fraction carried by the parton Q = scale of the scattering Paolo Gunnellini QCD at cosmic energies May 6

he underlying event at the LHC Paolo Gunnellini QCD at cosmic energies May 6

Final states at the LHC What can be produced in proton-proton collisions? From soft to hard.. Soft particles Jets Heavy flav. jets Vector boson op quark Vector boson pair Higgs... Standard Model otal Production Cross Section Measurements Status: Nov 5 σ [pb] 8 µb ALAS Preliminary heory Run, s = 7, 8, ev LHC pp s = 7 ev 6 Data.5.9 fb LHC pp s = 8 ev 5 5 pb Data. fb 5 pb LHC pp s = ev Data 85 pb. fb total. fb VBF VH t th pp W Z t t tt chan WW H Wt WZ ZZ ts chan t tw t tz Huge collection of measurements about these final states..in this talk only some of them with focus on our understanding of the experimental results! Paolo Gunnellini QCD at cosmic energies May 6 5

MC generator types Pure matrix element (ME) simulation MC integration of cross section + PDF No hadronization nor UE Useful for theoretical studies, no exclusive events generated Event generators Combination of ME and parton showers Generator for fixed-order ME combined with leading log (LL) parton shower MC Exclusive events, useful for experimentalists Paolo Gunnellini QCD at cosmic energies May 6 6

MC generator types Paolo Gunnellini QCD at cosmic energies May 6 7

MC generator types Paolo Gunnellini QCD at cosmic energies May 6 8

MC generator types Paolo Gunnellini QCD at cosmic energies May 6 9

Details of the Monte Carlo generators Example: multijet final state PYHIA8 and HERWIG++: LO MC generators with extra jets from PS & MPI SHERPA, MADGRAPH: matrix element with N-jets (extra real emission) POWHEG: matrix element with a hard emission @ NLO (real & virtual) PYHIA, HERWIG SHERPA MADGRAPH@LO POWHEG Paolo Gunnellini QCD at cosmic energies May 6

he Underlying Event at the LHC [mb] σ eff CMS (W + jets) ALAS (W + jets) 5 CDF ( jets) CDF (γ + jets) Corrected CDF (γ + jets) D (γ + jets) UA ( jets - lower limit) 5 AFS ( jets - no errors given) 5 5 Hard scattering Initial and Final State Radiation Multiple Parton Interactions (MPI) Beam-beam remnants Hadronization... 5 s [ev] In general, the UE is a softer contribution but.. some MPI can be hard! Double Parton Scattering σ DPS AB σ DPS AB = m σ eff P A = σ A σ pp tot P B = σ B σ pp tot m P AP B σ pp tot σ pp tot σ A σ B σ eff Need for correlations! Paolo Gunnellini QCD at cosmic energies May 6

How do we deal with that? Montecarlo event generators (PYHIA, HERWIG, SHERPA..) Parameters need to be adjusted (tuned) to describe data MPI Primordial k Parton shower Hadronization e.g. p = pref (E/E ref ) ɛ Proton matter distribution profile Colour reconnection e.g. Width of the gaussian used for modelling the parton primordial k inside the proton e.g. Strong coupling value Regularization cut-off Upper scale e.g. Length of fragmentation strings Strange baryon suppression How does one tune all these? Choice of parameter ranges and sensitive observables Predictions for different parameter choices and interpolation of the MC response Data-MC difference and minimisation over parameter space Paolo Gunnellini QCD at cosmic energies May 6

Some official tunes from the authors.. PYHIA 8 HERWIG++ Monash une - PDF: NNPDF.LO (EPJ C7 () 8) UE-EE-5C - PDF: CEQ6L (JHEP () ) PYHIA 8 Monash HERWIG++ UE-EE-5C (soft) MPI UE pp( p) data at various s UE pp( p) data at various s Value of measured σ eff Primordial k p spectrum of lepton pair p spectrum of Z boson from Z decays in hadronic collisions in hadronic collisions Parton shower Event shapes in p p interactions Jet multiplicity, jet rates and (taken from previous tune) shapes at various colliders Hadronization Particle multiplicities in hadronic Particle production at Z decays in e + e collisions various colliders General approach is a factorized tuning procedure with only some of the components investigated N.B. For the DPS simulation, generators normally use parameters of soft MPI and extrapolate to harder scales Paolo Gunnellini QCD at cosmic energies May 6

Can they be refined? How well do they describe observables at different energy? N ch and p sum as a function of the leading charged particle RANS MIN: sensitive to MPI RANS MAX: sensitive to MPI and PS RANS DIF: sensitive to PS RANS AVE: sensitive to MPI and PS PURPOSE: uning MPI and colour reconnection parameters Leading Object Direction Δφ ransverse- oward Away ransverse- Jet # Pmax Direction ransmax oward Away oward-side Jet Δφ ransmin Away-Side Jet Paolo Gunnellini QCD at cosmic energies May 6

Results of the energy-dependence tuning Charged particle mult. in the MAX reg. @.9 (left) and 7 (right) ev (/Nevents)dN ch/dηdφ MC/data (/Nevents)dN ch/dηdφ MC/data ransmaxcharged-particledensity s =9GeV.9.7.5.......9.7.5...... CDF data PYHIA8 une C CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 6 8 6 8 p max [GeV] ransmaxcharged-particledensity s =9GeV CDF data HERWIG++ une UE-EE-5C CUEHppS 6 8 6 8 p max [GeV] MC/data (/Nevents)dN ch/dηdφ MC/data (/Nevents)dN ch/dηdφ ransmaxcharged-particledensity s =7eV.8.6.. CMS data PYHIA8 une C CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO. CUEP8M... 5 5 5 p max [GeV] ransmaxcharged-particledensity s =7eV.8.6.. CMS data HERWIG++ une UE-EE-5C CUEHppS.... 5 5 5 p max [GeV] New tunes! PYHIA 8 (CUEP8) HERWIG++ (CUEHpp) with various PDFs Better constrain of the energy extrapolation CR changes with the choice of the PDF Rising part and plateaux region are well predicted by the new tunes (arxiv 5.85) Paolo Gunnellini QCD at cosmic energies May 6 5

une performance at the new energy N ch /dη dφ> <d / dη ch dn /N ev.5.5. MC / Data.5.5.5. n ch, ALAS p > 5 MeV, Preliminary s <.5 η = ev Data PYHIA 8 A PYHIA 8 Monash HERWIG++ UE-EE5 EPOS LHC QGSJE II-.5.5.5.5.5.5 η.5 ransverse region ALAS Preliminary...9 p >.5 GeV, η <.5 lead p > GeV DAA (uncorrected) PYHIA 8 A HERWIG++ EE5 s = ev EPOS PYHIA 8 A PYHIA 8 Monash 5 5 5 dn ch /dη CMS 8 7 6 5 pp s = ev inelastic data PYHIA8 CUEP8S EPOS LHC - - - (GeV) η ( φ) / p Σ data/mc data.8 Pythia8 + CUEP8M CMS.6 Pythia8 + Monash Preliminary EPOS. Herwig + CUEHS....6....7 6 8 6 8 5 5 5 Leading rack p (GeV) plead [GeV] Paolo Gunnellini QCD at cosmic energies May 6 6 η 6 8 6 8 HIS_data_syserror - 8 nb transmin Stat. Uncertainty Stat. + Syst. Uncertainty ( ev) s = ev OP: dn/dη ALAS-CONF-5-8, PLB75 (5) BOOM: N ch vs p lead ALAS-PHYS-5-9, CMS-FSQ-5-7 Current tunes reproduce the data at ev! May the energy dependence of the MPI be improved in the generators? p = pref (E/E ref ) ɛ

What about other observables? (arxiv 5.85) dn ch/dη Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5 Forward region UE vs p jet MC/data.. ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections Z-boson observables DPS observables Paolo Gunnellini QCD at cosmic energies May 6 7

What about other observables? (arxiv 5.85) dn ch/dη MC/data. Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5. ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections dn ch/d η MC/data.5 Forward region.5.5.5.5.. Charged-particlemultiplicity, s =7eV OEM data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 5. 5.6 5.8 6 6. 6. η Z-boson observables UE vs p jet DPS observables Paolo Gunnellini QCD at cosmic energies May 6 8

What about other observables? (arxiv 5.85) dn ch/dη MC/data. Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5. ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections dn ch/d η MC/data.5.5.5.5.5. Forward region. Charged-particlemultiplicity, s =7eV OEM data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 5. 5.6 5.8 6 6. 6. η Z-boson observables d N ch/dηdφ UE vs p jet.... ransaven chdensity, s =.76eV CMS data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M CUEHppS 5 6 7 8 9 p(leading charged-particle jet)[gev] DPS observables Paolo Gunnellini QCD at cosmic energies May 6 9

What about other observables? (arxiv 5.85) dn ch/dη MC/data d σ/dpdy[pb/gev] MC/data. Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5. 6 5 ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections CMS,Inclusivejets, s =7eV,.5 < y <. 7.5.5.5 CMS data PH+CUEP8M PH+CUEP8S-HERAPDF.5LO p[gev] dn ch/d η MC/data.5.5.5.5.5. Forward region. Charged-particlemultiplicity, s =7eV OEM data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 5. 5.6 5.8 6 6. 6. η Z-boson observables d N ch/dηdφ UE vs p jet.... ransaven chdensity, s =.76eV CMS data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M CUEHppS 5 6 7 8 9 p(leading charged-particle jet)[gev] DPS observables Paolo Gunnellini QCD at cosmic energies May 6

What about other observables? (arxiv 5.85) dn ch/dη MC/data d σ/dpdy[pb/gev] MC/data. Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5. 6 5 ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections CMS,Inclusivejets, s =7eV,.5 < y <. 7.5.5.5 CMS data PH+CUEP8M PH+CUEP8S-HERAPDF.5LO p[gev] dn ch/d η MC/data /σ dσ/dp(z)[/gev].5.5.5.5.5. Forward region. 5 6 Charged-particlemultiplicity, s =7eV OEM data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 5. 5.6 5.8 6 6. 6. η Z-boson observables Zbosonpinleptonicdecay.. CMS data CUEP8M PH+CUEP8M PH+CUEP8S-HERAPDF.5LO p(z)[gev] d N ch/dηdφ UE vs p jet.... ransaven chdensity, s =.76eV CMS data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M CUEHppS 5 6 7 8 9 p(leading charged-particle jet)[gev] DPS observables Paolo Gunnellini QCD at cosmic energies May 6

What about other observables? (arxiv 5.85) dn ch/dη MC/data d σ/dpdy[pb/gev] MC/data. Min. Bias observables Charged-particlemultiplicity, s =7eV 7 6 5. 6 5 ALICE data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M - -.5.5 η Incl. jet cross sections CMS,Inclusivejets, s =7eV,.5 < y <. 7.5.5.5 CMS data PH+CUEP8M PH+CUEP8S-HERAPDF.5LO p[gev] dn ch/d η MC/data /σ dσ/dp(z)[/gev].5.5.5.5.5. Forward region. 5 6 Charged-particlemultiplicity, s =7eV OEM data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M 5. 5.6 5.8 6 6. 6. η Z-boson observables Zbosonpinleptonicdecay.. CMS data CUEP8M PH+CUEP8M PH+CUEP8S-HERAPDF.5LO p(z)[gev] d N ch/dηdφ (/σ)dσ/d S MC/data UE vs p jet.... ransaven chdensity, s =.76eV CMS data CUEP6S CUEP8S-CEQ6L CUEP8S-HERAPDF.5LO CUEP8M CUEHppS 5 6 7 8 9 p(leading charged-particle jet)[gev] DPS observables X.. Normalized Sinpp j, s=7ev CMS data CUEP8M CUEHppS.5.5.5 S Paolo Gunnellini QCD at cosmic energies May 6

Further look at UE/DPS comparisons (arxiv 5.85) une name σ eff value (mb) CUEP8M 6. +. CUEHppS 5. +.5 ransaven ch density, s = 7eV Dedicated tune to DPS-sensitive observables in four-jet final state CDPSP8S σ eff = 9. +.7. mb Normalized Sinpp j, s=7ev d N ch /dηdφ..... ALAS data CDPSP8S-j 6 8 6 8 p (leadingtrack)[gev] (/σ)dσ/d S MC/data.. CMS data CUEP8M CUEHppS.5.5.5 S Not able to describe both UE and DPS observables at with the same set of tunes Indication for need of a refinement of the current MPI model? Paolo Gunnellini QCD at cosmic energies May 6

High multiplicity scenarios /N dn /dp - ) ch.jet ch.jet (GeV/c MC/data >(GeV/c) <p MC/data - - - - -5.5.5.5 Low multiplicities mainly softer MPI components High multiplicities contribution of harder MPI ch.jet CMS, pp s = 7 ev η <.9 8 5 < N ch data PYHIA8 MPI-off PYHIA8 C PYHIA6 Z* Herwig++.5 UE_EE_M 5 5 5 5 5 5 p (GeV/c) CMS, pp s = 7 ev all ch. particles ch.particle ch.particle η <., p >.5 GeV/c data PYHIA8 MPI-off PYHIA8 C PYHIA6 Z* Herwig++.5 UE_EE_M 6 8 N ch 6 8 N ch /N dn /dp - ) ch.jet ch.jet (GeV/c MC/data /N ch.jetdnch.jet/dp[gev/c] - - - - -5 ch.jet CMS, pp s = 7 ev η <.9.8.6.. < N ch data PYHIA8 C PYHIA6 Z* Herwig++.5 UE_EE_M 5 5 5 5 5 5 p (GeV/c) 5 5 5 5 5 5 p (GeV/c) CMS, s =7eV, η ch.jet <.9, <N ch CMS data CUEP8S-CEQ6L CDPSP8S-j CUEP6S Monash 5 5 5 5 5 p[gev/c] - Charged-particle measurement - Jet measurement Good agreement for UE tunes at low multiplicities but worse for increasing multiplicities Better description at high multiplicities for the DPS tunes! Normalized jet p spectrum at different multiplicities (top and bottom right) Charged particle spectra p as a function of multiplicity (bottom left) Eur.Phys.J. C7 () 67 Paolo Gunnellini QCD at cosmic energies May 6

Diffractive-enhanced final states (I) Diffractive dissociation cross sections as a function of the fraction of the dissociated mass in different ranges dominated by different components est of models for diffractive scenarios (mb) ξ dσ/dlog X.8.6.... CMS <.5 log M Y - 6. µb (7 ev) -6-5.5-5 -.5 - -.5 - -.5 - log (a) ξ X (mb) ξ dσ/dlog X.8.6.... CMS Data PYHIA version: P8-MBR (ε=.8) P8-MBR (ε=.) P8-C P6-Z* <..5 < log M Y - 6. µb (7 ev) -6-5.5-5 -.5 - -.5 - -.5 - log (c) ξ X Best description is provided by MBR generator (arxiv.) which fully simulates the diffractive components and is tuned to evatron data It uses different hadronization parameters to describe charged particle p spectra LEF: SD-enhanced, RIGH: DD-enhanced Phys. Rev. D 9 (5) Paolo Gunnellini QCD at cosmic energies May 6 5

Diffractive-enhanced final states (II) /dη ch ) dn events (/N.6.. CMS Preliminary N ch in η <. p >.5 GeV ( ev) SD selection data PYHIA8 CUEM PYHIA8 CUES PYHIA8 CUEM MBR PYHIA8 C MBR /dη ch ) dn events (/N.8.6. CMS Preliminary N ch in η <. p >.5 GeV ( ev) SD selection data PYHIA8 CUEM PYHIA8 CUES PYHIA8 CUEM MBR PYHIA8 C MBR SD-enhanced selection: activity in one region, veto in the other... d N ch/dηdp [(GeV/c) ]...5.5.5.5 Charged hadron p for η =. at s = 7eV Data CUEP8M CUEP8M-MBR CUEP8M-PomFlux.....6.8 p [GeV/c] η (/πp ) d Nch/dηdp [(GeV/c) ].5.5.5.5 Charged hadron p for η <. at.. s = 7eV Data CUEP8M CUEP8M-MBR CUEP8M-PomFlux 5 6 p [GeV/c] η MBR diffraction model and parameters help to improve the SD observables, due to a better description of low trans.mom. region CMS-FSQ-5-8 Need for different hadronization parameters for diffractive observables than for nondiffractive? Ongoing work! Paolo Gunnellini QCD at cosmic energies May 6 6

Jet observables at ev dy (pb/gev) σ / dp d 5 9 7 5 - - CMS Preliminary C NP anti-k R =.7-7 pb ( ev) 6 y <.5 (x ) 5.5< y <. (x ).< y <.5 (x ).5< y <. (x ).< y <.5 (x ).5< y <. (x ).< y <.7 (x ) Jet p (GeV) Inclusive jet cross section in different rapidity bins Various cone sizes probe different aspects of parton evolution Jets clustered with R =.7 are described by NLO fixed-order calculations and predictions from NLO MC generators LO MC event generators are not able to reproduce the measurement Ratio to C Ratio to PH+P8 CUEM.5.5.5.5.5.5 CMS Preliminary R =.7 anti-k y <.5-7 pb ( ev) Data HERAPDF.5 NNPDF. MMH Exp. uncert. heory uncert. Jet p (GeV) - 7 pb ( ev) CMS Preliminary R =.7 anti-k y <.5 Data PH+P8 CUES-CEQ6L PH+P8 CUES-HERAPDF P8 CUEM Hpp CUES Exp. uncert. Jet p (GeV) Fixed-order NLO ME corr. for NP effects with diff. PDF sets MC generators with LO or NLO dijet ME + PS and UE sim. CMS-PAS-SMP-5-7 ALAS-CONF-5- Paolo Gunnellini QCD at cosmic energies May 6 7

Jet observables at ev dy (pb/gev) σ / dp d 5 9 7 5 - - CMS Preliminary C NP anti-k R =. - 7 pb ( ev) 6 y <.5 (x ) 5.5< y <. (x ).< y <.5 (x ).5< y <. (x ).< y <.5 (x ).5< y <. (x ).< y <.7 (x ) Jet p (GeV) Inclusive jet cross section in different rapidity bins Various cone sizes probe different aspects of parton evolution Jets clustered with R =. are not described by NLO fixed-order calculations but only by predictions from NLO MC generators Effect of missing PS contributions relevant for smaller cones Ratio to C Ratio to PH+P8 CUEM.5.5.5.5.5.5 CMS Preliminary anti-k R =. y <.5-7 pb ( ev) Data HERAPDF.5 NNPDF. MMH Exp. uncert. heory uncert. Jet p (GeV) - 7 pb ( ev) CMS Preliminary anti-k R =. y <.5 Data PH+P8 CUES-CEQ6L PH+P8 CUES-HERAPDF P8 CUEM Hpp CUES Exp. uncert. 5 6 7 8 p [GeV] Paolo Gunnellini QCD at cosmic energies May 6 8 Ratio w.r.t. NLO pqcd (C)... Relative uncertainty of 9% in the integrated luminosity not included Jet p ALAS - 78 pb ev anti-k t jets, y <.5 (GeV) Preliminary R=. DAA (syst., total) NLOJE++ Non-pert. corr. max µ = µ = p F R C MMH NNPDF

Multijet observables (I) σ/dm dy* [pb/ev] d 7 - -5-8 - - -7 ALAS - L dt =.5 fb s = 7 ev anti-k t jets, R =. Systematic uncertainties NLOJE++ C, µ=p exp(. y*) Non-pert. & EW corr...5..5..5 - y * <.5 ( ) - y * <. ( ) -6 y * <.5 ( ) -9 y * <. ( ) - y * <.5 ( ) -5 y * <. ( ) - 5 6 7 Inclusive dijet cross section in different rapidity bins as a function of dijet mass Jets clustered with R =. are reproduced by NLO fixed-order calculations JHEP5()59 [ev] m heory/data y* <.5 P obs =.5 =.6.5 C.5 y* <. C HERA P obs HERA P obs =.98 P obs = 6.5 P obs =.68 P obs =.5.5. y* <.5 C HERA - [ev] m heory/data.5 y* <. C HERA P obs =. P obs =.8. y* <.5 C HERA P obs =. P obs = 6.5 y* <. C HERA P obs =.96 P obs =.98-8 5 [ev] m ALAS - L dt =.5 fb s = 7 ev anti-k t jets, R =. Statistical uncertainty Systematic uncertainties NLOJE++ µ=p exp(. y*) Non-pert. & EW corr. C HERAPDF.5 epaljet exp. only HERAPDF.5 exp. only Paolo Gunnellini QCD at cosmic energies May 6 9

Multijet observables (II) σ/dm dy* [pb/ev] d 7 - -5-8 - - -7 ALAS - L dt =.5 fb s = 7 ev anti-k t jets, R =. Systematic uncertainties NLOJE++ C, µ=p exp(. y*) Non-pert. & EW corr...5..5..5 - y * <.5 ( ) - y * <. ( ) -6 y * <.5 ( ) -9 y * <. ( ) - y * <.5 ( ) -5 y * <. ( ) - 5 6 7 Inclusive dijet cross section in different rapidity bins as a function of dijet mass Jets clustered with R =. are reproduced by MC event generators using NLO dijet ME but sensitivity to UE tunes JHEP5()59 [ev] m heory/data y* <.5...5 y* <... y* <.5 - [ev] m heory/data.5.5.5.5.5.5 y* <.. y* <.5.5 y* <..5-8 5 [ev] m ALAS - L dt =.5 fb s = 7 ev anti-k t jets, R =. Statistical uncertainty Systematic uncertainties POWHEG Born C, µ=p EW corr. + PYHIA shower AUEB Perugia Paolo Gunnellini QCD at cosmic energies May 6

Multijet observables (III) ) [fb/gev] () dσ / d(p heory/data ) [fb/gev] () dσ / d(p heory/data 6 5.5.5 6 5.5.5 ALAS ALAS s=8 ev, 95 pb. fb p [GeV] () s=8 ev, 95 pb. fb () p [GeV] Data HEJ (.9) BlackHat/Sherpa (.) NJet/Sherpa (.) () p > GeV otal experimental systematic uncertainty NLO (scale PDF) uncertainty Data Pythia 8 ( ) Herwig++ (.) MadGraph+Pythia (.) () p > GeV otal experimental systematic uncertainty Four-jet scenarios also well described in a wide region of the phase space JHEP (5) 5 Phys.Rev.D89()9 η <.7 SHERPA st nd, jet: POWHEG+P6 Z' MADGRAPH+P6 Z* p > 5 GeV PYHIA8 C rd th, jet: HERWIG++ UE-EE- p > GeV otal Uncertainty - CMS, s = 7 ev, L = 6 pb, pp j+x 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 Subleading jet Jet #pt Paolo Gunnellini QCD at cosmic energies May 6.8.6.....8.6.....8.6.... Leading jet 5 5 5 5 5 5 Jet #pt hird jet 5 5 5 5 5 5.8 Fourth jet Jet #pt.6.... 5 5 5 5 Jet 5 p (GeV) 5

Z boson observables Jets associated to a Z boson Measurement of differential cross sections for different jet multiplicities Crucial test of our QCD predictions! dσ/dh (jets) [pb/gev] CMS Preliminary -.5 fb ( ev) Data (5ns) amc@nlo + PY8 ( j NLO + PS) dσ/dh (jets) [pb/gev] CMS Preliminary -.5 fb ( ev) Data (5ns) amc@nlo + PY8 ( j NLO + PS) ) [pb] jets σ(z( µµ)+n - Z µ + µ + jets Data ALAS Preliminary - ev, 85 pb Sherpa Madgraph anti-k t, R=. jet p > GeV jet y <.5 amc@nlo/data.5.5 anti-k (R =.) Jets jet jet p > GeV, y <. Z/γ* µµ channel Stat. unc. (gen) 6 8 H, N [GeV] jets amc@nlo/data 5.5.5 anti-k (R =.) Jets jet jet p > GeV, y <. Z/γ* µµ channel Stat. unc. (gen) 6 8 H, N [GeV] jets Data / Pred... CMS-SMP-5- ALAS-CONF-5- Predictions describe measurements as long as the multiplicity matches the number of partons in the matrix element N jets N jets Paolo Gunnellini QCD at cosmic energies May 6

op observables (I) Underlying event in top pair events Measurement of event content in terms of charged particle multiplicity and p est of final-state universality of UE contribution ν μ μ + W + b t t t t W - b ) [/GeV] ch dn ev /d(p /N ev q q CMS Preliminary overall away transverse toward tt -. fb ( ev) Powheg + Pythia 8 (CUEP8M) [GeV] ch p.5 CMS.5.5 -. fb ( ev) Preliminary detector-level data tt Powheg + Pythia 8 (CUEP8M) tt Powheg + Pythia 8 (CUEP8M) (Q) tt Powheg + Pythia 8 (CUEP8M) (Q/) tt Powheg + Herwig++ (EE5C) transverse CMS- OP-5-7.5.5 5 5 5 ch p [GeV].5. tt p [GeV] Predictions of POWHEG NLO matrix-element interfaced to P8 tunes from hadronic events are able to reproduce all measured observables in each considered region Paolo Gunnellini QCD at cosmic energies May 6

op observables (II) ] op pair in lepton + jets final states Measurement of differential cross sections as a function of top pair p and jet multiplicity est of reliability of predictions in top sector - [pb GeV (tt) dσ dp CMS l+jets particle Preliminary -. fb ( ev) data sys stat stat. MG5_aMC@NLO P8 [FxFx] MG5_aMC@NLO P8 [MLM] Powheg P8 Powheg H++ MG5_aMC@NLO P8 MG5_aMC@NLO H++ dσ d n-jet 5 CMS l+jets particle Preliminary -. fb ( ev) data sys stat stat. MG5_aMC@NLO P8 [FxFx] MG5_aMC@NLO P8 [MLM] Powheg P8 Powheg H++ MG5_aMC@NLO P8 MG5_aMC@NLO H++ theory data.. 5 p (tt) [GeV] theory data.5 >= additional jets NLO predictions describe well the measurements but sensitivity to matching scheme and PS generator CMS-OP-6-8 Paolo Gunnellini QCD at cosmic energies May 6

Parton shower and MPI tuning UNING OF PYHIA 8 O OBSERVABLES MEASURED IN DIFFEREN PROCESSES Observables rack jet properties Jet shapes Dijet decorrelations Multijets Z boson p t t gap and jet shapes rack-jet and jet UE Study of the interplay between MPI and parton shower Various PDF sets investigated Extremely important for: testing the universality of the parton shower in leptonic and hadronic collisions testing the performance of UE simulation for different hard scattering processes AL-PHYS-PUB-- Paolo Gunnellini QCD at cosmic energies May 6 5

Parton shower and MPI tuning Significant improvement of description of observables in: Drell-Yan (left), top-antitop (center) and jet substructure (right) sectors dσ fid. σ fid. dp Z p 5 6 7..5..95.9 5.75.7 dressed muons ALAS Data AU CEQ6L Monash A-CEQ A-MSW A-NNPDF A-HERA ALAS Simulation Z p f gap..95.9 5.75.7....98.96 Gap fraction vs. Q for veto region: y < ALAS Simulation ALAS Data AU CEQ6L Monash A-CEQ A-MSW A-NNPDF A-HERA 5 5 5 Q [GeV] /σdσ/dτ.8.6..... Cambridge-Aachen jets, R =., GeV < p < GeV α S values are similar for all PDF used quite high for the hard processes lower for initial- and final-state radiation significantly lower than previous tunes for ISR and FSR Damped shower needed to describe gap fraction in t t events and p Z simultaneously. ALAS Data AU CEQ6L Monash A-CEQ A-MSW A-NNPDF A-HERA ALAS Simulation... N-subjettiness τ Paolo Gunnellini QCD at cosmic energies May 6 6

uning higher-order ME matched to parton showers OBSERVABLES: gap fraction, jet shapes and jet multiplicity in t t events GENERAORS: PYHIA 8 standalone, MADGRAPH amc@nlo and POWHEG + PYHIA 8 Different steps of tuning: tuning of ISR and FSR separately, then simultaneous tune to account for their interplay application of tune to matched generators tune of the matched MADGRAPH amc@nlo + PYHIA 8 (Some of the) Outcomes Significant improvement in the description of t t observables Parameters of simultaneous ISR-FSR tune do not differ much from the separate tunes une of matched MADGRAPH+PYHIA 8 has similar parameters to standalone PYHIA 8 AL-PHYS-PUB-5-7, AL-PHYS-PUB-5-8 Paolo Gunnellini QCD at cosmic energies May 6 7

Summary and conclusions High energy physics strongly relies on predictions from Monte Carlo event simulation Very sophisticated tunes are available and are able to describe a wide range of observables at different collision energies Actual measurements are sensitive to ME, UE and PS contributions and are able to disentangle them unes obtained at 7 ev (or below) are able to reproduce data at ev in a good level of agreement It is difficult to reproduce data relative to some corners of the phase space within a unique set of parameters (e.g. DPS-sensitive observables, diffraction) Need for more sophisticated models/parametrizations? New measurements, more advanced simulation.. Stay tuned! Paolo Gunnellini QCD at cosmic energies May 6 8

Summary and conclusions High energy physics strongly relies on predictions from Monte Carlo event simulation Very sophisticated tunes are available and are able to describe a wide range of observables at different collision energies Actual measurements are sensitive to ME, UE and PS contributions and are able to disentangle them unes obtained at 7 ev (or below) are able to reproduce data at ev in a good level of agreement It is difficult to reproduce data relative to some corners of the phase space within a unique set of parameters (e.g. DPS-sensitive observables, diffraction) Need for more sophisticated models/parametrizations? New measurements, more advanced simulation.. Stay tuned! HANK YOU FOR YOUR AENION Paolo Gunnellini QCD at cosmic energies May 6 9

BACKUP SLIDES Paolo Gunnellini QCD at cosmic energies May 6

A double parton scattering (DPS) wo different x values for the two partons in each proton Impact parameter b σ DPS A,B = m dx dx ˆσ A(x, x ) dx dx ˆσ B(x, x ) d b f (x, x, b) f (x, x, b) DPS Cross Section Partonic cross sections Double parton distribution functions If the partons are assumed to be uncorrelated: f (x, x, b) = f (x )f (x )F (b) he two scatterings factorize in the cross section formula: σab DPS = m dx dx ˆσ A f (x )f (x ) dx dx ˆσ B f (x )f (x ) d b [F (b)] It is thus defined: σ eff = d b [F (b)] Paolo Gunnellini QCD at cosmic energies May 6

Combined fits to whole MPI spectrum Combined fits of observables sensitive to soft, semi-hard and hard MPI (/σ)dσ/d S [/rad] Normalized S in pp j in η <.7 at s = 7 ev CMS data DPS-UE-MB une dn/dη 7 6 5 Pseudorapidity (s) = 7 ev, INEL > ALICE data DPS-UE-MB une (/Nevents) dn ch/dηdφ ransmin charged particle density s = 7eV CMS data DPS-UE-MB une (/σ)dσ/d S [/rad]....9.5.5.5 S (rad) Normalized S in pp j in η <.7 at s = 7 ev....9 CMS data DPS-UE-MB-Weighted une.5.5.5 S (rad) dn/dη.. - -.5.5 η Pseudorapidity (s) = 7 ev, INEL > 7 6 5.. ALICE data DPS-UE-MB-Weighted une - -.5.5 η No hope to get the measurements well described altogether (/Nevents) dn ch/dηdφ.. 5 5 5 pmax[gev/c] ransmin charged particle density s = 7eV.. CMS data DPS-UE-MB-Weighted une 5 5 5 pmax[gev/c] Paolo Gunnellini QCD at cosmic energies May 6

Why do we actually tune? Not only for fun! Correct description of the data Pile-up simulation Evaluation of detector effects and unfolding Estimation of background (in MC-driven approach) Models are not allowed to fail Good physics predictions Correct evaluation of physics effects Models are allowed to fail he danger is overtuning! Paolo Gunnellini QCD at cosmic energies May 6