Low-energy QCD at the high-energy frontier Andy Buckley University of Edinburgh Higgs-Maxwell Particle Physics Workshop, RSE, 2011-02-09 1/27
Soft QCD at a hard collider ◮ The LHC is the highest-energy particle collider ever made – built to -1 -1 10 10 ] ] directly produce new particles at the -1 -1 [GeV [GeV ATLAS Preliminary -2 -2 TeV scale. 10 10 s = 7 TeV anti-k jets R=0.6 jet jet t T T ◮ But the dominant interactions are still -3 -3 10 10 jet p p jet p >30 GeV | y |<2.8 /d /d T jet jet overwhelmingly soft! -4 -4 N N 10 10 ∫ -1 Data L dt=1 nb d d jet jet ◮ Usually dismiss this stuff as “just min PYTHIA N N -5 -5 10 10 1/ 1/ bias” – but that means it’s collective -6 -6 10 10 QCD interactions of whole nucleon -7 -7 10 10 systems. A theory nightmare! -8 -8 10 10 ◮ In these early days of running we 0 0 100 100 200 200 300 300 400 400 500 500 jet jet p p [GeV] [GeV] want to understand it as well as we T T can: as a background and for pure interest. This talk will be ATLAS-dominated: sorry! But it’s not all that unfair. 2/27
Multiple parton interactions (MPI) Number of parton interactions connected to the ratio of parton–parton cross-sections ˆ σ and total p–p cross-section, σ . QCD total cross-section evolves with √ s , e.g. 1992 Donnachie–Landshoff parameterisation, from S -matrix analyticity: tot ( s ) � 21 . 7 mb · ( s / GeV 2 ) 0 . 0808 → σ tot ( 14 TeV ) = 101–164 mb σ pp New ATLAS measurement of σ inel evolution to 7 TeV: [mb] Schuler and Sj o strand 100 -5 ξ Schuler and Sj o strand: > 10 inel Block 2010 σ -5 Data 2010 s = 7 TeV: ξ > 10 2 80 Data 2010 s = 7 TeV: extrap. to ξ > m /s p pp Data p p Data 60 40 ATLAS Preliminary 20 2 Theoretical predictions and data are shown for ξ > m /s unless specified otherwise p σ ξ ATLAS data extrapolated using Pythia prediction for d /d 0 2 3 4 1 10 10 10 10 s [GeV] 3/27
Minimum bias vs. “underlying event” Soft QCD is interesting because it’s not just a single-parton interaction: instead we have multiple, correlated interactions. Correlations are non-perturbatively generated. Minimum bias is purely soft MPI; underlying event (UE) is soft QCD in the presence of a hard scattering, such as hard QCD, EW boson production. . . or Higgs production/new physics! UE = “partial bias”. There is no sharp distinction. TeV-scale new physics searches are mostly designed to be pretty insensitive to soft QCD, but it’s still important to describe the QCD structure of the events as well as possible. UE could be important for e.g. analyses based on jet-structure. 4/27
MC models of soft QCD UE/MB models in MC generators are based on several things: ◮ Multiple parton interactions (in an eikonal approximation formalism) ◮ Regularised cross-section ( gg → 2 QCD naïvely diverges for low p T , in both cross-section and PDF) ◮ Hadronic transverse matter distribution ◮ (Colour topology rearrangement between all scattered partons) ◮ Black magic! Implemented in PYTHIA, JIMMY, Herwig++, Pythia 8, Sherpa, PHOJET, EPOS, (more?) MPI models are the least predictive part of MC event generators! Lots of non-perturbative QCD, but very dynamic so lattice/semi-analytic methods don’t work (even if they were tractable on MC event CPU timescales) MC models are the place where theory meets experiment – close interaction. 5/27
More MC model details Many variations: basic PYTHIA model is the most used/familiar: ◮ Ansatz: apply a ˆ p ⊥ cutoff, ˆ p 0 ⊥ , below which scatterings are vetoed or their cross-sections are suppressed. PYTHIA uses special p 0 “soft-scattering” matrix elements below the ˆ ⊥ cutoff. p 0 ◮ Another ansatz: assume that ˆ ⊥ evolves with energy with a power law “inspired” by the original Donnachie–Landshoff pomeron fit: � s � e / 2 ⊥ ( √ s ) = ˆ ⊥ ( √ s 0 ) · p 0 p 0 ˆ s 0 ⊥ ( √ s 0 ) and e are user-configurable parameters. Usually set p 0 ˆ √ s 0 = 1800 GeV. DL pomeron e ∼ 0 . 16. ◮ Finally, a configurable nucleon hadronic mass distribution in impact parameter space. PYTHIA has several variants, the most-used being a 2-parameter double-Gaussian. 6/27
Tuning the PYTHIA and JIMMY MPI models Need to fit pheno parameters of asymptotic MPI models to describe data. Model tuning is best done as the final stage of a wider tune. The first stages constrain hadronisation (flavour + kinematics), and initial/final-state parton showers: leave as little room as possible for the MPI to exceed its mandate. ATLAS has driven tuning of the Fortran PYTHIA and HERWIG/JIMMY generators to LHC data: new set of tunes for each generator, using early ATLAS data: diffractive-reduced MB data with N ch ≥ 6, ATLAS UE (limited stats) + CDF MB & UE data. Tuning done using the Rivet analysis system to produce data at lots of points in the tuning parameter space, then parameterisation of the observables is done with the Professor tool to find optimal parameters. 7/27
Minimum bias and PYTHIA AMBT1 Minimum bias data from ATLAS covering quite inclusive charged particle observables using the central tracker. Mainly with a track cut of p T > 500 MeV, but also at 100 MeV: a challenge for the models. Various phase spaces, such as diffraction-suppressing N ch cuts. 4 4 η η η η / d / d / d / d 5 5 ≥ 6, > 500 MeV, | η | < 2.5 ≥ 6, > 500 MeV, | η | < 2.5 n p n p ch ch T T ch ch ch ch 3.5 3.5 ATLAS s = 0.9 TeV ATLAS s = 7 TeV N N N N 4.5 4.5 d d d d ⋅ ⋅ ⋅ ⋅ ev ev ev ev N N N N 3 3 4 4 1/ 1/ 1/ 1/ 3.5 3.5 2.5 2.5 3 3 2 2 2.5 2.5 Data 2009 Data 2010 1.5 1.5 PYTHIA ATLAS AMBT1 PYTHIA ATLAS AMBT1 PYTHIA ATLAS MC09 PYTHIA ATLAS MC09 2 2 PYTHIA DW PYTHIA DW 1 1 PYTHIA 8 PYTHIA 8 1.5 1.5 PHOJET PHOJET 1.2 1.2 1.2 1.2 Data Uncertainties Data Uncertainties MC / Data MC / Data Ratio Ratio Ratio Ratio 1 1 1 1 0.8 0.8 0.8 0.8 -2.5 -2 -1.5 -1 -0.5 -2.5 -2 -1.5 -1 -0.5 0 0 0.5 0.5 1 1 1.5 1.5 2 2 2.5 2.5 -2.5 -2 -1.5 -1 -0.5 -2.5 -2 -1.5 -1 -0.5 0 0 0.5 0.5 1 1 1.5 1.5 2 2 2.5 2.5 η η η η 8/27
Minimum bias and PYTHIA AMBT1 Minimum bias data from ATLAS covering quite inclusive charged particle observables using the central tracker. Mainly with a track cut of p T > 500 MeV, but also at 100 MeV: a challenge for the models. Various phase spaces, such as diffraction-suppressing N ch cuts. ch ch ch ch n n 1 1 n n 1 1 ≥ 6, > 500 MeV, | η | < 2.5 ≥ 6, > 500 MeV, | η | < 2.5 /d /d n p /d /d n p ch ch T T ev ev ev ev ATLAS s = 0.9 TeV ATLAS s = 7 TeV N N N N d d -1 -1 d d -1 -1 10 10 10 10 ⋅ ⋅ ⋅ ⋅ ev ev ev ev N N N N 1/ 1/ 1/ 1/ -2 -2 -2 -2 10 10 10 10 -3 -3 -3 -3 10 10 10 10 -4 -4 -4 -4 10 10 10 10 Data 2009 Data 2010 PYTHIA ATLAS AMBT1 PYTHIA ATLAS AMBT1 -5 -5 PYTHIA ATLAS MC09 -5 -5 PYTHIA ATLAS MC09 10 10 10 10 PYTHIA DW PYTHIA DW PYTHIA 8 PYTHIA 8 -6 -6 PHOJET -6 -6 PHOJET 10 10 10 10 Data Uncertainties Data Uncertainties 1.5 1.5 1.5 1.5 MC / Data MC / Data Ratio Ratio Ratio Ratio 1 1 1 1 0.5 0.5 0.5 0.5 10 10 15 15 20 20 25 25 30 30 35 35 40 40 45 45 20 20 40 40 60 60 80 80 100 100 120 120 n n n n ch ch ch ch 9/27
Minimum bias and PYTHIA AMBT1 Minimum bias data from ATLAS covering quite inclusive charged particle observables using the central tracker. Mainly with a track cut of p T > 500 MeV, but also at 100 MeV: a challenge for the models. Various phase spaces, such as diffraction-suppressing N ch cuts. ] ] ] ] -2 -2 2 2 -2 -2 2 2 10 10 10 10 [ GeV [ GeV ≥ 6, > 500 MeV, | η | < 2.5 [ GeV [ GeV ≥ 6, > 500 MeV, | η | < 2.5 n p n p ch ch T T 10 10 10 10 ATLAS s = 0.9 TeV ATLAS s = 7 TeV 1 1 1 1 T T T T p p p p d d -1 -1 d d -1 -1 10 10 10 10 η η η η /d /d -2 -2 /d /d -2 -2 10 10 10 10 ch ch ch ch -3 -3 -3 -3 N N N N 10 10 10 10 2 2 2 2 ) d ) d ) d ) d -4 -4 -4 -4 10 10 10 10 T T T T p p -5 -5 p p -5 -5 10 10 10 10 π π π π 1/(2 1/(2 1/(2 1/(2 -6 -6 -6 -6 10 10 10 10 -7 -7 -7 -7 ev ev 10 10 Data 2009 ev ev 10 10 Data 2010 N N N N -8 -8 PYTHIA ATLAS AMBT1 -8 -8 PYTHIA ATLAS AMBT1 10 10 10 10 1/ 1/ 1/ 1/ PYTHIA ATLAS MC09 PYTHIA ATLAS MC09 -9 -9 -9 -9 10 10 10 10 PYTHIA DW PYTHIA DW -10 -10 -10 -10 10 10 10 10 PYTHIA 8 PYTHIA 8 -11 -11 PHOJET -11 -11 PHOJET 10 10 10 10 2 2 2 2 Data Uncertainties Data Uncertainties MC / Data MC / Data 1.5 1.5 1.5 1.5 Ratio Ratio Ratio Ratio 1 1 1 1 0.5 0.5 0.5 0.5 1 1 10 10 1 1 10 10 p p [GeV] [GeV] p p [GeV] [GeV] T T T T 10/27
Recommend
More recommend
