Monte Carlo Tuning The HERA Experience Monte Carlo

Monte Carlo Tuning: The HERA Experience • Monte Carlo Models for DIS events • Description of inclusive hadronic final state • Parameter tuning for Ariadne, Herwig, Lepto • Jets at high Q 2 and small x Ursula Bassler, LPNHE-Paris, RUN II MC workshop 1

Modeling ep interactions • proton structure: pdf • hard interaction: LO ME calculation at O( S) • QCD radiation: Parton Shower Models, Color Dipole Model • hadronisation: String or Cluster fragmentation 2

Parton Density Functions • strong constraints from structure function measurements • pdf’s determined with global fit programs: MRST, CTEQ hadronic final state is a good probe for QCD models independent of pdf’s. 3

MC Models at HERA: • MC Models used for DIS: Lepto, Ariadne, Herwig, Rapgap • MC Models used for p: Pythia, Phojet • MC Models at Small x: LDCMC, Smallx, Cascade • MC Models for diffraction: Rapgap, Lepto SCI, Ridi, Diff. VM 4

Where it started from… 92 • first hadronic final state measurements with Lint= 1. 6 nb-1 • transverse energy flow in the laboratory frame w. r. t. and e • comparison to various models: • Leading Log Parton Showers with max. virtuality scale Q 2 (LEP) or W 2 (Lepto 5. 2) • O( s) matrix element and parton shower (Lepto 6. 1) • Color Dipole Model (Ariadne 4. 03)! 5

Where it got to…. • transverse energy flow from 1994 data L=2. 7 pb-1 • 3. 2 < Q 2 < 2200 Ge. V 2 8· 10 -5 < x < 0. 11 increased precision requires improved understanding of Monte Carlo Models fine tuning of MC parameters possible and necessary 6

7

Inclusive hadronic final state G. Grindhammer et al: Comparison of energy flow and particle spectra in the hadronic CMS p * • Lorentz transformation from lab frame • Ariadne, Lepto, Rapgap and Herwig compared for various parameter sets 8

Lepto 6. 5 • ME calculation reproduce cross-sections • QCD cascade: – DGLAP based leading-log parton showers – strong ordering of gluons in kt • fragmentation: – JETSET - string model • parameters: – “Soft Color Interaction” between partons from hard interaction and proton remnant – “Generalized Area Law”: allows interactions between color string pieces 9

Rapgap 2. 06/48 • originally developed for description of diffractive events • takes into account direct and resolved virtual photon contributions • QCD cascade/fragmentation: – similar to Lepto • parameters: – resolved process scale = pt(jet)2+Q 2 – matrix element cut-off: PT 2 CUT=4 Ge. V 2 10

Herwig 5. 9 • QCD cascade: – coherent parton cascade with LO ME corrections – LO shower, but NLO S running • fragmentation: – cluster fragmentation • parameters: – strongly constraint from e+e- data – CLMAX: maximum cluster mass – PSPLT: cluster splitting 11

Ariadne 4. 10 • QCD cascade: based on the color dipole model – gluon emission from independently radiating dipoles – no ordering of gluons in k. T, BFKL emulation – gluon emission corrected to reproduce ME O( s) • fragmentation: JETSET • parameters: – PARA(10): suppression of soft gluon emission for proton remnant – PARA(15): for the struck quark – PARA(25): gluon emission outside suppression cut 12

Transverse Energy Flow Q 2 = 3. 2 Ge. V 2 14. 1 Ge. V 2 175 Ge. V 2 x= 0. 8 10 -4 0. 63 10 -3 0. 4 10 -2 proton remnant 2200 Ge. V 2 0. 11 • peaking ET in “current jet” region with rising Q 2 • plateau behavior at low Q 2 G. Grindhammer et al. A: 99/1 p(10) 1. 6 p(15) 0. 5 p(25) 1. 4 99/2 p(10) 1. 2 p(15) 1. 0 p(25) 1. 0 sgsr sgsc prob H: LO: CLMAX 3. 35 PSPLT 1. 0 96: CLMAX 5. 5 PSPLT 0. 65 99/1: CLMAX 3. 0 PSPLT 1. 2 99/2: CLMAX 5. 0 PSPLT 1. 0 Data: H 1 Eur. Phys. J C 12 (2000) 13

Charged particle multiplicity Q 2 = 7 Ge. V 2 x= 1. 6 10 -4 14 Ge. V 2 0. 64 10 -3 32 Ge. V 2 2. 1 10 -3 • reasonable descriptions can be found for all models • Herwig shows large variations depending on input parametrs G. Grindhammer et al. proton remnant Data: H 1 Nucl. Phys. B 485 (1997) 14

Charged particles multiplicities Q 2 = 7 Ge. V 2 x= 1. 6 10 -4 G. Grindhammer et al. proton remnant 14 Ge. V 2 0. 64 10 -3 32 Ge. V 2 2. 1 10 -3 • p*t > 1 Ge. V • only Ariadne and the high CLMAX parameter sets of Herwig give a good description Data: H 1 Nucl. Phys. B 485 (1997) 15

Pt spectrum Q 2 = 7 Ge. V 2 x= 1. 6 10 -4 14 Ge. V 2 0. 64 10 -3 32 Ge. V 2 2. 1 10 -3 • 0. 5 < * < 1. 5 G. Grindhammer et al. • difficulties at high pt for low Q 2 • only Ariadne describes the full phase space Data: H 1 Eur. Phys. J C 12 (2000) 16

MC parameter tuning N. H Brook et al. : tuning on hadronic final state variables in various Q 2 regions: • x. P in current region of the Breit frame • ET flow in hadronic center of mass system • event shape variables: thrust TC and TZ, jet broadening Bc, jet mass C • fragmentation function • differential and integrated jet shapes • di-jet production cross-sections • charged particle distributions compute combined 2 for all variables difficulties in describing simultaneously jets and charged particle distributions 17

NH. Brook et al. Ariadne: suppression of soft gluon emission for proton remnant P(10) • sensitive to di-jet cross-section • default parameter: Et spectra too hard at low Q 2 • increasing P(10): - suppression of ET over whole range - effect at low and high ET 18

Ariadne: gluon emission outside suppression cut-off P(25) • decreasing P(25): - larger changes at high ET - effect larger in fwd region • less sensitive to ET flow N. H. Brook et al. default tuned P(10) 1. 0 P(15) 1. 0 P(25) 2. 0 1. 6 0. 5 1. 4 19

Herwig: fragmentation parameters • LO s improves agreement PSPLT: increases ET flow CLMAX: • broader jets • harder momentum spectra for charger particles • no parameter set has been found describing all aspects of DIS data 20

Lepto: improved SCI • modified SCI (Lepto 6. 5. 2 ) suppressing SCI at high Q 2 • improved 2 by a factor ~5 • further improvement on (2+1) jet data varying PARL(8)=zpmin PARL(9)=ŝmin • But: other hadronic final state variables better described by default setting = 1/2(1 -cos *) 21

Jets at high Q 2 • 640 < Q 2 < 35000 Ge. V 2 • MC models used with optimized parameters • zp, xp distributions most sensitive to differences in the models • best description of data by Ariadne modified Durham algorithm 22

Jets in Charged Current Events • event selection in same kinematic region, but smaller cross-section • similar behavior of jets than in Neutral Current • stronger deviations seen for LEPTO w. r. t to data and other models 23

Parton Cascades at small x DGLAP: resummation of ln. Q 2 strong ordering in k. T BFKL: resummation of ln 1/x no ordering in k. T CCFM: color coherence strong angular ordering additional transverse energy in forward direction produced for BFKL and CCFM approach BFKL/CCFM in MC models: Ariadne, LDCMC, Smallx, Cascade 24

Forward Jets at small x • rise of jet crosssection with decreasing x, underestimated by MC Models • Lepto/Herwig and LDCMC predict smaller cross-sections • Ariadne and Rapgap show reasonable agreement 25

CCFM evolution - Cascade • CCFM equation implemented in backward evolution schema • forward jets: - good description for H 1 cross-section - above ZEUS measurement H. Jung, G. P Salam 26

Conclusions • MC tuning at HERA not yet to the precision of LEP, but – hadronic environment probed with a lepton – ongoing progress in understanding of various aspects in hadronic final state – further high precision measurements • ARIADNE gives overall a good picture of DIS events • useful experience for hadron colliders? ! 27