The Underlying Event in Hard Scattering Processes What

The “Underlying Event” in Hard Scattering Processes Æ What happens when a proton and an antiproton collide with a center-ofmass energy of 2 Te. V? Æ Most of the time the proton and antiproton ooze through each other and fall apart (i. e. no hard scattering). The outgoing particles continue in roughly the same direction as initial proton and antiproton. Æ Occasionally there will be a “hard” parton-parton collision resulting in large transverse momentum outgoing partons. Æ The “underlying event” is everything except the two outgoing hard scattered “jets”. It is an unavoidable background to many collider observables. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 1

Beam-Beam Remnants Æ The underlying event in a hard scattering process has a “hard” component (particles that arise from initial & final-state radiation and from the outgoing hard scattered partons) and a “soft” component (beam-beam remnants). Æ However the “soft” component is color connected to the “hard” component so this separation is (at best) an approximation. Min-Bias? Æ For ISAJET (no color flow) the “soft” and “hard” components are completely independent and the model for the beam-beam remnant component is the same as for min-bias (“cut pomeron”) but with a larger <PT>. Æ HERWIG breaks the color connection with a soft q-qbar pair and then models the beam-beam remnant component the same as HERWIG min-bias (cluster decay). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 2

Studying the “Underlying Event” at CDF The Underlying Event: beam-beam remnants initial-state radiation multiple-parton interactions Æ The underlying event in a hard scattering process is a complicated and not very well understood object. It is an interesting region since it probes the interface between perturbative and non-perturbative physics. Æ There are two CDF analyses which quantitatively study the underlying event and compare with the QCD Monte-Carlo models. Æ It is important to model this region well since it is an unavoidable background to all collider observables. Also, we need a good model of min-bias (zero-bias) collisions. CDF Cone Analysis Valeria Tano Eve Kovacs Joey Huston Anwar Bhatti Ph. D. Thesis D 0 Meeting September 6, 2002 Evolution of Charged Jets Rick Field - Florida/CDF Rick Field David Stuart Rich Haas PRD 65: 092002, 2002 3

Evolution of Charged Jets “Underlying Event” Charged Particle Df Correlations PT > 0. 5 Ge. V/c |h| < 1 Toward-side “jet” (always) Perpendicular to the plane of the 2 -to-2 hard scattering Very sensitive to the “underlying event” Away-side “jet” Æ Look at charged particle correlations in the azimuthal angle Df relative to the leading charged (sometimes) Æ Æ particle jet. Define |Df| < 60 o as “Toward”, 60 o < |Df| < 120 o as “Transverse”, and |Df| > 120 o as “Away”. All three regions have the same size in h-f space, Dhx. Df = 2 x 120 o = 4 p/3. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 4

Charged Multiplicity versus PT(chgjet#1) Underlying Event “plateau” Factor of 2 more active than an average Min-Bias event! Æ Data on the average number of “toward” (|Df|<60 o), “transverse” (60<|Df|<120 o), and “away” (|Df|>120 o) charged particles (PT > 0. 5 Ge. V, |h| < 1, including jet#1) as a function of the transverse momentum of the leading charged particle jet. Each point corresponds to the <Nchg> in a 1 Ge. V bin. The solid (open) points are the Min-Bias (JET 20) data. The errors on the (uncorrected) data include both statistical and correlated systematic uncertainties. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 5

“Transverse” PT Distribution PT(charged jet#1) > 30 Ge. V/c “Transverse” <Nchg> = 2. 3 PT(charged jet#1) > 5 Ge. V/c “Transverse” <Nchg> = 2. 2 Æ Comparison of the “transverse” <Nchg> versus PT(charged jet#1) with the PT distribution of the “transverse” <Nchg>, d. Nchg/d. PT. The integral of d. Nchg/d. PT is the “transverse” <Nchg>. Shows how the “transverse” <Nchg> is distributed in PT. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 6

“Max/Min Transverse” Nchg versus PT(chgjet#1) Bryan Webber idea! More sensitive to the “hard scattering” component Area Dh. Df 2 x 60 o = 2 p/3 “Trans. MAX” More sensitive to the “beam-beam remnants” “Trans. MIN” Æ Define “Trans. MAX” and “Trans. MIN” to be the maximum and minimum of the region Æ 60 o<Df<120 o (60 o<-Df<120 o) on an event by event basis. The overall “transverse” region is the sum of “Trans. MAX” and “Trans. MIN”. The plot shows the average “Trans. MAX” Nchg and “Trans. MIN” Nchg versus PT(charged jet#1). The solid (open) points are the Min-Bias (JET 20) data. The errors on the (uncorrected) data include both statistical and correlated systematic uncertainties. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 7

ISAJET: “Transverse” Nchg versus PT(chgjet#1) ISAJET Outgoing Jets plus Initial & Final-State Radiation Beam-Beam Remnants Æ Plot shows the “transverse” <Nchg> vs PT(chgjet#1) compared to the QCD hard Æ scattering predictions of ISAJET 7. 32 (default parameters with PT(hard)>3 Ge. V/c). The predictions of ISAJET are divided into two categories: charged particles that arise from the break-up of the beam and target (beam-beam remnants); and charged particles that arise from the outgoing jet plus initial and final-state radiation (hard scattering component). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 8

HERWIG: “Transverse” Nchg versus PT(chgjet#1) HERWIG Outgoing Jets plus Initial & Final-State Radiation Beam-Beam Remnants Æ Plot shows the “transverse” <Nchg> vs PT(chgjet#1) compared to the QCD hard Æ scattering predictions of HERWIG 5. 9 (default parameters with PT(hard)>3 Ge. V/c). The predictions of HERWIG are divided into two categories: charged particles that arise from the break-up of the beam and target (beam-beam remnants); and charged particles that arise from the outgoing jet plus initial and final-state radiation (hard scattering component). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 9

HERWIG: “Transverse” PT Distribution HERWIG has the too steep of a PT dependence of the “beam-beam remnant” component of the “underlying event”! PT(charged jet#1) > 30 Ge. V/c “Transverse” <Nchg> = 2. 2 PT(charged jet#1) > 5 Ge. V/c “Transverse” <Nchg> = 1. 7 Æ Data on the “transverse” <Nchg> versus PT(charged jet#1) and the PT distribution of the “transverse” <Nchg>, d. Nchg/d. PT, compared with the QCD Monte-Carlo predictions of HERWIG 5. 9 (default parameters with PT(hard) > 3 Ge. V/c). The integral of d. Nchg/d. PT is the “transverse” <Nchg>. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 10

MPI: Multiple Parton Interactions Æ PYTHIA models the “soft” component of the underlying event with color string fragmentation, but in addition includes a contribution arising from multiple parton interactions (MPI) in which one interaction is hard and the other is “semi-hard”. Æ The probability that a hard scattering events also contains a semi-hard multiple parton interaction can be varied but adjusting the cut-off for the MPI. Æ One can also adjust whether the probability of a MPI depends on the PT of the hard scattering, PT(hard) (constant cross section or varying with impact parameter). Æ One can adjust the color connections and flavor of the MPI (singlet or nearest neighbor, q-qbar or glue-glue). Æ Also, one can adjust how the probability of a MPI depends on PT(hard) (single or double Gaussian matter distribution). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 11

PYTHIA: Multiple Parton Interactions Pythia uses multiple parton interactions to enhace the underlying event. Parameter Value MSTP(81) 0 Multiple-Parton Scattering off 1 Multiple-Parton Scattering on 1 Multiple interactions assuming the same probability, with an abrupt cut-off PTmin=PARP(81) 3 Multiple interactions assuming a varying impact parameter and a hadronic matter overlap consistent with a single Gaussian matter distribution, with a smooth turnoff PT 0=PARP(82) 4 Multiple interactions assuming a varying impact parameter and a hadronic matter overlap consistent with a double Gaussian matter distribution (governed by PARP(83) and PARP(84)), with a smooth turn-off PT 0=PARP(82) MSTP(82) D 0 Meeting September 6, 2002 Description Rick Field - Florida/CDF and now HERWIG ! Herwig MPI J. M. Butterworth J. R. Forshaw M. H. Seymour Multiple parton interaction more likely in a hard (central) collision! Hard Core 12

PYTHIA Multiple Parton Interactions Note: Multiple parton interactions depend sensitively on the PDF’s! Æ Plot shows “transverse” <Nchg> versus PT(chgjet#1) compared to the QCD hard Æ Æ scattering predictions of PYTHIA with PT(hard) > 0 Ge. V/c. PYTHIA 6. 115: GRV 94 L, MSTP(82)=3, PT 0=PARP(82)=1. 55 Ge. V/c. PYTHIA 6. 115: CTEQ 3 L, MSTP(82)=3, PT 0=PARP(82)=1. 35 Ge. V/c. PYTHIA 6. 115: CTEQ 4 L, MSTP(82)=3, PT 0=PARP(82)=1. 8 Ge. V/c. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF Varying Impact Parameter 13

PYTHIA Multiple Parton Interactions Note dependence on PT 0. Larger PT 0 means less multiple parton interactions. Æ Plots shows data on the “trans. MAX/MIN” <PTsum> vs PT(chgjet#1) compared to the Æ Æ Æ QCD hard scattering predictions of PYTHIA with PT(hard) > 0 Ge. V/c. PYTHIA 6. 115: CTEQ 4 L, MSTP(82)=3, PT 0=PARP(82)=1. 6 Ge. V/c (solid). PYTHIA 6. 115: CTEQ 4 L, MSTP(82)=3, PT 0=PARP(82)=1. 8 Ge. V/c (dashed). PYTHIA 6. 115: CTEQ 4 L, MSTP(82)=3, PT 0=PARP(82)=2. 0 Ge. V/c (dotted). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 14

PYTHIA 6. 206 Defaults PYTHIA default parameters Parameter 6. 115 6. 125 6. 158 6. 206 MSTP(81) 1 1 MSTP(82) 1 1 PARP(81) 1. 4 1. 9 PARP(82) 1. 55 2. 1 1. 9 PARP(89) 1, 000 PARP(90) 0. 16 4. 0 1. 0 PARP(67) 4. 0 Æ Plot shows “Transverse” <Nchg> versus PT(chgjet#1) compared to the QCD hard scattering predictions of PYTHIA 6. 206 (PT(hard) > 0) using the default parameters for multiple parton interactions and CTEQ 3 L, CTEQ 4 L, and CTEQ 5 L. Constant Note Change PARP(67) = 4. 0 (< 6. 138) PARP(67) = 1. 0 (> 6. 138) D 0 Meeting September 6, 2002 Version 6. 120 PT 0(Ecm) = PT 0(Ecm/E 0)e E 0 = PARP(89) e = PARP(90) Rick Field - Florida/CDF Default parameters give very poor description of the “underlying event”! Probability Scattering 15

Azimuthal Correlations Æ Predictions of PYTHIA 6. 158 (CTEQ 4 L, PARP(67)=1) for the azimuthal angle, Df, between a bquark with PT 1 > 5 Ge. V/c and |y 1| < 1 and a bbar-quark with PT 2 > 0 Ge. V/c and |y 2|<1 in protonantiproton collisions at 1. 8 Te. V. The curves correspond to ds/d. Df (mb/o) for flavor creation, flavor excitation, shower/fragmentation, and the resulting total. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 16

Azimuthal Correlations New PYTHIA default (less initial-state radiation) Old PYTHIA default (more initial-state radiation) Æ Predictions of PYTHIA 6. 206 (CTEQ 5 L) with PARP(67)=1 (new default) and PARP(67)=4 (old default) for the azimuthal angle, Df, between a b-quark with PT 1 > 15 Ge. V/c, |y 1| < 1 and bbar-quark with PT 2 > 10 Ge. V/c, |y 2|<1 in proton-antiproton collisions at 1. 8 Te. V. The curves correspond to ds/d. Df (mb/o) for flavor creation, flavor excitation, shower/fragmentation, and the resulting total. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 17

Azimuthal Correlations Old PYTHIA default (more initial-state radiation) Æ Predictions of HERWIG 6. 4 (CTEQ 5 L) for the azimuthal angle, Df, between a b-quark with PT 1 > 15 Ge. V/c, |y 1| < 1 and bbar-quark with PT 2 > 10 Ge. V/c, |y 2|<1 in proton-antiproton collisions at 1. 8 Te. V. The curves correspond to ds/d. Df (mb/o) for flavor creation, flavor excitation, shower/fragmentation, and the resulting total. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF New PYTHIA default (less initial-state radiation) “Flavor Creation” 18

Di. Photon Correlations Æ Predictions of PYTHIA 6. 158 (CTEQ 5 L) with PARP(67)=1 (new default) and PARP(67)=4 (old default) for diphoton system PT and the azimuthal angle, Df, between a photon with PT 1 > 12 Ge. V/c, |y 1| < 0. 9 and photon with PT 2 > 12 Ge. V/c, |y 2|< 0. 9 in protonantiproton collisions at 1. 8 Te. V compared with CDF data. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 19

Tuned PYTHIA 6. 206 CTEQ 5 L Parameter Tune 1 Tune 2 MSTP(81) 1 MSTP(82) 3 3 PARP(82) 1. 6 Ge. V 1. 7 Ge. V PARP(85) 1. 0 PARP(86) 1. 0 PARP(89) 1. 8 Te. V PARP(90) 0. 16 PARP(67) 1. 0 4. 0 Bulk of Min-Bias 1 events! Æ Plot shows “Transverse” <Nchg> versus PT(chgjet#1) compared to the QCD hard scattering predictions of two tuned versions of PYTHIA 6. 206 (CTEQ 5 L, PARP(67)=1 and PARP(67)=4). New PYTHIA default (less initial-state radiation) D 0 Meeting September 6, 2002 Old PYTHIA default Old (less initial-state radiation) (less Rick Field - Florida/CDF Can describe transition between “soft” and “hard” regime! 20

Tuned PYTHIA 6. 206 “Transverse” PT Distribution PT(charged jet#1) > 30 Ge. V/c PARP(67)=4. 0 (old default) is favored over PARP(67)=1. 0 (new default)! Æ Data on the “transverse” <Nchg> versus PT(charged jet#1) and the PT distribution of the “transverse” <Nchg>, d. Nchg/d. PT, compared with the QCD Monte-Carlo predictions of two tuned versions of PYTHIA 6. 206 (PT(hard) > 0, CTEQ 5 L, PARP(67)=1 and PARP(67)=4). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 21

Tuned PYTHIA 6. 206 vs HERWIG 6. 4 “Trans. MAX/MIN” vs PT(chgjet#1) <Nchg> <PTsum> Æ Plots shows data on the Æ “trans. MAX/MIN” <Nchg> and “trans. MAX/MIN” <PTsum> vs PT(chgjet#1). The solid (open) points are the Min-Bias (JET 20) data. The data are compared with the QCD Monte-Carlo predictions of HERWIG 6. 4 (CTEQ 5 L, PT(hard) > 3 Ge. V/c) and two tuned versions of PYTHIA 6. 206 (PT(hard) > 0, CTEQ 5 L, PARP(67)=1 and PARP(67)=4). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 22

Tuned PYTHIA 6. 206 vs HERWIG 6. 4 “Trans. SUM/DIF” vs PT(chgjet#1) <Nchg> <PTsum> Æ Plots shows data on the Æ “trans. SUM/DIF” <Nchg> and “trans. SUM/DIF” <PTsum> vs PT(chgjet#1). The solid (open) points are the Min-Bias (JET 20) data. The data are compared with the QCD Monte-Carlo predictions of HERWIG 6. 4 (CTEQ 5 L, PT(hard) > 3 Ge. V/c) and two tuned versions of PYTHIA 6. 206 (PT(hard) > 0, CTEQ 5 L, PARP(67)=1 and PARP(67)=4). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 23

Tuned PYTHIA 6. 206 vs HERWIG 6. 4 “Transverse” PT Distribution Æ Data on the PT distribution of the “transverse” <Nchg>, d. Nchg/d. PT, compared with the QCD Monte-Carlo predictions of HERWIG 6. 4 (CTEQ 5 L, PT(hard) > 3 Ge. V/c) and two tuned versions of PYTHIA 6. 206 (PT(hard) > 0, CTEQ 5 L, PARP(67)=1 and PARP(67)=4). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 24

The Underlying Event: Summary & Conclusions The “Underlying Event” Æ Combining the two CDF analyses gives a quantitative study of the underlying event from Æ Æ very soft collisions to very hard collisions. ISAJET (with independent fragmentation) produces too many (soft) particles in the underlying event with the wrong dependence on PT(jet#1). HERWIG and PYTHIA modify the leading-log picture to include “color coherence effects” which leads to “angle ordering” within the parton shower and do a better job describing the underlying event. Both ISAJET and HERWIG have the too steep of a PT dependence of the beam-beam remnant component of the underlying event and hence do not have enough beam-beam remnants with PT > 0. 5 Ge. V/c. PYTHIA (with multiple parton interactions) does the best job in describing the underlying event. Perhaps the multiple parton interaction approach is correct or maybe we simply need to improve the way the Monte-Carlo models handle the beam-beam remnants (or both!). D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 25

Multiple Parton Interactions: Summary & Conclusions Proton Multiple Parton Interactions Anti. Proton Energy dependence? Hard Core Æ The increased activity in the underlying event in a hard scattering over a soft collision Æ Æ Æ cannot be explained by initial-state radiation. No time to discuss Multiple parton interactions gives a natural way of explaining the increased in the thisactivity here! underlying event in a hard scattering. A hard scattering is more likely to occur when the hard cores overlap and this is also when the probability of a multiple parton interaction is greatest. For a soft grazing collision the probability of a multiple parton interaction is small. PYTHIA (with varying impact parameter) describes the underlying event data fairly well and will also fit the min-bias data (must use MSTP(82)=4 “double Gaussian” and tune the parameters). More work is needed on the energy dependence. A. Moraes, I. Dawson, and C. Buttar (University of Sheffield) have also been working on tuning PYTHIA to fit the underlying event using the CDF data with the goal of extrapolating to the LHC. D 0 Meeting September 6, 2002 Rick Field - Florida/CDF 26