Differential cross sections of global event variables of ttbar
SSNSN-323-63November 8, 2018
Differential cross sections of global event variables of tt Tae Jeong Kim on behalf of the ATLAS and CMS collaborations
Hanyang University, Department of PhysicsWangsimniro 222, Seoul 04763, Korea
During the Run 2 period of the LHC, the ATLAS and CMS experi-ments have accumulated proton-proton collision data corresponding to anintegrated luminosity of around 3 fb − in 2015 and 36 fb − in 2016 at acenter of mass energy of 13 TeV. In the journey of finding new physics,it is of importance to understand the standard model which seems to becomplete after the Higgs discovery. Precision tests must be performed inevery corner of the phase space since new physics can appear in any dif-ferent places. In this proceedings, measurements of the differential crosssections of global event variables from the top quark pair productions atthe both experiments using the data collected during the Run 2 period bythe time of the year 2016 are presented.PRESENTED AT th International Workshop on Top Quark PhysicsBraga, Portugal, September 17–22, 2017 a r X i v : . [ h e p - e x ] D ec Introduction
The ATLAS [1] and CMS experiments [2] at the CERN LHC have collected proton-proton collision data corresponding to an integrated luminosity of around 3 fb − in2015 and 36 fb − in 2016 during the Run 2 period at the center of mass energy of13 TeV. In the journey of finding new physics, it is of importance to understandthe standard model which seems to be complete after the Higgs discovery. Precisiontest must be performed in every corner of the phase space since new physics canappear in any different places. In particular, the tt process is the main backgroundfor many searches for beyond the standard models. Measurements with respect toglobal event variables not requiring kinematic reconstruction such as jet multiplicityand missing transverse momentum (MET) from the top quark pair productions arecomplementary. These global event variables are sensitive to dark matter searches.Measurements with additional jets can allow us to check the validity of the QCDcalculation involving a top quark pair plus additional quarks or gluons by comparinghigher-order predictions against parton shower models and will lead us to betterunderstanding of systematic uncertainties on the theory modelling.In this proceedings, measurements of the differential cross sections of global eventvariables and also the variables of ∆ R = (cid:113) (∆ φ ) + (∆ η ) and invariant mass of thetwo additional jets are presented based on data collected the Run 2 period by theATLAS and CMS experiments. In the ATLAS and CMS experiments, the Monte Carlo (MC) simulated samplesfor the tt signal are generated using the POWHEG (v2) event generator at next-to-leading-order (NLO), interfaced with PYTHIA8 to provide the showering of thepartons and to match soft radiation with the contributions from the matrix elements.For cross-checks and studies of systematic uncertainties, various event generatorsare used and described in Table 1. Note that not all of them are used to evaluatesystematic uncertainties.Differential cross sections for the tt production are measured not only at partonlevel but also at particle level. One advantage of measuring the cross section atparticle level is to decrease the uncertainty which can arise from extrapolating tounmeasurable phase space. The particle-level jets are reconstructed by clustering allstable particle except the selected e , µ and radiated γ as well as neutrinos (but doinclude those from hadron decay) using the anti- k T with a parameter r = 0.4. Toidentify b-jets unambiguously, so called “ghost matching” is used by scaling downthe b hadron momentum to a negligible value and including in the jet clustering.The b-jets are then identified by the presence of the corresponding “ghost” hadrons1mong the jet constituents. The MET in Section 4 is defined as the vectorial sum ofthe transverse momenta of all neutrinos in the events, regardless of origin. Jet and b-jet multiplicities in the tt process are measured using data correspondingto an integrated luminosity of 3.2 fb − and 2.3 fb − at the ATLAS [3] and CMSexperiment [4], respectively. The comparisons between data and MC simulations areshown in Fig. 1. Overall, the prediction has a tendency to underestimate data. Inthis measurement, the largest uncertainty comes from the jet energy scale/resolutionand flavor tagging. The multiplicity for additional jets not from the tt system havealso been measured at the ATLAS experiment with 3.2 fb − [5] and at the CMS ex-periment with 36 fb − [6]. Figure 2 (left) shows that the multiplicity of additionaljets is reasonably modelled by the MC simulation of the POWHEG event gener-ator with PYTHIA8. The best agreement with data in this case comes from theMG5 aMC@NLO with PYTHIA8. The simulated samples in this measurement aretuned for h damp and α ISRs using the results of the differential cross section at 8 TeV [4].In the ATLAS collaboration, the study of the differential cross section of theanglur distribution between two additional b-jets has been performed [7]. Figure 2(right) shows the ∆ R between the two additional b-jets together with various NLOpredictions with 4 flavor scheme (massive b quarks) indicated by the black-colorline. A measurement of the differential cross section of the same variable in the ttbb process at 8 TeV from the CMS experiment is also available in the Ref. [8]. Bothmeasurements show that the parton shower event generator HERWIG++ shows somedeviations in the ttbb process.Highly boosted top quark in the all hadronic channel is also reconstructed withthe threshold of p T >
500 GeV and 350 GeV for a leading and second leading topEvent generator Parton showerPOWHEG (v2) PYTHIA8 (default) / HERWIG++ / HERWIG7MG5 aMC@NLO (NLO) PYTHIA8 with FxFx / HERWIG++ / HERWIG7MG5 (LO) PYTHIA8 with MLMSHERPA default SHERPA tuneTable 1: The Monte Carlo simulated samples used for the differential cross sectionmeasurements at the ATLAS and CMS experiments. Both experiments use the PO-HWEG event generator as a default tt simulated sample and SHERPA as cross checks.The different jet merging schemes for MG5 aMC@NLO are used in the CMS experi-ment. 2uark, respectively. The scalar sum of the transverse momenta of the two top-quarkjets is compared together with the distributions created with the POWHEG andaMC@NLO event generators interfaced to the parton shower models of PYTHIA6 orHERWIG++ in Ref. [9]. E v en t s · DatattSingle topW+jetsZ+jetsDibosonVttMultijetsStat.+Syst. Unc. -1 = 13 TeV, 3.2 fbs ATLAS
Resolved
Jet multiplicity D a t a / P r ed . E v en t s · DatattSingle topW+jetsZ+jetsDibosonVttMultijetsStat.+Syst. Unc. -1 = 13 TeV, 3.2 fbs ATLAS
Resolved b-jet multiplicity D a t a / P r ed . Figure 1: Jet and b-jet multiplicity at the ATLAS experiment [3]
In the CMS experiment, the differential cross sections of global event variables thatdo not require the reconstruction of the tt system are measured based on the full2016 data corresponding to an integrated luminosity of 36 fb − [10]. Measured vari-ables are the MET ( p missT ), the scalar sum of the transverse jet momenta ( H T ), thetransverse momentum of the leptonically decaying boson ( p WT ), the scalar sum of alltransverse momenta of the particles ( S T ), the jet multiplicity ( N jets ), and the trans-verse momentum of the lepton ( p l T ). Figures 3 show the H T (left) and P missT (right)distributions. The MG5 aMC@NLO (leading-order) simulation does not accuratelydescribe distributions in data while three other predictions have general agreement.Using these 6 global variable distributions, a global χ test between the absolute crosssections in data and several simulation models was performed assuming there is nocorrelation between the variables. Without including uncertainties in the predictions,the POWHEG+HERWIG++ simulation has the best agreement in the studied distri-butions with χ /ndf = 62.5/62. The POWHEG+PYTHIA8 simulation has a generalagreement with χ /ndf = 100.4/62. 3 -
10 110 [ pb ] s (13 TeV) -1 particle level+jets m e/ CMS
Preliminary > 100 GeV T p Data stat ¯ Sys Stat P8
OWHEG
P CS
HERPA
S H++
OWHEG
PMG5 P8 [FxFx] ‡ additional jets D a t a T heo r y F r a c t i on o f e v en t s ATLAS
Simulation Preliminary = 13 TeVstt+bb category +jets Powheg+P6tt +jets Powheg+P6 radHitt +jets Powheg+P6 radLowtt +jets Powheg+Hpptt bb R ∆ M C / P o w heg + P Figure 2: Additional jet (left) [6] and ∆ R between two additional b-jets distribution(right) [7]. Measurements of the differential cross section for jet multiplicity and additional jetsas well as the global event variables from the tt are presented by comparing withvarious MC predictions. It has been shown that generally the event generators ofPOWHEG with PYTHIA8, POWHEG with HERWIG++, MG5 aMC@NLO (FxFx)have a general agreement with data. When it comes to the differential cross sectionof the ∆ R of the additional b-jets, the parton shower event generator HERWIG++shows some deviations. There is no single simulation that can provide a good de-scription of all variables simultaneously. In the ATLAS and CMS experiments, manydifferential measurements at 13 TeV with high precision are available and need to becompared with each other to improve the theory modelling. More differential crosssection measurements with the full data of 36 fb − are expected to come soon. References [1] ATLAS Collaboration, JINST 3:S08003,2008.[2] CMS Collaboration, JINST 3:S08004,2008.[3] ATLAS Collaboration, arXiv:1708.00727v2.[4] CMS Collaboration, CMS-PAS-TOP-16-021, https://cds.cern.ch/record/2235192.4 .0000.0010.0020.0030.0040.005 σ d σ d H T ( G e V − ) Preliminary
CMS e, µ + jets combined35.9 fb − (13 TeV) Unfolded dataP
OWHEG + P
YTHIA X F X ) + P YTHIA OWHEG + H
ERWIG ++MG5 aMC@NLO (MLM) + P
YTHIA
500 1000 1500 H T (GeV) p r ed . da t a Stat. Stat. ⊕ Syst.Stat. Stat. ⊕ Syst. σ d σ dp m i ss T ( G e V − ) Preliminary
CMS e, µ + jets combined35.9 fb − (13 TeV) Unfolded dataP
OWHEG + P
YTHIA X F X ) + P YTHIA OWHEG + H
ERWIG ++MG5 aMC@NLO (MLM) + P
YTHIA p missT (GeV) p r ed . da t a Stat. Stat. ⊕ Syst.Stat. Stat. ⊕ Syst.
Figure 3: H T (left) and P missT (right) distributions at the CMS experiment [10].[5] ATLAS Collaboration, Eur. Phys. J. C , 220 (2017).[6] CMS Collaboration, CMS-PAS-TOP-17-002, https://cds.cern.ch/record/2284596.[7] ATLAS Collaboration, ATL-PHYS-PUB-2016-016,http://cds.cern.ch/record/2205262.[8] CMS Collaboration, Eur. Phys. J. C76