t\bar{t}W Production: a very complex process
MMarch 1, 2021 ttW
Production: a very complex process
Marcos Miralles L´opez, on behalf of the ATLAS Collaboration Instituto de F´ısica Corpuscular (IFIC)Universitat de Val`encia – CSIC, Valencia, SPAIN
These Monte Carlo studies describe the impact of higher order effectsin both QCD and EW ttW production. Both next-to-leading inclusiveand multileg setups are studied for ttW
QCD production.PRESENTED AT th International Workshop on Top Quark PhysicsDurham, UK (videoconference), 14–18 September, 2020 This project was funded by LCF/BQ/PI19/11690014 a r X i v : . [ h e p - ph ] F e b Introduction
The ttW process is very interesting from the phenomenological point of view [1]. It isfor instance a main background for some beyond the Standard Model (SM) searchesand other rare top SM processes such as ttH and tttt . Moreover, ttW production rateshave been measured at the LHC by CMS and ATLAS as inclusive cross–sections [2, 3]and those measurements yield larger values than the SM predictions from the CERNYellow Report 4 [4]. This motivates the in–depth study of this process.For these Monte Carlo (MC) studies, the event selection is as follows: the tt pair is decayed semileptonically and the associated W boson is decayed leptonically,being both leptons of the same charge. In addition, the following particle level jetcuts p T ( j ) >
25 GeV and | η | < . . < | η | < . Higher order effects (in the quantum chromodynamic (QCD) strong coupling constant α S and the electro–weak (EW) coupling constant α ) are very important for ttW production and can significantly modify leading order cross–sections. Figure 1 showsthe Born level diagrams due to these higher order corrections that enter the MCsimulations.Figure 1: Examples of ttW Feynman diagrams relevant for these studies: LO QCD( O ( α S α )) and NLO QCD ( O ( α S α )) production in the top part, and the “tree–levelEW” contributions ( O ( α + α S α )) in the bottom.The MG5 aMC@NLO [5] generator is used interfaced with the
Pythia8 [6] par-ton shower (PS) for both multileg and inclusive setups. The following items are ex-plored: scale variations of the renormalisation and factorisation scales ( µ R and µ F ) inthe matrix elements (ME) (for inclusive setups), where up to three different functionalforms are used; multileg setups (with the FxFx [7] algorithm), using NLO–accuratematrix elements for up to one additional jet and LO–accurate matrix elements for up1o two additional jets ( ttW +0 , j NLO +2 j LO); parameter variations that impactthe FxFx matching algorithm.
The QCD corrections have been studied with both NLO inclusive and multileg mergedsetups. Figure 2 shows the studies performed at NLO QCD accuracy for the threepoints mentioned in the previous section. The following conclusions may be extracted: • (a): there is a 10% increase in the cross–section between the (green) defaultdynamical scale used in MG5 aMC@NLO and the (blue) fixed scale used inthe CERN YR4. For all functional forms, there is a big dependence of thecross–section with the chosen value of the scale with σ ( µ i, / /σ ( µ i, ) ∼ . • (b): the nominal multileg (FxFx) sample has a merging scale µ Q = 30 GeV anda p minT ( j ) = 8 GeV. No significant shape effects and a cross–section differenceof about 2% is observed when changing the merging scale. This configurationyields a cross–section of σ F xF xttW = 614 . +12% − fb. • (c) and (d): there is good agreement between both MG5 aMC@NLO and
Sherpa2.2.8 [8, 9] multileg setups inside the uncertainty bands. These showcorrelated scale variations in the ME and PS.
The EW corrections to the ttW process have been recently calculated to increase thecross–section by around 10% [10] which is much bigger than naively expected. This iscaused by the appearance of tW → tW scattering diagrams ( O ( α S α )), as those onthe bottom right of Figure 1. Such corrections and their effects are shown in Figure 3from which the following conclusions may be extracted: • (a): in addition to the strong µ R and µ F scale dependance, the EW correctionspredict a 10% increase in the cross–section throughout for all scale values. Asimilar study has been performed using Sherpa2.2.8 where the effect on thecross–section is of about 5%. • (b) to (d): shape effects of around a 20% are observed for events in the highcentral and forward jet multiplicity regions, as well as in the high pseudo rapidityregion (2 . < | η | < .
5) where the extra jet in tW → tW scattering is expected.2 m / m W ) [f b ]tt fi ( pp s )/2 W + m top = (2m m,0 m fixed T,W m (cid:215) tT, m (cid:215) T,t m = g,0 m dyn. geom. T,i m i (cid:229) T = H a,0 m dyn. arit. MG5_aMC+Py8 m m NLOinc QCD, g m NLOinc QCD, a m NLOinc QCD,
Phys. Rev. D 99 (2019) 072009 -1 = 36fb int ATLAS, L
Preliminary
ATLAS = 13 TeVs jets N [f b ] j e t s d N s d (ptj=8 GeV)MG5_aMC+Py8 FxFx = 30 GeV Q m = 20 GeV Q m = 40 GeV Q m Generator Level
ATLAS = 13 TeVs Wtt fi pp jets N R a t i o (a) (b) jets N [f b ] j e t s d N s d MG5_aMC+Py8 FxFxSherpa 2.2.8 MEPS@NLOaMC+Py8 scale var. ME+PSSherpa scale var. ME+PS
Generator Level
ATLAS = 13 TeVs Wtt fi pp jets N R a t i o [GeV] T Leading jet p ] - G e V (cid:215) [f b T dp s d MG5_aMC+Py8 FxFxSherpa 2.2.8 MEPS@NLOaMC+Py8 scale var. ME+PSSherpa scale var. ME+PS
Generator Level
ATLAS = 13 TeVs Wtt fi pp [GeV] T Leading jet p R a t i o (c) (d) Figure 2: (a): Cross–section dependence with three functional forms of the µ R and µ F scales. (b): MG5 aMC@NLO
FxFx samples parameter variations. (c) and (d):Comparisons between the
MG5 aMC@NLO and
Sherpa2.2.8
MC generators. Thevertical error lines show the 7–point scale variations for (a), while for the rest theyindicate the MC statistical uncertainty and the shaded bands represent these scalevariations. Figures from Ref. [11].
From these studies including higher order effects in both QCD and EW it is clear thatwe still don’t have the whole picture for the ttW process. The choice of the functional3 W ) [f b ]tt fi ( pp s Preliminary
ATLAS = 13 TeVs
Phys. Rev. D 99 (2019) 072009 -1 = 36fb int ATLAS, L )/2 W + m top = (2m m,0 m fixed T,i m i (cid:229) T = H a,0 m dyn. arit. MG5_aMC+Py8 NLOinc m m QCD+tree-level EW, m m QCD, a m QCD+tree-level EW, a m QCD, m / m R a t i o Q CD Q CD + E W jets N [f b ] j e t s d N s d MG5_aMC+Py8 NLOinc
QCDQCD+tree-level EW
Generator Level
ATLAS = 13 TeVs Wtt fi pp jets N R a t i o (a) (b) forward jets N [f b ] f o r w a r d j e t s d N s d MG5_aMC+Py8 NLOinc
QCDQCD+tree-level EW
Generator Level
ATLAS = 13 TeVs Wtt fi pp forward jets N R a t i o jet h max h [f b ] j e t h m a x h d s d MG5_aMC+Py8 NLOinc
QCDQCD+tree-level EW
Generator Level
ATLAS = 13 TeVs Wtt fi pp jet h max h R a t i o (c) (d) Figure 3: (a): Cross–section dependence with two functional forms of the µ R and µ F scales. (b) to (d): Effect of the “tree–level EW” contribution for MG5 aMC@NLO for some kinematic variable distributions. The vertical error lines show the 7–pointscale variations for (a), while for the rest they indicate the MC statistical uncertaintyand the shaded bands represent these scale variations. Figures from Ref. [11].form of the µ R and µ F scales as well as their values can change the predictionssubstantially. The addition of multileg setups and EW corrections also have a 10%impact on the cross–section values. The former seem to be in agreement across4ifferent MC generators and also consistent within relevant parameter variations (suchas µ Q ); while the latter further increase the cross–section and have considerable shapeeffects in some kinematic distributions. These results are documented in Ref. [11].Copyright 2021 CERN for the benefit of the ATLAS Collaboration. CC-BY-4.0license. References [1] F. Maltoni, M. L. Mangano, I. Tsinikos and M. Zaro, Phys. Lett. B (2014),252-260 doi:10.1016/j.physletb.2014.07.033 [arXiv:1406.3262 [hep-ph]].[2] CMS Collaboration, JHEP (2018), 011 doi:10.1007/JHEP08(2018)011[arXiv:1711.02547 [hep-ex]].[3] ATLAS Collaboration, Phys. Rev. D (2019) no.7, 072009doi:10.1103/PhysRevD.99.072009 [arXiv:1901.03584 [hep-ex]].[4] D. de Florian et al. [LHC Higgs Cross Section Working Group],doi:10.23731/CYRM-2017-002 [arXiv:1610.07922 [hep-ph]].[5] J. Alwall, R. Frederix, S. Frixione, V. Hirschi, F. Maltoni, O. Mattelaer,H. S. Shao, T. Stelzer, P. Torrielli and M. Zaro, JHEP (2014), 079doi:10.1007/JHEP07(2014)079 [arXiv:1405.0301 [hep-ph]].[6] T. Sj¨ostrand, S. Ask, J. R. Christiansen, R. Corke, N. Desai, P. Ilten, S. Mrenna,S. Prestel, C. O. Rasmussen and P. Z. Skands, Comput. Phys. Commun. (2015), 159-177 doi:10.1016/j.cpc.2015.01.024 [arXiv:1410.3012 [hep-ph]].[7] R. Frederix and S. Frixione, JHEP (2012), 061 doi:10.1007/JHEP12(2012)061[arXiv:1209.6215 [hep-ph]].[8] E. Bothmann et al. [Sherpa], SciPost Phys. (2019) no.3, 034doi:10.21468/SciPostPhys.7.3.034 [arXiv:1905.09127 [hep-ph]].[9] C. G¨utschow, J. M. Lindert and M. Sch¨onherr, Eur. Phys. J. C (2018) no.4,317 doi:10.1140/epjc/s10052-018-5804-2 [arXiv:1803.00950 [hep-ph]].[10] R. Frederix, D. Pagani and M. Zaro, JHEP02