t\bar{t}W^\pm at NLO accuracy with realistic final states
MMarch 1, 2021 ttW ± at NLO accuracy with realistic final states Giuseppe Bevilacqua MTA-DE Particle Physics Research GroupH-4002 Debrecen, PO Box 400, Hungary
In this proceedings we summarize state-of-the-art theoretical predictions for theprocess of ttW ± production at the LHC. After a review of the status of inclusivecalculations, we discuss recent advancements in exclusive predictions.PRESENTED AT th International Workshop on Top Quark PhysicsDurham, UK (videoconference), 14–18 September, 2020 Work supported by grant K 125105 of the National Research, Development and Innovation Office in Hun-gary. a r X i v : . [ h e p - ph ] F e b Introduction
The associated production of a tt pair and a W ± gauge boson ( ttW ) is one of the heaviestprocesses that can be studied at the LHC using the full Run II data. For this reason it is anoptimal candidate for probing the electroweak (EW) scale by means of precision measurementsof top-quark couplings to EW bosons [1, 2]. Particularly interesting are the experimental sig-natures induced by leptonic decays of top quarks and of the W boson, which can lead to finalstates with three light leptons or two same-charge light leptons. The latter are relatively rarein the Standard Model (SM) and thus well suited to look for signatures of new physics as pre-dicted by several BSM scenarios, such as Supersymmetry, minimal Universal Extra Dimensionsor non-standard Higgs models (see e.g. [3–5]). Other than being interesting in its own right, ttW is also an important background for Higgs boson searches in the ttH channel as well as forfour-top production. Last but not least, the absence of contributions from gg -initiated processesup to NNLO in QCD confers distinctively large charge asymmetries to ttW in comparison toinclusive tt production. Precision studies of these observables offer another opportunity to lookfor new physics [6, 7, 27].On the experimental side, the ATLAS and CMS Collaborations have carried out directmeasurements of the cross section for ttW ± production at 13 TeV using an integrated luminosityof L = 36 fb − [8, 9]. More recently, ttW has been scrutinized in the context of ttH searches inmultilepton final states with L = 80 fb − [10] and L = 137 fb − [11]. It also played a role in thefirst evidence of tttt production based on L = 139 fb − [12]. All these analyses report sustainedtensions where the measured ttW rates exceed SM predictions by 25% −
70% depending on thesignal cathegory. This raises the need for increased theoretical accuracy.On the theory side, several directions have been undertaken in efforts for improving thedescription of this process at both inclusive and exclusive levels. This proceedings brieflysummarizes the present status and future perspectives of the theoretical modeling of ttW . ttW ± production When dealing with fully inclusive predictions, the approximation of stable top quarks and W bosons is usually preferred to keep the computational burden of higher-order corrections assmall as possible. This strategy allows to achieve predictions for total production rates withparticularly high accuracy. At the present state-of-the-art, the ttW cross section is knownperturbatively up to the complete-NLO , which consists of all possible QCD and EW correc-tions to the full LO cross section [13–16]. The various NLO contributions are classified intofour groups according to the perturbative order: O ( α s α ) (dubbed NLO ), O ( α s α ) ( NLO ), O ( α s α ) ( NLO ) and O ( α ) ( NLO ). By power counting one would naively expect that thecontributions above appear in order of decreasing importance, however as shown in Ref. [16]a different hierarchy is realized. While NLO dominates as expected ( ∼
50% of the LO crosssection),
NLO ( ∼ NLO ( ∼ − NLO (permille level). The importance of NLO is understood as due tothe opening of diagrams embedding the electroweak subprocess of tW → tW scattering, whichis particularly sensitive to the large top-quark Yukawa coupling [1, 16].1omplete-NLO results have been also matched to threshold resummation of soft gluoncorrections up to NNLL accuracy [17–21], resulting in the most accurate predictions to date fortotal production rates. After resummation, the ttW cross section receives a modest shift of fewpercents in the central value and no substantial reduction of scale uncertainties, which remainat the level of 10-20% (see Figure 1, left plot). For comparison, in the case of ttZ productionthe central predictions are brought much closer to data and theoretical errors are reduced byalmost half [20]. The weaker impact of resummation on ttW is in line with the fact that thisprocess, in contrast with the gg -dominated ttZ , receives only contributions from qq (cid:48) channel atLO. Correspondingly only soft-gluon emissions from incoming quark lines are considered in theresummation frameworks.To summarize this part, sub-leading EW contributions impact on the ttW cross sectionat the level of 10% and must be necessarily included in experimental analyses. The presentstate-of-the-art description (NLO+NNLL) does not solve the slight tension observed betweentheoretical predictions and measured production rates. Much of the attention is on the po-tential role of missing higher-order corrections, mainly from NNLO QCD. Systematic analysesof ttW j and ttW jj subprocesses, based on NLO multi-jet (FxFx) matching, show that higherjet multiplicities impact the total cross section at the level of 10% [22, 23] (see Figure 1, rightplot). This raises a strong motivation for a full NNLO QCD calculation of ttW which, althoughhighly desirable, remains beyond reach. N L O N L O+ N LL N L O+ N LL wC N L O+ NN LL σ ( pp → t ¯ tW + X ) [fb] √ S = 13 TeV µ = M/ µ = H T / µ = Q/ µ = H T µ = Q
600 800 1000 1200 1400) [fb] – Wt(t s Z ) [f b ] t ( t s NLOEW sd (This work) + FxFx1jQCD s - – Wtt
NLOEW sd + NLOQCD s - – Wtt
NLOEW sd (This work) + FxFx1jQCD s Z - tt
NLOEW sd + NLOQCD s Z - tt s – ATLAS+CMS best-fit ‹ [CONF-2019-045] -1 = 80 fb ATLASttW L [CONF-2018-047] -1 =36.1 fb ATLASttZ L [PAS-HIG-19-008] -1 =137 fb CMSttV L Figure 1:
Left plot: impact of soft-gluon resummation on cross section predictions for ttW ± at √ s = 13 TeV [21] . Right plot: comparison of ttW and ttZ total cross sections at NLOQCD+EW as well as with FxFx matching [22].2 ttW ± with realistic final states Since the lifetime of top quarks and W bosons is extremely short, including effects of decays is anequally important side of the modelling process. Spin correlations should be taken into accountas they can have a sizeable impact on the decay products, indeed top-quark polarization effectsare particularly important in ttW production [6]. Also, in view of more accurate comparisonswith LHC data, higher-order effects should be included in decays when possible. First attemptsin this direction have been made in Ref. [29] with the first NLO calculation in full Narrow WidthApproximation (NWA), including QCD corrections to both production and decay stages.At the same time, continuous efforts have been made to improve the modelling of hadronicobservables by interfacing calculations with parton showers. The first study at NLO was pre-sented in Ref. [30] using the POWHEG method as implemented in the PowHel framework.The
NLO accuracy was employed for the production process and LO for spin-correlated de-cays. Studies based on MC@NLO matching have been also carried out using same perturbativeaccuracy [6, 14]. In a number of recent papers [7, 23, 24] the impact of the sub-leading NLO contributions has been examined in multilepton final states. It is worth to notice that NLO ,like NLO , can be matched to available shower frameworks such as PYTHIA8 since it rep-resents pure QCD corrections to ttW . In contrast
NLO and NLO contributions - not yetconsidered at this stage - are genuine EW corrections and a consistent matching will requireto include EW effects in the shower framework. NLO effects have not always a flat impacton the phase space and may lead to visible shape distortions. This has been reported e.g. forjet multiplicities or dijet invariant mass distributions [7, 24]. In these cases, including NLO as a flat +10% rescaling is clearly not as accurate as including the same effects in a differentialmanner.Complementary to parton shower aspects, another direction which can be tackled for im-proving final-state modelling is to overcome the limit Γ t /m t → on-shell predictions. In off-shell calculations, the complete set of resonantand non-resonant diagrams is computed for a given final state ( e.g. bbe + ν e e − ν e µ − ν µ + X formultilepton channels). Based as it is on a complete calculation at fixed perturbative order, it isalso more challenging compared to NWA. Off-shell effects are expected to impact inclusive crosssections at the level of O (Γ t /m t ) ≈ .
8% but can reach tens of percents in tails of distributions.Two independent off-shell calculations of ttW in the multilepton channel are currently availableat NLO QCD accuracy [25–27]. Very recently further progress has been obtained with the fullcombination of QCD+EW effects in the same channel [28].In Ref. [25], a systematic comparison between off-shell predictions and NWA has been carriedout for different accuracies in the modelling of decays: LO (”NWA
LOdecay ”) and NLO QCD(”full NWA”). Inclusive fiducial cross sections show a very good agreement: off-shell resultsdiffer from NWA
LOdecay by 3-5% and just by 0.2% from the full NWA. These numbers shouldbe compared to theoretical uncertainties estimated from scale variation, which amount to 11%(NWA
LOdecay ), 7% (full NWA) and 7% (off-shell). Interestingly, the accuracy of decays is foundto impact scale uncertainties in NWA. Much larger discrepancies are observed between off-shellpredictions and NWA when looking at differential cross sections: as an example we highlightthe case of the transverse momentum ( p T ) of the leading b -jet. The tail of the distribution3s enhanced up to 50% by off-shell effects, to be compared with ∼
20% scale uncertainties inthe same region [25]. An alternative approach for assessing non-resonant effects is discussedin Ref. [26], where Double Pole Approximation (DPA) is used for the virtual and for the I -operator contributions to the NLO cross section. Comparing with full off-shell results, thefiducial cross-section agrees at percent-level while discrepancies up to 10% are found in regionsnot dominated by the tt resonance. Both Refs. [25, 26] propose dynamical scales which help toobtain moderate QCD corrections and reduced shape distortions for most observables. Thereare notable exceptions, e.g. the p T of the bb system, where QCD corrections reach up to300% in suppressed phase space regions [26]. The large K -factors are consistent with thedominance of effects from real radiation, which is also confirmed by the observation of LO-likescale uncertainties in the same regions. Sustained tensions with measurements of enhanced production rates at the LHC raise the needfor improved theoretical modeling of the ttW process. The state-of-the-art accuracy for inclusivepredictions is NLO+NNLL. On a more exclusive ground, parton-shower Monte Carlo generatorscan now model on-shell ttW production including EW effects at O ( α s α ). Additionally, theimpact of off-shell and non-resonant effects in ttW multilepton final states can be studied bymeans of complete fixed-order calculations.Achieving full NNLO QCD predictions, while beyond present reach, will be an importantstep forward to reduce theoretical uncertainties as well as to assess the possible role of higher-order effects in the observed tension with data. In view of accurate comparisons with high-luminosity data, it will be also appropriate to include the full set of sub-leading EW effects inMonte Carlo generators. Matching off-shell calculations to parton showers is an ambitious goalwhich will lead to the most complete NLO description of ttW final states. References [1] J. A. Dror, M. Farina, E. Salvioni and J. Serra, JHEP (2016), 071[2] O. Bessidskaia Bylund, F. Maltoni, I. Tsinikos, E. Vryonidou and C. Zhang, JHEP (2016), 052[3] R. M. Barnett, J. F. Gunion and H. E. Haber, Phys. Lett. B (1993), 349-354[4] H. C. Cheng, K. T. Matchev and M. Schmaltz, Phys. Rev. D (2002), 056006[5] R. Contino and G. Servant, JHEP (2008), 026[6] F. Maltoni, M. L. Mangano, I. Tsinikos and M. Zaro, Phys. Lett. B (2014), 252-260[7] R. Frederix and I. Tsinikos, Eur. Phys. J. C (2020) no.9, 80348] A. M. Sirunyan et al. [CMS], JHEP (2018), 011[9] M. Aaboud et al. [ATLAS], Phys. Rev. D (2019) no.7, 072009[10] ATLAS Collaboration, ATLAS-CONF-2019-045.[11] CMS Collaboration, arXiv:2011.03652 [hep-ex].[12] ATLAS Collaboration, Eur. Phys. J. C (2020) no.11, 1085[13] V. Hirschi, R. Frederix, S. Frixione, M. V. Garzelli, F. Maltoni and R. Pittau, JHEP (2011), 044[14] F. Maltoni, D. Pagani and I. Tsinikos, JHEP (2016), 113[15] S. Frixione, V. Hirschi, D. Pagani, H. S. Shao and M. Zaro, JHEP (2015), 184[16] R. Frederix, D. Pagani and M. Zaro, JHEP (2018), 031[17] H. T. Li, C. S. Li and S. A. Li, Phys. Rev. D (2014) no.9, 094009[18] A. Broggio, A. Ferroglia, G. Ossola and B. D. Pecjak, JHEP (2016), 089[19] A. Broggio, A. Ferroglia, R. Frederix, D. Pagani, B. D. Pecjak and I. Tsinikos, JHEP (2019), 039[20] A. Kulesza, L. Motyka, D. Schwartl¨ander, T. Stebel and V. Theeuwes, Eur. Phys. J. C (2019) no.3, 249[21] A. Kulesza, L. Motyka, D. Schwartl¨ander, T. Stebel and V. Theeuwes, Eur. Phys. J. C (2020) no.5, 428[22] S. von Buddenbrock, R. Ruiz and B. Mellado, Phys. Lett. B (2020), 135964[23] ATLAS Collaboration, ATL-PHYS-PUB-2020-024.[24] F. F. Cordero, M. Kraus and L. Reina, arXiv:2101.11808 [hep-ph].[25] G. Bevilacqua, H. Y. Bi, H. B. Hartanto, M. Kraus and M. Worek, JHEP (2020), 043[26] A. Denner and G. Pelliccioli, JHEP (2020), 069[27] G. Bevilacqua, H. Y. Bi, H. B. Hartanto, M. Kraus, J. Nasufi and M. Worek,arXiv:2012.01363 [hep-ph].[28] A. Denner and G. Pelliccioli, arXiv:2102.03246 [hep-ph][29] J. M. Campbell and R. K. Ellis, JHEP (2012), 052[30] M. V. Garzelli, A. Kardos, C. G. Papadopoulos and Z. Trocsanyi, JHEP11