Wing Sheung Chan

86 Signal and background modelling 5.1. Monte Carlo simulations Different Monte Carlo samples are generated for different physical processes. These processes have either different final states or very little interference with each other, meaning that their contributions to the total event yields can simply be summed together (represented by histograms stacked on top of each other in plots). The signal Z → eτ and Z → µτ samples are simulated Z +jets events where the Z boson undergoes a two-body decay into a eτ or µτ final state. These events are generated using Pythia 8.2 [98] with matrix elements calculated at leading order (LO) in the strong coupling constant α s . Parameters for initial-state radiations, multiparton interactions and beam remnants are set according to the A14 set of tuned parameters (tune) [99] with the NNPDF2.3 LO parton distribution function (PDF) set [100] . Nominal signal samples are generated with unpolarised τ leptons, which correspond to the specific scenario where the effective Z`τ interaction vertex is parity-conserving. Scenarios with polarised τ leptons and parity-violating Z`τ vertices are considered by reweighting the nominal samples according to the probability of occurrence of each generated events under different assumptions. The reweighting method will be described in Section 5.2. The Z ( → τ τ ) +jets and W +jets events are generated using Sherpa 2.2.1 [101] with matrix elements calculated using the Comix [102] and OpenLoops [103– 105] libraries at next-to-leading-order (NLO) in α s for up to two partons, and at LO for up to four partons. They are matched with the Sherpa parton shower [106] using the MEPS@NLO prescription [107– 110] with the default Sherpa tune. The Z ( → `` ) +jets events are generated using Powheg-Box [111] for calculating the matrix elements and Pythia 8 for modelling the parton showers, hadronisation, and underlying-event activity ( Powheg+Pythia 8). The matrix elements are calculated at NLO in α s . The CT10 PDF set [112] is used for the hard-scattering processes, whereas the CTEQ6L1 PDF set [113] and the AZNLO tune [114] are used for the parton showers. The overall prefit (see Section 6.1) yield of each of the abovementioned Z -decay samples is normalised to match the measured Z -production cross section σ ( Z ) = 1 . 981 nb [62] . The t ¯ t , single-top and Wt events are generated using Powheg+Pythia 8 with the NNPDF3.0 NLO PDF set [115] and the A14 tune. The diboson events are generated using Sherpa 2.2.1 for purely leptonic decays and Sherpa 2.2.2 for semileptonic and semihadronic decays with the NNPDF3.0 next-to-next- to-leading-order (NNLO) PDF and a set-up for parton showers modelling similar to that for the Z → τ τ and W +jets samples. The H → τ τ and H → WW events are generated using Powheg+Pythia 8 with the NNPDF3.0 NNLO PDF set and a set-up for parton showers modelling similar to that for the Z → `` samples. Samples are generated separately for gluon-fusion and vector- boson-fusion productions. For gluon-fusion productions, the NNLOPS program [116, 117] is used to yield a fully exclusive, hadron-level description of the final state. No