HerwigDibosonHighPtMC

Introduction

This page is intended to provide information about samples of Herwig diboson production with high-pT boson filter to improve statistics at high-pT/high-mass regions. The samples are primarily used for diboson resonance searches in the WZ/ZZ -> qqll channel, so one of the Z-boson in the WZ/ZZ processes is forced to decay into leptons (electrons or muons). The WW process is not considered here as it is negligible after applying the baseline selection cuts for the llqq channel.

Herwig WZ/ZZ Samples

The official Herwig inclusive WZ/ZZ samples (DSID 105986-7) are produced with the following conditions:

• Process ID 12820 (for WZ), 12810 (for ZZ)
• Both W and Z bosons decay inclusively
• Single truth lepton (electron or muon) with pT > 10 GeV and |eta| < 2.8 is required
• CTEQ6L1 PDF and AUET2 tune

About 1 M (250 K) events (after lepton filter) have been produced in MC12a for the WZ (ZZ) sample, corresponding to about 270 (250) fb-1 of luminosity. They are used in the llqq search, but as presented in e.g, the DBL pre-approval meeting on 20 Dec, 2013 (link), the statistics is not sufficient to model those backgrounds in the high-pT resolved and merged signal regions (slide 15).

Additional WZ/ZZ Samples

Based on this, we have looked into additional MC samples with minimum boson-pT cuts at the generator level. The pT cuts are chosen such that the high-pT resolved and merged regions are most efficiently populated while avoiding bias caused by the cut on the second boson in the events. As mentioned above at least one of Z bosons in events is forced to decay into electrons or muons, because the WZ with Z->qq or ZZ->qqqq processes have negligible contribution after the final llqq analysis selections. The same PDF, underlying event tune and the lepton filter as those used in the official inclusive samples are used.

Boson pT-slicing with PTMIN and PTMAX cuts is checked but it turns out that the PTMAX does not work in diboson production.

The dilepton-pT filter (similar to the high-mass dilepton filter in concept) is also considered instead of boson-pT cut, but the dilepton cut could cause bias to events in WZ with leptonic W decay or ZZ->4l events. So, this option is not used here.

Monte Carlo Studies

Monte Carlo truth studies to determine the boson pT cuts and required statistics are performed by generating events in AtlasProduction 17.8.0.1. The pT of dilepton system and the 3-object (4-object) mass for the merged (high-pT resolved) regime with and without boson-pT cuts are checked at the truth level. The Z-mass window cut (mZ +/- 25 GeV) is applied for all the plots, and the standard high-pT resolved and merged signal selections (including the pT(Z) cuts of >250 and >400 GeV, respectively), as described in ATL-COM-PHYS-2013-1556, are applied to the mass distributions. Below is the summary table of the samples used in this study : cross section times branching fraction, lepton filter efficiency, luminosity, number of events after the lepton filter. The inclusive samples include all W and Z boson decays while the samples with boson pT cuts include at least one Z-boson with Z->ll decay (included in the branching fraction in the table).

WZ Production

Boson pT Cut	JobOption	xsec x BR	Lepton Filter Efficiency	Luminosity [fb-1]	#Events (after lepton filter)
Inclusive	JO	12.01 pb	0.305	327	1.2M
>150 GeV	JO	25.2 fb	0.997	3983	100K
>300 GeV	JO	1.79 fb	1.000	22288	40K

ZZ Production

Boson pT Cut	JobOption	xsec x BR	Lepton Filter Efficiency	Luminosity [fb-1]	#Events (after lepton filter)
Inclusive	JO	4.69 pb	0.211	606	600K
>150 GeV	JO	22.6 fb	0.998	4437	100K
>300 GeV	JO	1.63 fb	1.000	24562	40K

The numbers of events in these tables are for the samples used in this study, not for the requested samples. The requested number of events will be given later. The official inclusive WZ and ZZ samples (DSID 105986-7) consist of 1M and 250K events, respectively.

High-pT Resolved Channel

Comparison between inclusive and boson-pT > 150 GeV samples:

Channel	Validation plots
ee	dilepton pT (log-scale, lin-scale), 4-object mass
mumu	dilepton pT (log-scale, lin-scale), 4-object mass

Merged Channel

Comparison between inclusive and boson-pT > 300 GeV samples:

Channel	Validation plots
ee	dilepton pT (log-scale, lin-scale), 3-object mass
mumu	dilepton pT (log-scale, lin-scale), 3-object mass

Requested Statistics

Given the smoothness of the boson-pT sample distributions, it appears to be sufficient to have about a factor 5 more events that those used to make the plots. This would lead to 500K (200K) events each for the electron and muon channels for the boson-pT cut of >150 (300) GeV, in total of 1.4M events.

---

Major updates:
-- KojiTerashi - 29 Dec 2013