| 1 |
\section{Signal and Background Modeling}
|
| 2 |
\label{sec:gen}
|
| 3 |
\subsection{Monte Carlo generators}
|
| 4 |
The signal and background samples for the full detector simulation
|
| 5 |
are generated with the leading order (LO) event generators
|
| 6 |
{\sl PYTHIA}~\cite{Sjostrand:2003wg}, {\sl ALPGEN} and {\sl COMPHEP}.
|
| 7 |
To accommodate the next-to-leading (NLO) effects, constant $k$-factors are applied
|
| 8 |
except for the signal where a $p_T$-dependence has been taken into account
|
| 9 |
and some of the backgrounds, $e.g.$ $t\bar{t}$, $W+jets$, and $Z+jets$ samples,
|
| 10 |
officially produced with NLO effects taken into account.
|
| 11 |
|
| 12 |
The $p_T$-dependent $k$-factor for the signal is estimated using
|
| 13 |
the NLO cross section calculator {\sl MCFM}~\cite{Campbell:2005}.
|
| 14 |
We estimate the PDF uncertainty on the cross-section using
|
| 15 |
{\sl MC@NLO 3.1}~\cite{Frixione:2002ik} NLO event generator
|
| 16 |
together with CTEQ6M PDF set.
|
| 17 |
|
| 18 |
\subsection{Signal definition}
|
| 19 |
The goal of this analysis is to study the associative production of the on-shell
|
| 20 |
$\W$ and $\Z$ bosons that decay into three leptons and a neutrino. In the
|
| 21 |
following we refer to a lepton to as either a muon or an electron, unless
|
| 22 |
specified otherwise.
|
| 23 |
|
| 24 |
Using {\sl MCFM} we estimate the total NLO $\WZ$ cross-section to be
|
| 25 |
\begin{equation}
|
| 26 |
\sigma_{NLO} ( pp \rightarrow W^+\Z; \sqrt{s}=14~{\rm TeV}) = 30.5~{\rm pb},
|
| 27 |
\end{equation}
|
| 28 |
\begin{equation}
|
| 29 |
\sigma_{NLO} ( pp \rightarrow W^-\Z; \sqrt{s}=14~{\rm TeV}) = 19.1~{\rm pb}.
|
| 30 |
\end{equation}
|
| 31 |
|
| 32 |
The LO and NLO distributions of the \Z boson transverse momentum are
|
| 33 |
shown in Fig.~\ref{fig:LOvsNLO} with the case of $W^+$ on the left and $W^-$
|
| 34 |
on the right side. The NLO/LO ratio, $k$-factor, is also presented on the figure,
|
| 35 |
and it is increasing with $p_T(\Z)$. We take into account the $p_T$ dependence
|
| 36 |
by re-weighting the LO Monte Carlo simulation as a function of the $p_T(\Z)$.
|
| 37 |
%
|
| 38 |
%
|
| 39 |
%The $p_T$ dependence of the $k$-factor
|
| 40 |
%becomes important when a proper NLO description of the $\Z$ boson transverse
|
| 41 |
%momentum must be obtained, $e.g$ to measure the strength of the $WWZ$ coupling.
|
| 42 |
%As the focus of this analysis is to prepare for the cross-section measurement,
|
| 43 |
%we take a $p_{T}$-averaged value of the $k$-factor, equal to 1.84.
|
| 44 |
|
| 45 |
\begin{figure}[!bt]
|
| 46 |
\begin{center}
|
| 47 |
\scalebox{0.8}{\includegraphics{figs/LOvsNLOZPtWminuns.eps}\includegraphics{figs/LOvsNLOZPtWplus.eps}}
|
| 48 |
\caption{$p_T(Z)$ distribution for LO (solid black histogram) and NLO (dashed black histogram)
|
| 49 |
in $W^-\Z$ events (left) and $W^+\Z$ events (right). The ratio NLO/LO is also given as a red
|
| 50 |
solid line.
|
| 51 |
}
|
| 52 |
\label{fig:LOvsNLO}
|
| 53 |
\end{center}
|
| 54 |
\end{figure}
|
| 55 |
|
| 56 |
%# for bbll:
|
| 57 |
%#CS NLO ((Z/gamma*->l+l-)bb) = 830pb = 345 pb * 2.4, where:
|
| 58 |
%#- 345 pb is LO CS calculated with precision of ~0.15%
|
| 59 |
%#- 2.4 is MCMF calculated k-factor with precision ~30% (!)
|
| 60 |
%# 830x0.173 (== XS x eff.) = 143.59pb
|
| 61 |
|
| 62 |
|
| 63 |
\subsection{Signal and background Monte Carlo samples}
|
| 64 |
|
| 65 |
The signal Monte Carlo sample is produced using {\sl PYTHIA}
|
| 66 |
generator. The decay for the \W lepton is forced to $e\nu_e$,
|
| 67 |
$\mu\nu_{\mu}$ or $\tau\nu_{\tau}$ final state, while the \Z decays
|
| 68 |
into electrons or muons only.
|
| 69 |
|
| 70 |
The background to the \WZ final state can be divided in physics and
|
| 71 |
instrumental. The only physics background is from $Z^0Z^0$ production
|
| 72 |
where one of the leptons is either mis-reconstructed or lost.
|
| 73 |
|
| 74 |
The instrumental backgrounds are all due to mis-identified electron candidates
|
| 75 |
from either jets or photons. These backgrounds include production of $\W$ and $\Z$ bosons
|
| 76 |
with jets and $t\bar{t}$ processes and $Z^0\gamma$ process. The background from $W\gamma$
|
| 77 |
production, where the $\gamma$ converts and produces a dielectron system is neglected
|
| 78 |
due to a requirement on the $\ell^+\ell^-$ invariant mass to be consistent with the nominal \Z boson mass.
|
| 79 |
|
| 80 |
All non-negligible instrumental backgrounds are summarized below.
|
| 81 |
\begin{itemize}
|
| 82 |
\item $\Z + jets$: this background is one of the dominant to the \WZ final state. Although
|
| 83 |
the misidentification rate for a jet to be misidentified as a lepton is quite small, the
|
| 84 |
$\Z+jets$ cross-section is 35 times larger than the signal one. We use the {\sl ALPGEN}
|
| 85 |
generated official samples of $\Z+jet$ production Monte Carlo samples for different
|
| 86 |
values of the jet transverse momentum.
|
| 87 |
\item $t\bar{t}$: each of the top quarks decay into a $\W b$ pair producing at least two
|
| 88 |
leptons and two $b$-quark jets. Although this process does not have a genuine $\Z$
|
| 89 |
candidate and can be suppressed be a $\Z$ candidate invariant mass requirement,
|
| 90 |
the probability for a $b$-quark jet to decay semi-leptonically and be misidentified
|
| 91 |
as a lepton is higher than that from a light-quark jets. The cross-section of the $t\bar{t}$
|
| 92 |
production is also exceed by about 15 times the cross-section of the \WZ production.
|
| 93 |
Thus, this background is also one of the most dominant. We use the official $t\bar{t}$
|
| 94 |
samples produced with {\sl ALPGEN} generator to estimate this background.
|
| 95 |
\item $\Z + b\bar{b}$: this process is produced by the {\sl COMPHEP}
|
| 96 |
generator and have a genuine $\Z$ candidate in the final state. One of the $b$-quark
|
| 97 |
jets are misidentified as the third lepton from the $\W$ boson.
|
| 98 |
\item $\W+jets$: in this process, the \W boson produces a genuine lepton,
|
| 99 |
while the other two leptons are misidentified jets. As the misidentification
|
| 100 |
probability is low, this channel does not contribute significantly to the \WZ
|
| 101 |
final state. The additional \Z candidate invariant mass requirement suppresses
|
| 102 |
this background further. We use the officially produced sample of $\W+jets$ processes
|
| 103 |
for different number of jets in the final state generated by the {\sl ALPGEN}
|
| 104 |
generator.
|
| 105 |
\item $Z^0\gamma$: this process is calculated with {\sl PYTHIA}.
|
| 106 |
\end{itemize}
|
| 107 |
|
| 108 |
All the samples we use in this study are a part of the CSA07 production and
|
| 109 |
are generated using $\mathrm{CMSSW}\_1\_4_\_6$ using the full {\sl GEANT}
|
| 110 |
simulation of the CMS detector. The digitization and reconstruction are
|
| 111 |
done using a newer $\mathrm{CMSSW}\_1\_6_\_7$ release with a
|
| 112 |
misalignment/miscalibration of the detector scenario expected
|
| 113 |
to be achieved after collection of $\sim$ 100~pb$^{-1}$ of data.
|
| 114 |
All {\sl ALPGEN} samples are mixed together in further referred to as to a
|
| 115 |
``Chowder soup''.
|
| 116 |
|
| 117 |
The summary of all datasets used for signal and background is given in
|
| 118 |
Table~\ref{tab:MC}. We use the RECO production level to access to
|
| 119 |
low-level detector information, such as reconstructed hits. This lets
|
| 120 |
us to use full granularity of the CMS sub-detectors to use isolation
|
| 121 |
discriminants.
|
| 122 |
|
| 123 |
Analysis of the samples is done using CMSSW$\_1\_6\_7$ CMS software
|
| 124 |
release. The information is stored in ROOT trees using a code in
|
| 125 |
CVS:/UserCode/Vuko/WZAnalysis, which is based on Physics Tools candidates.
|
| 126 |
|
| 127 |
\begin{table}[!tb]
|
| 128 |
%\begin{tabular}{llllll} \hline
|
| 129 |
%Sample & Generator & Sample name & Events & $\sigma \cdot \epsilon
|
| 130 |
%\cdot k$ & k-factor \\ \hline WZ & Pythia &
|
| 131 |
%/WZ/CMSSW\_1\_6\_7-CSA07-1195663763/RECO & 58897 & 0.585 pb & 1.92 \\
|
| 132 |
%$Zb\bar{b}$ & COMPHEP &
|
| 133 |
%/comphep-bbll/CMSSW\_1\_6\_7-CSA07-1198677426/RECO & 143.59 pb & 2.4
|
| 134 |
%\\ ``Chowder'' & ALPGEN &
|
| 135 |
%/CSA07AllEvents/CMSSW\_1\_6\_7-CSA07-Chowder-A1-PDAllEvents-ReReco-100pb/RECO
|
| 136 |
%& 25 M & event weights & - \\
|
| 137 |
\begin{tabular}{|c|c|c|c|c|} \hline
|
| 138 |
Sample & cross section, pb & Events & Dataset name \\ \hline
|
| 139 |
$\WZ$ & 1.12 & 59K & /WZ/CMSSW$\_1\_6\_7$-CSA07-1195663763\\ \hline
|
| 140 |
$\Z b\bar{b}$ & 830*0.173 (NLO) & 1.9M & /comphep-bbll/CMSSW$\_1\_6\_7$-CSA07-1198677426\\ \hline
|
| 141 |
Chowder & Event Weight & $\sim$ 21M & /CSA07AllEvents/\\ & & & CMSSW$\_1\_6\_7$-CSA07-Chowder-A1-PDAllEvents-ReReco
|
| 142 |
-100pb\\ \hline
|
| 143 |
$\Z\Z$ inclusive & 16.1 (NLO) & $\sim$ 140k & /ZZ$\_$incl/CMSSW$\_1\_6\_7$-CSA07-1194964234/RECO\\ \hline
|
| 144 |
$\Z\gamma \rightarrow e^+e^-\gamma$ & 1.08 (NLO) & $\sim$125k &/Zeegamma/CMSSW$\_1\_6\_7$-CSA07-1198935518/RECO \\ \hline
|
| 145 |
$\Z\gamma \rightarrow \mu^+\mu^-\gamma$ & 1.08 (NLO) & $\sim$ 93k & /Zmumugamma/CMSSW$\_1\_6\_7$-CSA07-1194806860/RECO\\ \hline
|
| 146 |
\end{tabular}
|
| 147 |
\label{tab:MC}
|
| 148 |
\caption{Monte Carlo samples used in this analysis using 100 pb$^{-1}$ scenario}
|
| 149 |
\end{table}
|
| 150 |
|
| 151 |
|
| 152 |
|
| 153 |
|
